The original code had an implicit assumption that if the test for
allocas or globals was reached, the two pointers were not equal. With my
changes to make the pointer analysis more powerful here, I also had to
guard against circumstances where the results weren't useful. That in
turn violated the assumption and gave rise to a circumstance in which we
could have a store with both the queried pointer and stored pointer
rooted at *the same* alloca. Clearly, we cannot ignore such a store.
There are other things we might do in this code to better handle the
case of both pointers ending up at the same alloca or global, but it
seems best to at least make the test explicit in what it intends to
check.
I've added tests for both the alloca and global case here.
llvm-svn: 220190
r220178. First, the creation routine doesn't insert prior to the
terminator of the basic block provided, but really at the end of the
basic block. Instead, get the terminator and insert before that. The
next issue was that we need to ensure multiple PHI node entries for
a single predecessor re-use the same cast instruction rather than
creating new ones.
All of the logic here was without tests previously. I've reduced and
added a test case from the test suite that crashed without both of these
fixes.
llvm-svn: 220186
logic to look through pointer casts, making them trivially stronger in
the face of loads and stores with intervening pointer casts.
I've included a few test cases that demonstrate the kind of folding
instcombine can do without pointer casts and then variations which
obfuscate the logic through bitcasts. Without this patch, the variations
all fail to optimize fully.
This is more important now than it has been in the past as I've started
moving the load canonicialization to more closely follow the value type
requirements rather than the pointer type requirements and thus this
needs to be prepared for more pointer casts. When I made the same change
to stores several test cases regressed without logic along these lines
so I wanted to systematically improve matters first.
llvm-svn: 220178
of InstCombine rather than just the bits enabled when datalayout is
optional.
The primary fixes here are because now things are little endian.
In good news, silliness like this seems like it will be going away as
we've got pretty stong consensus on dropping optional datalayout
entirely.
llvm-svn: 220176
loads.
This handles many more cases than just the AA metadata, some of them
suggested by Hal in his review of the AA metadata handling patch. I've
tried to test this behavior where tractable to do so.
I'll point out that I have specifically *not* included a test for
debuginfo because it was going to require 2 or 3 times as much work to
craft some input which would survive the "helpful" stripping of debug
info metadata that doesn't match the desired schema. This is another
good example of why the current state of write-ability for our debug
info metadata is unacceptable. I spent over 30 minutes trying to conjure
some test case that would survive, even copying from other debug info
tests, but it always failed to survive with no explanation of why or how
I might fix it. =[
llvm-svn: 220165
up to where it actually works as intended. The problem is that
a GlobalAlias isa GlobalValue and so the prior block handled all of the
cases.
This allows us to constant fold based on the actual constant expression
in the global alias. As an example, see the last function in the newly
added test case which explicitly aligns an unaligned pointer using
constant expression math. Without this change, we fail to see that and
fold an alignment test to zero.
llvm-svn: 220164
The following implements the transformation:
(sub (or A B) (xor A B)) --> (and A B).
Patch by Ankur Garg!
Differential Revision: http://reviews.llvm.org/D5719
llvm-svn: 220163
The following implements the optimization for sequences of the form:
icmp eq/ne (shl Const2, A), Const1
Such sequences can be transformed to:
icmp eq/ne A, (TrailingZeros(Const1) - TrailingZeros(Const2))
This handles only the equality operators for now. Other operators need
to be handled.
Patch by Ankur Garg!
llvm-svn: 220162
by my refactoring of this code.
The method isSafeToLoadUnconditionally assumes that the load will
proceed with the preferred type alignment. Given that, it has to ensure
that the alloca or global is at least that aligned. It has always done
this historically when a datalayout is present, but has never checked it
when the datalayout is absent. When I refactored the code in r220156,
I exposed this path when datalayout was present and that turned the
latent bug into a patent bug.
This fixes the issue by just removing the special case which allows
folding things without datalayout. This isn't worth the complexity of
trying to tease apart when it is or isn't safe without actually knowing
the preferred alignment.
llvm-svn: 220161
...)) and (load (cast ...)): canonicalize toward the former.
Historically, we've tried to load using the type of the *pointer*, and
tried to match that type as closely as possible removing as many pointer
casts as we could and trading them for bitcasts of the loaded value.
This is deeply and fundamentally wrong.
Repeat after me: memory does not have a type! This was a hard lesson for
me to learn working on SROA.
There is only one thing that should actually drive the type used for
a pointer, and that is the type which we need to use to load from that
pointer. Matching up pointer types to the loaded value types is very
useful because it minimizes the physical size of the IR required for
no-op casts. Similarly, the only thing that should drive the type used
for a loaded value is *how that value is used*! Again, this minimizes
casts. And in fact, the *only* thing motivating types in any part of
LLVM's IR are the types used by the operations in the IR. We should
match them as closely as possible.
I've ended up removing some tests here as they were testing bugs or
behavior that is no longer present. Mostly though, this is just cleanup
to let the tests continue to function as intended.
The only fallout I've found so far from this change was SROA and I have
fixed it to not be impeded by the different type of load. If you find
more places where this change causes optimizations not to fire, those
too are likely bugs where we are assuming that the type of pointers is
"significant" for optimization purposes.
llvm-svn: 220138
This test is pretty awesome. It is claiming to test devirtualization.
However, the code in question is not in fact devirtualized by LLVM. If
you take the original C++ test case and run it through Clang at -O3 we
fail to devirtualize it completely. It also isn't a sufficiently focused
test case.
The *reason* we fail to devirtualize it isn't because of any missing
instcombine though. Instead, it is because we fail to emit an available
externally vtable and thus the vtable is just an external and completely
opaque. If I cause the vtable to be emitted, we successfully
devirtualize things.
Anyways, I'm just removing it because it is providing negative value at
this point: it isn't representative of the output of Clang really, LLVM
isn't doing the transform it claims to be testing, LLVM's failure to do
the transform isn't actually an LLVM bug at all and we shouldn't be
testing for it here, and finally the test is written in such a way that
it will trivially pass even when the point of the test is failing.
llvm-svn: 220137
cases where the alloca type, the load types, and the store types used
all disagree.
Previously, the only way that vector-based promotion occured was if the
alloca type was a vector type. This was one of the *very* few remaining
uses of the alloca's type to guide SROA/mem2reg left in LLVM. It turns
out it was a bad idea.
The alloca type can change very easily based on the mixture of types
loaded and stored to that alloca. We shouldn't be relying on it as
a signal for very much. Instead, the source of truth should be loads and
stores. We should canonicalize the loads and stores as much as possible
and then rely on them exclusively in SROA.
When looking and loads and stores, we may find many different candidate
vector types. This change will let SROA try all of them to find a vector
type which is a viable way to promote the entire alloca to a vector
register.
With this change, it becomes possible to do better canonicalization and
optimization of loads and stores without breaking SROA in random ways,
and that should allow fixing a core source of performance loss in hot
numerical loops such as those in Eigen.
llvm-svn: 220116
This reverts commit r219899.
This also updates byval-tail-call.ll to make it clear what was breaking.
Adding r219899 again will cause the load/store to disappear.
llvm-svn: 220093
DSE's overlap checking contained special logic, used only when no DataLayout
was available, which inferred a complete overwrite when the pointee types were
equal. This logic seems fine for regular loads/stores, but does not work for
memcpy and friends. Instead of fixing this, I'm just removing it.
Philosophically, transformations should not contain enhanced behavior used only
when data layout is lacking (data layout should be strictly additive), and
maintaining these rarely-tested code paths seems not worthwhile at this stage.
Credit to Aliaksei Zasenka for the bug report and the diagnosis. The test case
(slightly reduced from that provided by Aliaksei) replaces the original
contents of test/Transforms/DeadStoreElimination/no-targetdata.ll -- a few
other tests have been updated to have a data layout.
llvm-svn: 220035
Summary:
Currently, call slot optimization requires that if the destination is an
argument, the argument has the sret attribute. This is to ensure that
the memory access won't trap. In addition to sret, we can also allow the
optimization to happen for arguments that have the new dereferenceable
attribute, which gives the same guarantee.
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D5832
llvm-svn: 219950
If a square root call has an FP multiplication argument that can be reassociated,
then we can hoist a repeated factor out of the square root call and into a fabs().
In the simplest case, this:
y = sqrt(x * x);
becomes this:
y = fabs(x);
This patch relies on an earlier optimization in instcombine or reassociate to put the
multiplication tree into a canonical form, so we don't have to search over
every permutation of the multiplication tree.
Because there are no IR-level FastMathFlags for intrinsics (PR21290), we have to
use function-level attributes to do this optimization. This needs to be fixed
for both the intrinsics and in the backend.
Differential Revision: http://reviews.llvm.org/D5787
llvm-svn: 219944
Make tail recursion elimination a bit more aggressive. This allows us to get
tail recursion on functions that are just branches to a different function. The
fact that the function takes a byval argument does not restrict it from being
optimised into just a tail call.
llvm-svn: 219899
For pointer-typed function arguments, enhanced alignment can be asserted using
the 'align' attribute. When inlining, if this enhanced alignment information is
not otherwise available, preserve it using @llvm.assume-based alignment
assumptions.
llvm-svn: 219876
If x is known to have the range [a, b) in a loop predicated by (icmp
ne x, a), its range can be sharpened to [a + 1, b). Get
ScalarEvolution and hence IndVars to exploit this fact.
This change triggers an optimization to widen-loop-comp.ll, so it had
to be edited to get it to pass.
phabricator: http://reviews.llvm.org/D5639
llvm-svn: 219834
Truncate the operands of a switch instruction to a narrower type if the upper
bits are known to be all ones or zeros.
rdar://problem/17720004
llvm-svn: 219832
The SLP vectorizer should not vectorize ephemeral values. These are used to
express information to the optimizer, and vectorizing them does not lead to
faster code (because the ephemeral values are dropped prior to code generation,
vectorized or not), and obscures the information the instructions are
attempting to communicate (the logic that interprets the arguments to
@llvm.assume generically does not understand vectorized conditions).
Also, uses by ephemeral values are free (because they, and the necessary
extractelement instructions, will be dropped prior to code generation).
llvm-svn: 219816
A few minor changes to prevent @llvm.assume from interfering with loop
vectorization. First, treat @llvm.assume like the lifetime intrinsics, which
are scalarized (but don't otherwise interfere with the legality checking).
Second, ignore the cost of ephemeral instructions in the loop (these will go
away anyway during CodeGen).
Alignment assumptions and other uses of @llvm.assume can often end up inside of
loops that should be vectorized (this is not uncommon for assumptions generated
by __attribute__((align_value(n))), for example).
llvm-svn: 219741
Eliminate library calls and intrinsic calls to fabs when the input
is a squared value.
Note that no unsafe-math / fast-math assumptions are needed for
this optimization.
Differential Revision: http://reviews.llvm.org/D5777
llvm-svn: 219717
We assumed that A must be greater than B because the right hand side of
a remainder operator must be nonzero.
However, it is possible for A to be less than B if Pow2 is a power of
two greater than 1.
Take for example:
i32 %A = 0
i32 %B = 31
i32 Pow2 = 2147483648
((Pow2 << 0) >>u 31) is non-zero but A is less than B.
This fixes PR21274.
llvm-svn: 219713
Reapply r216913, a fix for PR20832 by Andrea Di Biagio. The commit was reverted
because of buildbot failures, and credit goes to Ulrich Weigand for isolating
the underlying issue (which can be confirmed by Valgrind, which does helpfully
light up like the fourth of July). Uli explained the problem with the original
patch as:
It seems the problem is calling multiplySignificand with an addend of category
fcZero; that is not expected by this routine. Note that for fcZero, the
significand parts are simply uninitialized, but the code in (or rather, called
from) multiplySignificand will unconditionally access them -- in effect using
uninitialized contents.
This version avoids using a category == fcZero addend within
multiplySignificand, which avoids this problem (the Valgrind output is also now
clean).
Original commit message:
[APFloat] Fixed a bug in method 'fusedMultiplyAdd'.
When folding a fused multiply-add builtin call, make sure that we propagate the
correct result in the case where the addend is zero, and the two other operands
are finite non-zero.
Example:
define double @test() {
%1 = call double @llvm.fma.f64(double 7.0, double 8.0, double 0.0)
ret double %1
}
Before this patch, the instruction simplifier wrongly folded the builtin call
in function @test to constant 'double 7.0'.
With this patch, method 'fusedMultiplyAdd' correctly evaluates the multiply and
propagates the expected result (i.e. 56.0).
Added test fold-builtin-fma.ll with the reproducible from PR20832 plus extra
test cases to verify the behavior of method 'fusedMultiplyAdd' in the presence
of NaN/Inf operands.
This fixes PR20832.
llvm-svn: 219708
When LazyValueInfo uses @llvm.assume intrinsics to provide edge-value
constraints, we should check for intrinsics that dominate the edge's branch,
not just any potential context instructions. An assumption that dominates the
edge's branch represents a truth on that edge. This is specifically useful, for
example, if multiple predecessors assume a pointer to be nonnull, allowing us
to simplify a later null comparison.
The test case, and an initial patch, were provided by Philip Reames. Thanks!
llvm-svn: 219688
This is the same optimization of r219233 with modifications to support PHIs with multiple incoming edges from the same block
and a test to check that this condition is handled.
llvm-svn: 219656
We assumed that negation operations of the form (0 - %Z) resulted in a
negative number. This isn't true if %Z was originally negative.
Substituting the negative number into the remainder operation may result
in undefined behavior because the dividend might be INT_MIN.
This fixes PR21256.
llvm-svn: 219639
We have a transform that changes:
(x lshr C1) udiv C2
into:
x udiv (C2 << C1)
However, it is unsafe to do so if C2 << C1 discards any of C2's bits.
This fixes PR21255.
llvm-svn: 219634
Consider the case where X is 2. (2 <<s 31)/s-2147483648 is zero but we
would fold to X. Note that this is valid when we are in the unsigned
domain because we require NUW: 2 <<u 31 results in poison.
This fixes PR21245.
llvm-svn: 219568
consider:
C1 = INT_MIN
C2 = -1
C1 * C2 overflows without a doubt but consider the following:
%x = i32 INT_MIN
This means that (%X /s C1) is 1 and (%X /s C1) /s C2 is -1.
N. B. Move the unsigned version of this transform to InstSimplify, it
doesn't create any new instructions.
This fixes PR21243.
llvm-svn: 219567
consider:
mul i32 nsw %x, -2147483648
this instruction will not result in poison if %x is 1
however, if we transform this into:
shl i32 nsw %x, 31
then we will be generating poison because we just shifted into the sign
bit.
This fixes PR21242.
llvm-svn: 219566
The LLVM Lang Ref states for signed/unsigned int to float conversions:
"If the value cannot fit in the floating point value, the results are undefined."
And for FP to signed/unsigned int:
"If the value cannot fit in ty2, the results are undefined."
This matches the C definitions.
The existing behavior pins to infinity or a max int value, but that may just
lead to more confusion as seen in:
http://llvm.org/bugs/show_bug.cgi?id=21130
Returning undef will hopefully lead to a less silent failure.
Differential Revision: http://reviews.llvm.org/D5603
llvm-svn: 219542
It also makes it more aggressive in querying range information by
adding a call to isKnownPredicateWithRanges to
isLoopBackedgeGuardedByCond and isLoopEntryGuardedByCond.
phabricator: http://reviews.llvm.org/D5638
Reviewed by: atrick, hfinkel
llvm-svn: 219532
ScalarEvolution in the presence of multiple exits. Previously all
loops exits had to have identical counts for a loop trip count to be
considered computable. This pessimization was implemented by calling
getBackedgeTakenCount(L) rather than getExitCount(L, ExitingBlock)
inside of ScalarEvolution::getSmallConstantTripCount() (see the FIXME
in the comments of that function). The pessimization was added to fix
a corner case involving undefined behavior (pr/16130). This patch more
precisely handles the undefined behavior case allowing the pessimization
to be removed.
ControlsExit replaces IsSubExpr to more precisely track the case where
undefined behavior is expected to occur. Because undefined behavior is
tracked more precisely we can remove MustExit from ExitLimit. MustExit
was used to track the case where the limit was computed potentially
assuming undefined behavior even if undefined behavior didn't necessarily
occur.
llvm-svn: 219517
instead
We used to transform this:
define void @test6(i1 %cond, i8* %ptr) {
entry:
br i1 %cond, label %bb1, label %bb2
bb1:
br label %bb2
bb2:
%ptr.2 = phi i8* [ %ptr, %entry ], [ null, %bb1 ]
store i8 2, i8* %ptr.2, align 8
ret void
}
into this:
define void @test6(i1 %cond, i8* %ptr) {
%ptr.2 = select i1 %cond, i8* null, i8* %ptr
store i8 2, i8* %ptr.2, align 8
ret void
}
because the simplifycfg transformation into selects would happen to happen
before the simplifycfg transformation that removes unreachable control flow
(We have 'unreachable control flow' due to the store to null which is undefined
behavior).
The existing transformation that removes unreachable control flow in simplifycfg
is:
/// If BB has an incoming value that will always trigger undefined behavior
/// (eg. null pointer dereference), remove the branch leading here.
static bool removeUndefIntroducingPredecessor(BasicBlock *BB)
Now we generate:
define void @test6(i1 %cond, i8* %ptr) {
store i8 2, i8* %ptr.2, align 8
ret void
}
I did not see any impact on the test-suite + externals.
rdar://18596215
llvm-svn: 219462
This patch fixes a bug in method InstCombiner::FoldCmpCstShrCst where we
wrongly computed the distance between the highest bits set of two negative
values.
This fixes PR21222.
Differential Revision: http://reviews.llvm.org/D5700
llvm-svn: 219406
A function with discardable linkage cannot be discarded if its a member
of a COMDAT group without considering all the other COMDAT members as
well. This sort of thing is already handled by GlobalOpt/GlobalDCE.
This fixes PR21206.
llvm-svn: 219335
The icmp-select-icmp optimization targets select-icmp.eq
only. This is now ensured by testing the branch predicate
explictly. This commit also includes the test case for pr21199.
llvm-svn: 219282
`LoopUnrollPass` says that it preserves `LoopInfo` -- make it so. In
particular, tell `LoopInfo` about copies of inner loops when unrolling
the outer loop.
Conservatively, also tell `ScalarEvolution` to forget about the original
versions of these loops, since their inputs may have changed.
Fixes PR20987.
llvm-svn: 219241
This optimization tries to convert switch instructions that are used to select a value with only 2 unique cases + default block
to a select or a couple of selects (depending if the default block is reachable or not).
The typical case this optimization wants to be able to optimize is this one:
Example:
switch (a) {
case 10: %0 = icmp eq i32 %a, 10
return 10; %1 = select i1 %0, i32 10, i32 4
case 20: ----> %2 = icmp eq i32 %a, 20
return 2; %3 = select i1 %2, i32 2, i32 %1
default:
return 4;
}
It also sets the base for further optimizations that are planned and being reviewed.
llvm-svn: 219223
After some stellar (& inspired) help from Reid Kleckner providing a test
case for some rather unstable undefined behavior showing up as
assertions produced by r214761, I was able to fix this issue in DAE
involving the application of both varargs removal, followed by normal
argument removal.
Indeed I introduced this same bug into ArgumentPromotion (r212128) by
copying the code from DAE, and when I fixed the bug in ArgPromo
(r213805) and commented in that patch that I didn't need to address the
same issue in DAE because it was a single pass. Turns out it's two pass,
one for the varargs and one for the normal arguments, so the same fix is
needed (at least during varargs removal). So here it is.
(the observable/net effect of this bug, even when it didn't result in
assertion failure, is that debug info would describe the DAE'd function
in the abstract, but wouldn't provide high/low_pc, variable locations,
line table, etc (it would appear as though the function had been
entirely optimized away), see the original PR14016 for details of the
general problem)
I'm not recommitting the assertion just yet, as there's been another
regression of it since I last tried. It might just be a few test cases
weren't adequately updated after Adrian or Duncan's recent schema
changes.
llvm-svn: 219210
Takes care of the assert that caused build fails.
Rather than asserting the code checks now that the definition
and use are in the same block, and does not attempt
to optimize when that is not the case.
llvm-svn: 219175
Particularly, it addresses cases where Reassociate breaks Subtracts but then fails to optimize combinations like I1 + -I2 where I1 and I2 have the same rank and are identical.
Patch by Dmitri Shtilman.
llvm-svn: 219092
For any @llvm.assume intrinsic, if there is another which dominates it and uses
the same condition, then it is redundant and can be removed. While this does
not alter the semantics of the @llvm.assume intrinsics, it makes subsequent
handling more efficient (and the resulting IR easier to read).
llvm-svn: 219067
C++14 adds new builtin signatures for 'operator delete'. This change allows
new/delete pairs to be removed in C++14 onwards, as they were in C++11 and
before.
llvm-svn: 219014
This reverts commit r218918, effectively reapplying r218914 after fixing
an Ocaml bindings test and an Asan crash. The root cause of the latter
was a tightened-up check in `DILexicalBlock::Verify()`, so I'll file a
PR to investigate who requires the loose check (and why).
Original commit message follows.
--
This patch addresses the first stage of PR17891 by folding constant
arguments together into a single MDString. Integers are stringified and
a `\0` character is used as a separator.
Part of PR17891.
Note: I've attached my testcases upgrade scripts to the PR. If I've
just broken your out-of-tree testcases, they might help.
llvm-svn: 219010
This patch addresses the first stage of PR17891 by folding constant
arguments together into a single MDString. Integers are stringified and
a `\0` character is used as a separator.
Part of PR17891.
Note: I've attached my testcases upgrade scripts to the PR. If I've
just broken your out-of-tree testcases, they might help.
llvm-svn: 218914
When unsafe-fp-math is enabled, we can turn sqrt(X) * sqrt(X) into X.
This can happen in the real world when calculating x ** 3/2. This occurs
in test-suite/SingleSource/Benchmarks/BenchmarkGame/n-body.c.
Differential Revision: http://reviews.llvm.org/D5584
llvm-svn: 218906
My commit rL216160 introduced a bug PR21014: IndVars widens code 'for (i = ; i < ...; i++) arr[ CONST - i]' into 'for (i = ; i < ...; i++) arr[ i - CONST]'
thus inverting index expression. This patch fixes it.
Thanks to Jörg Sonnenberger for pointing.
Differential Revision: http://reviews.llvm.org/D5576
llvm-svn: 218867
As discussed here:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20140609/220598.html
And again here:
http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-September/077168.html
The sqrt of a negative number when using the llvm intrinsic is undefined.
We should return undef rather than 0.0 to match the definition in the LLVM IR lang ref.
This change should not affect any code that isn't using "no-nans-fp-math";
ie, no-nans is a requirement for generating the llvm intrinsic in place of a sqrt function call.
Unfortunately, the behavior introduced by this patch will not match current gcc, xlc, icc, and
possibly other compilers. The current clang/llvm behavior of returning 0.0 doesn't either.
We knowingly approve of this difference with the other compilers in an attempt to flag code
that is invoking undefined behavior.
A front-end warning should also try to convince the user that the program will fail:
http://llvm.org/bugs/show_bug.cgi?id=21093
Differential Revision: http://reviews.llvm.org/D5527
llvm-svn: 218803
argument of the llvm.dbg.declare/llvm.dbg.value intrinsics.
Previously, DIVariable was a variable-length field that has an optional
reference to a Metadata array consisting of a variable number of
complex address expressions. In the case of OpPiece expressions this is
wasting a lot of storage in IR, because when an aggregate type is, e.g.,
SROA'd into all of its n individual members, the IR will contain n copies
of the DIVariable, all alike, only differing in the complex address
reference at the end.
By making the complex address into an extra argument of the
dbg.value/dbg.declare intrinsics, all of the pieces can reference the
same variable and the complex address expressions can be uniqued across
the CU, too.
Down the road, this will allow us to move other flags, such as
"indirection" out of the DIVariable, too.
The new intrinsics look like this:
declare void @llvm.dbg.declare(metadata %storage, metadata %var, metadata %expr)
declare void @llvm.dbg.value(metadata %storage, i64 %offset, metadata %var, metadata %expr)
This patch adds a new LLVM-local tag to DIExpressions, so we can detect
and pretty-print DIExpression metadata nodes.
What this patch doesn't do:
This patch does not touch the "Indirect" field in DIVariable; but moving
that into the expression would be a natural next step.
http://reviews.llvm.org/D4919
rdar://problem/17994491
Thanks to dblaikie and dexonsmith for reviewing this patch!
Note: I accidentally committed a bogus older version of this patch previously.
llvm-svn: 218787
argument of the llvm.dbg.declare/llvm.dbg.value intrinsics.
Previously, DIVariable was a variable-length field that has an optional
reference to a Metadata array consisting of a variable number of
complex address expressions. In the case of OpPiece expressions this is
wasting a lot of storage in IR, because when an aggregate type is, e.g.,
SROA'd into all of its n individual members, the IR will contain n copies
of the DIVariable, all alike, only differing in the complex address
reference at the end.
By making the complex address into an extra argument of the
dbg.value/dbg.declare intrinsics, all of the pieces can reference the
same variable and the complex address expressions can be uniqued across
the CU, too.
Down the road, this will allow us to move other flags, such as
"indirection" out of the DIVariable, too.
The new intrinsics look like this:
declare void @llvm.dbg.declare(metadata %storage, metadata %var, metadata %expr)
declare void @llvm.dbg.value(metadata %storage, i64 %offset, metadata %var, metadata %expr)
This patch adds a new LLVM-local tag to DIExpressions, so we can detect
and pretty-print DIExpression metadata nodes.
What this patch doesn't do:
This patch does not touch the "Indirect" field in DIVariable; but moving
that into the expression would be a natural next step.
http://reviews.llvm.org/D4919
rdar://problem/17994491
Thanks to dblaikie and dexonsmith for reviewing this patch!
llvm-svn: 218778
In special cases select instructions can be eliminated by
replacing them with a cheaper bitwise operation even when the
select result is used outside its home block. The instances implemented
are patterns like
%x=icmp.eq
%y=select %x,%r, null
%z=icmp.eq|neq %y, null
br %z,true, false
==> %x=icmp.ne
%y=icmp.eq %r,null
%z=or %x,%y
br %z,true,false
The optimization is integrated into the instruction
combiner and performed only when all uses of the select result can
be replaced by the select operand proper. For this dominator information
is used and dominance is now a required analysis pass in the combiner.
The optimization itself is iterative. The critical step is to replace the
select result with the non-constant select operand. So the select becomes
local and the combiner iteratively works out simpler code pattern and
eventually eliminates the select.
rdar://17853760
llvm-svn: 218721
Summary:
This patch adds a threshold that controls the number of bonus instructions
allowed for folding branches with common destination. The original code allows
at most one bonus instruction. With this patch, users can customize the
threshold to allow multiple bonus instructions. The default threshold is still
1, so that the code behaves the same as before when users do not specify this
threshold.
The motivation of this change is that tuning this threshold significantly (up
to 25%) improves the performance of some CUDA programs in our internal code
base. In general, branch instructions are very expensive for GPU programs.
Therefore, it is sometimes worth trading more arithmetic computation for a more
straightened control flow. Here's a reduced example:
__global__ void foo(int a, int b, int c, int d, int e, int n,
const int *input, int *output) {
int sum = 0;
for (int i = 0; i < n; ++i)
sum += (((i ^ a) > b) && (((i | c ) ^ d) > e)) ? 0 : input[i];
*output = sum;
}
The select statement in the loop body translates to two branch instructions "if
((i ^ a) > b)" and "if (((i | c) ^ d) > e)" which share a common destination.
With the default threshold, SimplifyCFG is unable to fold them, because
computing the condition of the second branch "(i | c) ^ d > e" requires two
bonus instructions. With the threshold increased, SimplifyCFG can fold the two
branches so that the loop body contains only one branch, making the code
conceptually look like:
sum += (((i ^ a) > b) & (((i | c ) ^ d) > e)) ? 0 : input[i];
Increasing the threshold significantly improves the performance of this
particular example. In the configuration where both conditions are guaranteed
to be true, increasing the threshold from 1 to 2 improves the performance by
18.24%. Even in the configuration where the first condition is false and the
second condition is true, which favors shortcuts, increasing the threshold from
1 to 2 still improves the performance by 4.35%.
We are still looking for a good threshold and maybe a better cost model than
just counting the number of bonus instructions. However, according to the above
numbers, we think it is at least worth adding a threshold to enable more
experiments and tuning. Let me know what you think. Thanks!
Test Plan: Added one test case to check the threshold is in effect
Reviewers: nadav, eliben, meheff, resistor, hfinkel
Reviewed By: hfinkel
Subscribers: hfinkel, llvm-commits
Differential Revision: http://reviews.llvm.org/D5529
llvm-svn: 218711
This patch improves the target-specific cost model to better handle signed
division by a power of two. The immediate result is that this enables the SLP
vectorizer to do a better job.
http://reviews.llvm.org/D5469
PR20714
llvm-svn: 218607
Runtime unrolling will create a prologue to execute the extra
iterations which is can't divided by the unroll factor. It
generates an if-then-else sequence to jump into a factor -1
times unrolled loop body, like
extraiters = tripcount % loopfactor
if (extraiters == 0) jump Loop:
if (extraiters == loopfactor) jump L1
if (extraiters == loopfactor-1) jump L2
...
L1: LoopBody;
L2: LoopBody;
...
if tripcount < loopfactor jump End
Loop:
...
End:
It means if the unroll factor is 4, the loop body will be 7
times unrolled, 3 are in loop prologue, and 4 are in the loop.
This commit is to use a loop to execute the extra iterations
in prologue, like
extraiters = tripcount % loopfactor
if (extraiters == 0) jump Loop:
else jump Prol
Prol: LoopBody;
extraiters -= 1 // Omitted if unroll factor is 2.
if (extraiters != 0) jump Prol: // Omitted if unroll factor is 2.
if (tripcount < loopfactor) jump End
Loop:
...
End:
Then when unroll factor is 4, the loop body will be copied by
only 5 times, 1 in the prologue loop, 4 in the original loop.
And if the unroll factor is 2, new loop won't be created, just
as the original solution.
llvm-svn: 218604
The annotation instructions are dropped during codegen and have no
impact on size. In some cases, the annotations were preventing the
unroller from unrolling a loop because the annotation calls were
pushing the cost over the unrolling threshold.
Differential Revision: http://reviews.llvm.org/D5335
llvm-svn: 218525
The doFinalization method checks that the LoopToAliasSetMap is
empty. LICM populates that map as it runs through the loop nest,
deleting the entries for child loops as it goes. However, if a child
loop is deleted by another pass (e.g. unrolling) then the loop will
never be deleted from the map because LICM walks the loop nest to
find entries it can delete.
The fix is to delete the loop from the map and free the alias set
when the loop is deleted from the loop nest.
Differential Revision: http://reviews.llvm.org/D5305
llvm-svn: 218387
Rather than slurping in and splatting out the whole ctor list, preserve
the existing array entries without trying to understand them. Only
remove the entries that we know we can optimize away. This way we don't
need to wire through priority and comdats or anything else we might add.
Fixes a linker issue where the .init_array or .ctors entry would point
to discarded initialization code if the comdat group from the TU with
the faulty global_ctors entry was dropped.
llvm-svn: 218337
This improves other optimizations such as LSR. A sext may be added to the
compare's other operand, but this can often be hoisted outside of the loop.
llvm-svn: 217953
Example:
define i1 @foo(i32 %a) {
%shr = ashr i32 -9, %a
%cmp = icmp ne i32 %shr, -5
ret i1 %cmp
}
Before this fix, the instruction combiner wrongly thought that %shr
could have never been equal to -5. Therefore, %cmp was always folded to 'true'.
However, when %a is equal to 1, then %cmp evaluates to 'false'. Therefore,
in this example, it is not valid to fold %cmp to 'true'.
The problem was only affecting the case where the comparison was between
negative quantities where one of the quantities was obtained from arithmetic
shift of a negative constant.
This patch fixes the problem with the wrong folding (fixes PR20945).
With this patch, the 'icmp' from the example is now simplified to a
comparison between %a and 1. This still allows us to get rid of the arithmetic
shift (%shr).
llvm-svn: 217950
Some ICmpInsts when anded/ored with another ICmpInst trivially reduces
to true or false depending on whether or not all integers or no integers
satisfy the intersected/unioned range.
This sort of trivial looking code can come about when InstCombine
performs a range reduction-type operation on sdiv and the like.
This fixes PR20916.
llvm-svn: 217750
We used to crash processing any relevant @llvm.assume on a 32-bit target
(because we'd ask SE to subtract expressions of differing types). I've copied
our 'simple.ll' test, but with the data layout from arm-linux-gnueabihf to get
some meaningful test coverage here.
llvm-svn: 217574
The routine that determines an alignment given some SCEV returns zero if the
answer is unknown. In a case where we could determine the increment of an
AddRec but not the starting alignment, we would compute the integer modulus by
zero (which is illegal and traps). Prevent this by returning early if either
the start or increment alignment is unknown (zero).
llvm-svn: 217544
"Unroll" is not the appropriate name for this variable. Clang already uses
the term "interleave" in pragmas and metadata for this.
Differential Revision: http://reviews.llvm.org/D5066
llvm-svn: 217528
From a combination of @llvm.assume calls (and perhaps through other means, such
as range metadata), it is possible that all bits of a return value might be
known. Previously, InstCombine did not check for this (which is understandable
given assumptions of constant propagation), but means that we'd miss simple
cases where assumptions are involved.
llvm-svn: 217346
This change teaches LazyValueInfo to use the @llvm.assume intrinsic. Like with
the known-bits change (r217342), this requires feeding a "context" instruction
pointer through many functions. Aside from a little refactoring to reuse the
logic that turns predicates into constant ranges in LVI, the only new code is
that which can 'merge' the range from an assumption into that otherwise
computed. There is also a small addition to JumpThreading so that it can have
LVI use assumptions in the same block as the comparison feeding a conditional
branch.
With this patch, we can now simplify this as expected:
int foo(int a) {
__builtin_assume(a > 5);
if (a > 3) {
bar();
return 1;
}
return 0;
}
llvm-svn: 217345
This adds a ScalarEvolution-powered transformation that updates load, store and
memory intrinsic pointer alignments based on invariant((a+q) & b == 0)
expressions. Many of the simple cases we can get with ValueTracking, but we
still need something like this for the more complicated cases (such as those
with an offset) that require some algebra. Note that gcc's
__builtin_assume_aligned's optional third argument provides exactly for this
kind of 'misalignment' offset for which this kind of logic is necessary.
The primary motivation is to fixup alignments for vector loads/stores after
vectorization (and unrolling). This pass is added to the optimization pipeline
just after the SLP vectorizer runs (which, admittedly, does not preserve SE,
although I imagine it could). Regardless, I actually don't think that the
preservation matters too much in this case: SE computes lazily, and this pass
won't issue any SE queries unless there are any assume intrinsics, so there
should be no real additional cost in the common case (SLP does preserve DT and
LoopInfo).
llvm-svn: 217344
This builds on r217342, which added the infrastructure to compute known bits
using assumptions (@llvm.assume calls). That original commit added only a few
patterns (to catch common cases related to determining pointer alignment); this
change adds several other patterns for simple cases.
r217342 contained that, for assume(v & b = a), bits in the mask
that are known to be one, we can propagate known bits from the a to v. It also
had a known-bits transfer for assume(a = b). This patch adds:
assume(~(v & b) = a) : For those bits in the mask that are known to be one, we
can propagate inverted known bits from the a to v.
assume(v | b = a) : For those bits in b that are known to be zero, we can
propagate known bits from the a to v.
assume(~(v | b) = a): For those bits in b that are known to be zero, we can
propagate inverted known bits from the a to v.
assume(v ^ b = a) : For those bits in b that are known to be zero, we can
propagate known bits from the a to v. For those bits in
b that are known to be one, we can propagate inverted
known bits from the a to v.
assume(~(v ^ b) = a) : For those bits in b that are known to be zero, we can
propagate inverted known bits from the a to v. For those
bits in b that are known to be one, we can propagate
known bits from the a to v.
assume(v << c = a) : For those bits in a that are known, we can propagate them
to known bits in v shifted to the right by c.
assume(~(v << c) = a) : For those bits in a that are known, we can propagate
them inverted to known bits in v shifted to the right by c.
assume(v >> c = a) : For those bits in a that are known, we can propagate them
to known bits in v shifted to the right by c.
assume(~(v >> c) = a) : For those bits in a that are known, we can propagate
them inverted to known bits in v shifted to the right by c.
assume(v >=_s c) where c is non-negative: The sign bit of v is zero
assume(v >_s c) where c is at least -1: The sign bit of v is zero
assume(v <=_s c) where c is negative: The sign bit of v is one
assume(v <_s c) where c is non-positive: The sign bit of v is one
assume(v <=_u c): Transfer the known high zero bits
assume(v <_u c): Transfer the known high zero bits (if c is know to be a power
of 2, transfer one more)
A small addition to InstCombine was necessary for some of the test cases. The
problem is that when InstCombine was simplifying and, or, etc. it would fail to
check the 'do I know all of the bits' condition before checking less specific
conditions and would not fully constant-fold the result. I'm not sure how to
trigger this aside from using assumptions, so I've just included the change
here.
llvm-svn: 217343
This change, which allows @llvm.assume to be used from within computeKnownBits
(and other associated functions in ValueTracking), adds some (optional)
parameters to computeKnownBits and friends. These functions now (optionally)
take a "context" instruction pointer, an AssumptionTracker pointer, and also a
DomTree pointer, and most of the changes are just to pass this new information
when it is easily available from InstSimplify, InstCombine, etc.
As explained below, the significant conceptual change is that known properties
of a value might depend on the control-flow location of the use (because we
care that the @llvm.assume dominates the use because assumptions have
control-flow dependencies). This means that, when we ask if bits are known in a
value, we might get different answers for different uses.
The significant changes are all in ValueTracking. Two main changes: First, as
with the rest of the code, new parameters need to be passed around. To make
this easier, I grouped them into a structure, and I made internal static
versions of the relevant functions that take this structure as a parameter. The
new code does as you might expect, it looks for @llvm.assume calls that make
use of the value we're trying to learn something about (often indirectly),
attempts to pattern match that expression, and uses the result if successful.
By making use of the AssumptionTracker, the process of finding @llvm.assume
calls is not expensive.
Part of the structure being passed around inside ValueTracking is a set of
already-considered @llvm.assume calls. This is to prevent a query using, for
example, the assume(a == b), to recurse on itself. The context and DT params
are used to find applicable assumptions. An assumption needs to dominate the
context instruction, or come after it deterministically. In this latter case we
only handle the specific case where both the assumption and the context
instruction are in the same block, and we need to exclude assumptions from
being used to simplify their own ephemeral values (those which contribute only
to the assumption) because otherwise the assumption would prove its feeding
comparison trivial and would be removed.
This commit adds the plumbing and the logic for a simple masked-bit propagation
(just enough to write a regression test). Future commits add more patterns
(and, correspondingly, more regression tests).
llvm-svn: 217342
This adds a set of utility functions for collecting 'ephemeral' values. These
are LLVM IR values that are used only by @llvm.assume intrinsics (directly or
indirectly), and thus will be removed prior to code generation, implying that
they should be considered free for certain purposes (like inlining). The
inliner's cost analysis, and a few other passes, have been updated to account
for ephemeral values using the provided functionality.
This functionality is important for the usability of @llvm.assume, because it
limits the "non-local" side-effects of adding llvm.assume on inlining, loop
unrolling, etc. (these are hints, and do not generate code, so they should not
directly contribute to estimates of execution cost).
llvm-svn: 217335
The special case did not work when run under -reassociate and can easily
be expressed by a further generalization of an existing pattern.
llvm-svn: 217227
LinearFunctionTestReplace tries to use the *next* indvar to compare
against when possible. However, it may be the case that the calculation
for the next indvar has NUW/NSW flags and that it may only be safely
used inside the loop. Using it in a comparison to calculate the exit
condition could result in observing poison.
This fixes PR20680.
Differential Revision: http://reviews.llvm.org/D5174
llvm-svn: 217102
Fixes two latent bugs:
- There was no fence inserted before expanded seq_cst load (unsound on Power)
- There was only a fence release before seq_cst stores (again unsound, in particular on Power)
It is not even clear if this is correct on ARM swift processors (where release fences are
DMB ishst instead of DMB ish). This behaviour is currently preserved on ARM Swift
as it is not clear whether it is incorrect. I would love to get documentation stating
whether it is correct or not.
These two bugs were not triggered because Power is not (yet) using this pass, and these
behaviours happen to be (mostly?) working on ARM
(although they completely butchered the semantics of the llvm IR).
See:
http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-August/075821.html
for an example of the problems that can be caused by the second of these bugs.
I couldn't see a way of fixing these in a completely target-independent way without
adding lots of unnecessary fences on ARM, hence the target-dependent parts of this
patch.
This patch implements the new target-dependent parts only for ARM (the default
of not doing anything is enough for AArch64), other architectures will use this
infrastructure in later patches.
llvm-svn: 217076
The SLP vectorizer should propagate IR-level optimization hints/flags (nsw, nuw, exact, fast-math)
when converting scalar instructions into vectors. But this isn't a simple copy - we need to take
the intersection (the logical 'and') of the sets of flags on the scalars.
The solution is further complicated because we can have non-uniform (non-SIMD) vector ops after:
http://reviews.llvm.org/D4015http://llvm.org/viewvc/llvm-project?view=revision&revision=211339
The vast majority of changed files are existing tests that were not propagating IR flags, but I've
also added a new test file for focused testing of IR flag possibilities.
Differential Revision: http://reviews.llvm.org/D5172
llvm-svn: 217051
When folding a fused multiply-add builtin call, make sure that we propagate the
correct result in the case where the addend is zero, and the two other operands
are finite non-zero.
Example:
define double @test() {
%1 = call double @llvm.fma.f64(double 7.0, double 8.0, double 0.0)
ret double %1
}
Before this patch, the instruction simplifier wrongly folded the builtin call
in function @test to constant 'double 7.0'.
With this patch, method 'fusedMultiplyAdd' correctly evaluates the multiply and
propagates the expected result (i.e. 56.0).
Added test fold-builtin-fma.ll with the reproducible from PR20832 plus extra
test cases to verify the behavior of method 'fusedMultiplyAdd' in the presence
of NaN/Inf operands.
This fixes PR20832.
Differential Revision: http://reviews.llvm.org/D5152
llvm-svn: 216913
Summary:
BBs might contain non-LCSSA'd values after the LCSSA pass is run if they
are unreachable from the entry block.
Normally, the users of the instruction would be PHIs but the unreachable
BBs have normal users; rewrite their uses to be undef values.
An alternative fix could involve fixing this at LCSSA but that would
require this invariant to hold after subsequent transforms. If a BB
created an unreachable block, they would be in violation of this.
This fixes PR19798.
Differential Revision: http://reviews.llvm.org/D5146
llvm-svn: 216911
SROA may decide that it needs to insert a bitcast and would set it's
insertion point before a PHI. This will create an invalid module
right quick.
Instead, choose the first insertion point in the basic block that holds
our PHI.
This fixes PR20822.
Differential Revision: http://reviews.llvm.org/D5141
llvm-svn: 216891
This reverts commit r216698 which reverted r216523 and r216598.
We would attempt to perform the transformation even if the match()
failed because, as a side effect, it would set V. This would trick us
into believing that we correctly found a place to correctly apply the
transform.
An additional test case was added to getelementptr.ll so that we might
not regress in the future.
llvm-svn: 216890
The loop vectorizer preserves wrapping, exact, and fast-math properties of scalar instructions.
This patch adds a convenience method to make that operation easier because we need to do this
in the loop vectorizer, SLP vectorizer, and possibly other places.
Although this is a 'no functional change' patch, I've added a testcase to verify that the exact
flag is preserved by the loop vectorizer. The wrapping and fast-math flags are already checked
in existing testcases.
Differential Revision: http://reviews.llvm.org/D5138
llvm-svn: 216886
chain became completely broken here as *all* intrinsic users ended up
being skipped, and the ones that seemed to be singled out were actually
the exact wrong set.
This is a great example of why long else-if chains can be easily
confusing. Switch the entire code to use early exits and early continues
to have simpler (and more importantly, correct) logic here, as well as
fixing the reversed logic for detecting and continuing on lifetime
intrinsics.
I've also significantly cleaned up the test case and added another test
case demonstrating an example where the optimization is not (trivially)
safe to perform.
llvm-svn: 216871
Previously, the hint mechanism relied on clean up passes to remove redundant
metadata, which still showed up if running opt at low levels of optimization.
That also has shown that multiple nodes of the same type, but with different
values could still coexist, even if temporary, and cause confusion if the
next pass got the wrong value.
This patch makes sure that, if metadata already exists in a loop, the hint
mechanism will never append a new node, but always replace the existing one.
It also enhances the algorithm to cope with more metadata types in the future
by just adding a new type, not a lot of code.
Re-applying again due to MSVC 2013 being minimum requirement, and this patch
having C++11 that MSVC 2012 didn't support.
Fixes PR20655.
llvm-svn: 216870
This feeds AA through the IFI structure into the inliner so that
AddAliasScopeMetadata can use AA->getModRefBehavior to figure out which
functions only access their arguments (instead of just hard-coding some
knowledge of memory intrinsics). Most of the information is only available from
BasicAA; this is important for preserving alias scoping information for
target-specific intrinsics when doing the noalias parameter attribute to
metadata conversion.
llvm-svn: 216866
I thought that I had fixed this problem in r216818, but I did not do a very
good job. The underlying issue is that when we add alias.scope metadata we are
asserting that this metadata completely describes the aliasing relationships
within the current aliasing scope domain, and so in the context of translating
noalias argument attributes, the pointers must all be based on noalias
arguments (as underlying objects) and have no other kind of underlying object.
In r216818 excluding appropriate accesses from getting alias.scope metadata is
done by looking for underlying objects that are not identified function-local
objects -- but that's wrong because allocas, etc. are also function-local
objects and we need to explicitly check that all underlying objects are the
noalias arguments for which we're adding metadata aliasing scopes.
This fixes the underlying-object check for adding alias.scope metadata, and
does some refactoring of the related capture-checking eligibility logic (and
adds more comments; hopefully making everything a bit clearer).
Fixes self-hosting on x86_64 with -mllvm -enable-noalias-to-md-conversion (the
feature is still disabled by default).
llvm-svn: 216863
The previous implementation of AddAliasScopeMetadata, which adds noalias
metadata to preserve noalias parameter attribute information when inlining had
a flaw: it would add alias.scope metadata to accesses which might have been
derived from pointers other than noalias function parameters. This was
incorrect because even some access known not to alias with all noalias function
parameters could easily alias with an access derived from some other pointer.
Instead, when deriving from some unknown pointer, we cannot add alias.scope
metadata at all. This fixes a miscompile of the test-suite's tramp3d-v4.
Furthermore, we cannot add alias.scope to functions unless we know they
access only argument-derived pointers (currently, we know this only for
memory intrinsics).
Also, we fix a theoretical problem with using the NoCapture attribute to skip
the capture check. This is incorrect (as explained in the comment added), but
would not matter in any code generated by Clang because we get only inferred
nocapture attributes in Clang-generated IR.
This functionality is not yet enabled by default.
llvm-svn: 216818
consider: (and (icmp X, Y), (and Z, (icmp A, B)))
It may be possible to combine (icmp X, Y) with (icmp A, B).
If we successfully combine, create an 'and' instruction with Z.
This fixes PR20814.
N.B. There is room for improvement after this change but I'm not
convinced it's worth chasing yet.
llvm-svn: 216814
Even loads/stores that have a stronger ordering than monotonic can be safe.
The rule is no release-acquire pair on the path from the QueryInst, assuming that
the QueryInst is not atomic itself.
llvm-svn: 216771
Don't promote byval pointer arguments when when their size in bits is
not equal to their alloc size in bits. This can happen for x86_fp80,
where the size in bits is 80 but the alloca size in bits in 128.
Promoting these types can break passing unions of x86_fp80s and other
types.
Patch by Thomas Jablin!
Reviewed By: rnk
Differential Revision: http://reviews.llvm.org/D5057
llvm-svn: 216693
For a detailed description of the problem see the comment in the test file.
The problematic moveBefore() calls are not required anymore because the new
scheduling algorithm ensures a correct ordering anyway.
llvm-svn: 216656
Several combines involving icmp (shl C2, %X) C1 can be simplified
without introducing any new instructions. Move them to InstSimplify;
while we are at it, make them more powerful.
llvm-svn: 216642
We try to perform this transform in InstSimplify but we aren't always
able to. Sometimes, we need to insert a bitcast if X and Y don't have
the same time.
llvm-svn: 216598
It's incorrect to perform this simplification if the types differ.
A bitcast would need to be inserted for this to work.
This fixes PR20771.
llvm-svn: 216597
'shl nuw CI, x' produces [CI, CI << CLZ(CI)]
'shl nsw CI, x' produces [CI << CLO(CI)-1, CI] if CI is negative
'shl nsw CI, x' produces [CI, CI << CLZ(CI)-1] if CI is non-negative
llvm-svn: 216570
We supported transforming:
(gep i8* X, -(ptrtoint Y))
to:
(inttoptr (sub (ptrtoint X), (ptrtoint Y)))
However, this only fired if 'X' had type i8*. Generalize this to
support various types of different sizes. This results in much better
CodeGen, especially for pointers to packed structs.
llvm-svn: 216523
consider:
long long *f(long long *b, long long *e) {
return b + (e - b);
}
we would lower this to something like:
define i64* @f(i64* %b, i64* %e) {
%1 = ptrtoint i64* %e to i64
%2 = ptrtoint i64* %b to i64
%3 = sub i64 %1, %2
%4 = ashr exact i64 %3, 3
%5 = getelementptr inbounds i64* %b, i64 %4
ret i64* %5
}
This should fold away to just 'e'.
N.B. This adds m_SpecificInt as a convenient way to match against a
particular 64-bit integer when using LLVM's match interface.
llvm-svn: 216439
Summary:
There is no functionality change here except in the way we assemble and
dump musttail calls in variadic functions. There's really no need to
separate out the bits for musttail and "is forwarding varargs" on call
instructions. A musttail call by definition has to forward the ellipsis
or it would fail verification.
Reviewers: chandlerc, nlewycky
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D4892
llvm-svn: 216423
Adding, removing, or changing non-pack parameters can change the ABI
classification of pack parameters. Clang and other frontends encode the
classification in the IR of the call site, but the callee side
determines it dynamically based on the number of registers consumed so
far. Changing the prototype affects the number of registers consumed
would break such code.
Dead argument elimination performs a similar task and already has a
similar check to avoid this problem.
Patch by Thomas Jablin!
llvm-svn: 216421
GlobalDCE deletes global vars and updates their initializers to nullptr
while leaving underlying constants to be cleaned up later by its uses.
The clean up may never happen, fix this by forcing it every time it's
safe to destroy constants.
Final patch by Rafael Espindola
http://reviews.llvm.org/D4931
<rdar://problem/17523868>
llvm-svn: 216390
This patch adds support to recognize division by uniform power of 2 and modifies the cost table to vectorize division by uniform power of 2 whenever possible.
Updates Cost model for Loop and SLP Vectorizer.The cost table is currently only updated for X86 backend.
Thanks to Hal, Andrea, Sanjay for the review. (http://reviews.llvm.org/D4971)
llvm-svn: 216371
CFE, with -03, would turn:
bool f(unsigned x) {
bool a = x & 1;
bool b = x & 2;
return a | b;
}
into:
%1 = lshr i32 %x, 1
%2 = or i32 %1, %x
%3 = and i32 %2, 1
%4 = icmp ne i32 %3, 0
This sort of thing exposes a nasty pathology in GCC, ICC and LLVM.
Instead, we would rather want:
%1 = and i32 %x, 3
%2 = icmp ne i32 %1, 0
Things get a bit more interesting in the following case:
%1 = lshr i32 %x, %y
%2 = or i32 %1, %x
%3 = and i32 %2, 1
%4 = icmp ne i32 %3, 0
Replacing it with the following sequence is better:
%1 = shl nuw i32 1, %y
%2 = or i32 %1, 1
%3 = and i32 %2, %x
%4 = icmp ne i32 %3, 0
This sequence is preferable because %1 doesn't involve %x and could
potentially be hoisted out of loops if it is invariant; only perform
this transform in the non-constant case if we know we won't increase
register pressure.
llvm-svn: 216343
Summary:
Fixes PR20425.
During slice building, if all of the incoming values of a PHI node are the same, replace the PHI node with the common value. This simplification makes alloca's used by PHI nodes easier to promote.
Test Plan: Added three more tests in phi-and-select.ll
Reviewers: nlewycky, eliben, meheff, chandlerc
Reviewed By: chandlerc
Subscribers: zinovy.nis, hfinkel, baldrick, llvm-commits
Differential Revision: http://reviews.llvm.org/D4659
llvm-svn: 216299
Consider:
%add = add nuw i32 %a, -16777216
%and = and i32 %add, 255
Regardless of whether or not we demand the sign bit of %add, we cannot
replace -16777216 with 2130706432 without also removing 'nuw' from the
instruction.
llvm-svn: 216273
Consider:
%add = add nsw i32 %a, -16777216
%and = and i32 %add, 255
Regardless of whether or not we demand the sign bit of %add, we cannot
replace -16777216 with 2130706432 without also removing 'nsw' from the
instruction.
This fixes PR20377.
llvm-svn: 216261
In unreachable blocks it's legal to have instructions like "%x = op %x".
Such instuctions are not schedulable. Therefore the SLPVectorizer has to check for
unreachable blocks and ignore them.
Fixes bug 20646.
llvm-svn: 216256
Given something like X01XX + X01XX, we know that the result must look
like X1XXX.
Adapted from a patch by Richard Smith, test-case written by me.
llvm-svn: 216250
In this case, we are creating an x86_fp80 slice for a union from C where
the padding bytes may contain real data. An x86_fp80 alloca is 16 bytes,
and that's just fine. We can't, however, use regular loads and stores to
access the slice, because the store size is only 10 bytes / 80 bits.
Instead, use memcpy and memset.
Fixes PR18726.
Reviewed By: chandlerc
Differential Revision: http://reviews.llvm.org/D5012
llvm-svn: 216248
Somewhat unnoticed in the original implementation of discriminators, but
it could cause instructions to end up in new, small,
DW_TAG_lexical_blocks due to the use of DILexicalBlock to track
discriminator changes.
Instead, use DILexicalBlockFile which we already use to track file
changes without introducing new scopes, so it works well to track
discriminator changes in the same way.
llvm-svn: 216239
AtomicExpandLoadLinked is currently rather ARM-specific. This patch is the first of
a group that aim at making it more target-independent. See
http://lists.cs.uiuc.edu/pipermail/llvmdev/2014-August/075873.html
for details
The command line option is "atomic-expand"
llvm-svn: 216231
This does not require -ffast-math, and it gives CSE/GVN more options to
eliminate duplicate expressions in, e.g.:
return ((x + 0.1234 * y) * (x - 0.1234 * y));
Differential Revision: http://reviews.llvm.org/D4904
llvm-svn: 216169
Currently only "add nsw" are widened. This patch eliminates tons of "sext" instructions for 64 bit code (and the corresponding target code) in cases like:
int N = 100;
float **A;
void foo(int x0, int x1)
{
float * A_cur = &A[0][0];
float * A_next = &A[1][0];
for(int x = x0; x < x1; ++x).
{
// Currently only [x+N] case is widened. Others 2 cases lead to sext.
// This patch fixes it, so all 3 cases do not need sext.
const float div = A_cur[x + N] + A_cur[x - N] + A_cur[x * N];
A_next[x] = div;
}
}
...
> clang++ test.cpp -march=core-avx2 -Ofast -fno-unroll-loops -fno-tree-vectorize -S -o -
Differential Revision: http://reviews.llvm.org/D4695
llvm-svn: 216160
We can prove that a 'sub' can be a 'sub nsw' under certain conditions:
- The sign bits of the operands is the same.
- Both operands have more than 1 sign bit.
The subtraction cannot be a signed overflow in either case.
llvm-svn: 216037
Previously, the hint mechanism relied on clean up passes to remove redundant
metadata, which still showed up if running opt at low levels of optimization.
That also has shown that multiple nodes of the same type, but with different
values could still coexist, even if temporary, and cause confusion if the
next pass got the wrong value.
This patch makes sure that, if metadata already exists in a loop, the hint
mechanism will never append a new node, but always replace the existing one.
It also enhances the algorithm to cope with more metadata types in the future
by just adding a new type, not a lot of code.
llvm-svn: 215994
- add check for volatile (probably unneeded, but I agree that we should be conservative about it).
- strengthen condition from isUnordered() to isSimple(), as I don't understand well enough Unordered semantics (and it also matches the comment better this way) to be confident in the previous behaviour (thanks for catching that one, I had missed the case Monotonic/Unordered).
- separate a condition in two.
- lengthen comment about aliasing and loads
- add tests in GVN/atomic.ll
llvm-svn: 215943
While this might seem like an obvious canonicalization, there is one subtle problem with it. The result of the original expression
is undef when x is NaN (remember, fast math flags), but the result of the select is always defined when x is NaN. This means that the
new expression is strictly more defined than the original one. One unfortunate consequence of this is that the transform is not reversible!
It's always legal to make increase the defined-ness of an expression, but it's not legal to reduce it. Thus, targets that prefer the original
form of the expression cannot reverse the transform to recover it. Another way to think of it is that the transform has lost source-level
information (the fast math flags), which is undesirable.
llvm-svn: 215825
We can combne a mul with a div if one of the operands is a multiple of
the other:
%mul = mul nsw nuw %a, C1
%ret = udiv %mul, C2
=>
%ret = mul nsw %a, (C1 / C2)
This can expose further optimization opportunities if we end up
multiplying or dividing by a power of 2.
Consider this small example:
define i32 @f(i32 %a) {
%mul = mul nuw i32 %a, 14
%div = udiv exact i32 %mul, 7
ret i32 %div
}
which gets CodeGen'd to:
imull $14, %edi, %eax
imulq $613566757, %rax, %rcx
shrq $32, %rcx
subl %ecx, %eax
shrl %eax
addl %ecx, %eax
shrl $2, %eax
retq
We can now transform this into:
define i32 @f(i32 %a) {
%shl = shl nuw i32 %a, 1
ret i32 %shl
}
which gets CodeGen'd to:
leal (%rdi,%rdi), %eax
retq
This fixes PR20681.
llvm-svn: 215815
When a call site with noalias metadata is inlined, that metadata can be
propagated directly to the inlined instructions (only those that might access
memory because it is not useful on the others). Prior to inlining, the noalias
metadata could express that a call would not alias with some other memory
access, which implies that no instruction within that called function would
alias. By propagating the metadata to the inlined instructions, we preserve
that knowledge.
This should complete the enhancements requested in PR20500.
llvm-svn: 215676
When preserving noalias function parameter attributes by adding noalias
metadata in the inliner, we should do this for general function calls (not just
memory intrinsics). The logic is very similar to what already existed (except
that we want to add this metadata even for functions taking no relevant
parameters). This metadata can be used by ModRef queries in the caller after
inlining.
This addresses the first part of PR20500. Adding noalias metadata during
inlining is still turned off by default.
llvm-svn: 215657
v2: continue iterating through the rest of the bb
use for loop
v3: initialize FlattenCFG pass in ScalarOps
add test
v4: split off initializing flattencfg to a separate patch
add comment
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 215574
attribute and function argument attribute synthesizing and propagating.
As with the other uses of this attribute, the goal remains a best-effort
(no guarantees) attempt to not optimize the function or assume things
about the function when optimizing. This is particularly useful for
compiler testing, bisecting miscompiles, triaging things, etc. I was
hitting specific issues using optnone to isolate test code from a test
driver for my fuzz testing, and this is one step of fixing that.
llvm-svn: 215538
Correctness proof of the transform using CVC3-
$ cat t.cvc
A, B : BITVECTOR(32);
QUERY BVXOR(A | B, BVXOR(A,B) ) = A & B;
$ cvc3 t.cvc
Valid.
llvm-svn: 215524
What follows bellow is a correctness proof of the transform using CVC3.
$ < t.cvc
A, B : BITVECTOR(32);
QUERY BVPLUS(32, A & B, A | B) = BVPLUS(32, A, B);
$ cvc3 < t.cvc
Valid.
llvm-svn: 215400
and the lattice will be updated to be a state other than "undefined". This
limiation could miss some opportunities of lowering "overdefined" to be an
even accurate value. So this patch ask the algorithm to try to lower the
lattice value again even if the value has been lowered to be "overdefined".
llvm-svn: 215343
GlobalOpt didn't know how to simulate InsertValueInst or
ExtractValueInst. Optimizing these is pretty straightforward.
N.B. This came up when looking at clang's IRGen for MS ABI member
pointers; they are represented as aggregates.
llvm-svn: 215184
this case, the code path dealing with vector promotion was missing the explicit
checks for lifetime intrinsics that were present on the corresponding integer
promotion path.
llvm-svn: 215148
Optimize the following IR:
%1 = tail call noalias i8* @calloc(i64 1, i64 4)
%2 = bitcast i8* %1 to i32*
; This store is dead and should be removed
store i32 0, i32* %2, align 4
Memory returned by calloc is guaranteed to be zero initialized. If the value being stored is the constant zero (and the store is not otherwise observable across threads), we can delete the store. If the store is to an out of bounds address, it is undefined and thus also removable.
Reviewed By: nicholas
Differential Revision: http://reviews.llvm.org/D3942
llvm-svn: 214897
Some types, such as 128-bit vector types on AArch64, don't have any callee-saved registers. So if a value needs to stay live over a callsite, it must be spilled and refilled. This cost is now taken into account.
llvm-svn: 214859
When the cost model determines vectorization is not possible/profitable these remarks print an analysis of that decision.
Note that in selectVectorizationFactor() we can assume that OptForSize and ForceVectorization are mutually exclusive.
Reviewed by Arnold Schwaighofer
llvm-svn: 214599
The current remark is ambiguous and makes it sounds like explicitly specifying vectorization will allow the loop to be vectorized. This is not the case. The improved remark directs the user to -Rpass-analysis=loop-vectorize to determine the cause of the pass-miss.
Reviewed by Arnold Schwaighofer`
llvm-svn: 214445
We can only propagate the nsw bits if both subtraction instructions are
marked with the appropriate bit.
N.B. We only propagate the nsw bit in InstCombine because the nuw case
is already handled in InstSimplify.
This fixes PR20189.
llvm-svn: 214385
If the NUW bit is set for 0 - Y, we know that all values for Y other
than 0 would produce a poison value. This allows us to replace (0 - Y)
with 0 in the expression (X - (0 - Y)) which will ultimately leave us
with X.
This partially fixes PR20189.
llvm-svn: 214384
Before this patch we had
@a = weak global ...
but
@b = alias weak ...
The patch changes aliases to look more like global variables.
Looking at some really old code suggests that the reason was that the old
bison based parser had a reduction for alias linkages and another one for
global variable linkages. Putting the alias first avoided the reduce/reduce
conflict.
The days of the old .ll parser are long gone. The new one parses just "linkage"
and a later check is responsible for deciding if a linkage is valid in a
given context.
llvm-svn: 214355
While we can already transform A | (A ^ B) into A | B, things get bad
once we have (A ^ B) | (A ^ B ^ Cst) because reassociation will morph
this into (A ^ B) | ((A ^ Cst) ^ B). Our existing patterns fail once
this happens.
To fix this, we add a new pattern which looks through the tree of xor
binary operators to see that, in fact, there exists a redundant xor
operation.
What follows bellow is a correctness proof of the transform using CVC3.
$ cat t.cvc
A, B, C : BITVECTOR(64);
QUERY BVXOR(A, B) | BVXOR(BVXOR(B, C), A) = BVXOR(A, B) | C;
QUERY BVXOR(BVXOR(A, C), B) | BVXOR(A, B) = BVXOR(A, B) | C;
QUERY BVXOR(A, B) & BVXOR(BVXOR(B, C), A) = BVXOR(A, B) & ~C;
QUERY BVXOR(BVXOR(A, C), B) & BVXOR(A, B) = BVXOR(A, B) & ~C;
$ cvc3 < t.cvc
Valid.
Valid.
Valid.
Valid.
llvm-svn: 214342
The lifetime intrinsics need some work in order to make it clear which
optimizations are or are not valid.
For now dropping this optimization avoids a miscompilation.
Patch by Björn Steinbrink.
llvm-svn: 214336
The test being performed is just an approximation anyway, so it really
shouldn't crash when things don't go entirely as expected.
Should fix PR20474.
llvm-svn: 214177
This is the first commit in a series that add an @llvm.assume intrinsic which
can be used to provide the optimizer with a condition it may assume to be true
(when the control flow would hit the intrinsic call). Some basic properties are added here:
- llvm.invariant(true) is dead.
- llvm.invariant(false) is unreachable (this directly corresponds to the
documented behavior of MSVC's __assume(0)), so is llvm.invariant(undef).
The intrinsic is tagged as writing arbitrarily, in order to maintain control
dependencies. BasicAA has been updated, however, to return NoModRef for any
particular location-based query so that we don't unnecessarily block code
motion.
llvm-svn: 213973
This functionality is currently turned off by default.
Part of the motivation for introducing scoped-noalias metadata is to enable the
preservation of noalias parameter attribute information after inlining.
Sometimes this can be inferred from the code in the caller after inlining, but
often we simply lose valuable information.
The overall process if fairly simple:
1. Create a new unqiue scope domain.
2. For each (used) noalias parameter, create a new alias scope.
3. For each pointer, collect the underlying objects. Add a noalias scope for
each noalias parameter from which we're not derived (and has not been
captured prior to that point).
4. Add an alias.scope for each noalias parameter from which we might be
derived (or has been captured before that point).
Note that the capture checks apply only if one of the underlying objects is not
an identified function-local object.
llvm-svn: 213949
hint) the loop unroller replaces the llvm.loop.unroll.count metadata with
llvm.loop.unroll.disable metadata to prevent any subsequent unrolling
passes from unrolling more than the hint indicates. This patch fixes
an issue where loop unrolling could be disabled for other loops as well which
share the same llvm.loop metadata.
llvm-svn: 213900
This commit adds scoped noalias metadata. The primary motivations for this
feature are:
1. To preserve noalias function attribute information when inlining
2. To provide the ability to model block-scope C99 restrict pointers
Neither of these two abilities are added here, only the necessary
infrastructure. In fact, there should be no change to existing functionality,
only the addition of new features. The logic that converts noalias function
parameters into this metadata during inlining will come in a follow-up commit.
What is added here is the ability to generally specify noalias memory-access
sets. Regarding the metadata, alias-analysis scopes are defined similar to TBAA
nodes:
!scope0 = metadata !{ metadata !"scope of foo()" }
!scope1 = metadata !{ metadata !"scope 1", metadata !scope0 }
!scope2 = metadata !{ metadata !"scope 2", metadata !scope0 }
!scope3 = metadata !{ metadata !"scope 2.1", metadata !scope2 }
!scope4 = metadata !{ metadata !"scope 2.2", metadata !scope2 }
Loads and stores can be tagged with an alias-analysis scope, and also, with a
noalias tag for a specific scope:
... = load %ptr1, !alias.scope !{ !scope1 }
... = load %ptr2, !alias.scope !{ !scope1, !scope2 }, !noalias !{ !scope1 }
When evaluating an aliasing query, if one of the instructions is associated
with an alias.scope id that is identical to the noalias scope associated with
the other instruction, or is a descendant (in the scope hierarchy) of the
noalias scope associated with the other instruction, then the two memory
accesses are assumed not to alias.
Note that is the first element of the scope metadata is a string, then it can
be combined accross functions and translation units. The string can be replaced
by a self-reference to create globally unqiue scope identifiers.
[Note: This overview is slightly stylized, since the metadata nodes really need
to just be numbers (!0 instead of !scope0), and the scope lists are also global
unnamed metadata.]
Existing noalias metadata in a callee is "cloned" for use by the inlined code.
This is necessary because the aliasing scopes are unique to each call site
(because of possible control dependencies on the aliasing properties). For
example, consider a function: foo(noalias a, noalias b) { *a = *b; } that gets
inlined into bar() { ... if (...) foo(a1, b1); ... if (...) foo(a2, b2); } --
now just because we know that a1 does not alias with b1 at the first call site,
and a2 does not alias with b2 at the second call site, we cannot let inlining
these functons have the metadata imply that a1 does not alias with b2.
llvm-svn: 213864
We use gep to access the global array "switch.table", and the table index
should be treated as unsigned. When the highest bit is 1, this commit
zero-extends the index to an integer type with larger size.
For a switch on i2, we used to generate:
%switch.tableidx = sub i2 %0, -2
getelementptr inbounds [4 x i64]* @switch.table, i32 0, i2 %switch.tableidx
It is incorrect when %switch.tableidx is 2 or 3. The fix is to generate
%switch.tableidx = sub i2 %0, -2
%switch.tableidx.zext = zext i2 %switch.tableidx to i3
getelementptr inbounds [4 x i64]* @switch.table, i32 0, i3 %switch.tableidx.zext
rdar://17735071
llvm-svn: 213815
While the subprogram map cache used by Dead Argument Elimination works
there, I made a mistake when reusing it for Argument Promotion in
r212128 because ArgPromo may transform functions more than once whereas
DAE transforms each function only once, removing all the dead arguments
in one go.
To address this, ensure that the map is updated after each argument
promotion.
In retrospect it might be a little wasteful to create a map of all
subprograms when only handling a single CGSCC, but the alternative is
walking the debug info for each function in the CGSCC that gets updated.
It's not clear to me what the right tradeoff is there, but since the
current tradeoff seems to be working OK (and the code to keep things
updated is very cheap), let's stick with that for now.
llvm-svn: 213805
Also the debug location I had here was bogus, describing the location of
the call site as in the callee - and unnecessary, so just drop it.
llvm-svn: 213803
It handles the errors which were seen in PR19958 where wrong code was being emitted due to earlier patch.
Added code for lshr as well as non-exact right shifts.
It implements :
(icmp eq/ne (ashr/lshr const2, A), const1)" ->
(icmp eq/ne A, Log2(const2/const1)) ->
(icmp eq/ne A, Log2(const2) - Log2(const1))
Differential Revision: http://reviews.llvm.org/D4068
llvm-svn: 213678
"((~A & B) | A) -> (A | B)" and "((A & B) | ~A) -> (~A | B)"
Original Patch credit to Ankit Jain !!
Differential Revision: http://reviews.llvm.org/D4591
llvm-svn: 213676
We previously supported the align attribute on all (pointer) parameters, but we
only used it for byval parameters. However, it is completely consistent at the
IR level to treat 'align n' on all pointer parameters as an alignment
assumption on the pointer, and now we wll. Specifically, this causes
computeKnownBits to use the align attribute on all pointer parameters, not just
byval parameters. I've also added an explicit parameter attribute test for this
to test/Bitcode/attributes.ll.
And I've updated the LangRef to document the align parameter attribute (as it
turns out, it was not documented at all previously, although the byval
documentation mentioned that it could be used).
There are (at least) two benefits to doing this:
- It allows enhancing alignment based on the pointer alignment after inlining callees.
- It allows simplification of pointer arithmetic.
llvm-svn: 213670
Prior to this change, the loop vectorizer did not make use of the alias
analysis infrastructure. Instead, it performed memory dependence analysis using
ScalarEvolution-based linear dependence checks within equivalence classes
derived from the results of ValueTracking's GetUnderlyingObjects.
Unfortunately, this meant that:
1. The loop vectorizer had logic that essentially duplicated that in BasicAA
for aliasing based on identified objects.
2. The loop vectorizer could not partition the space of dependency checks
based on information only easily available from within AA (TBAA metadata is
currently the prime example).
This means, for example, regardless of whether -fno-strict-aliasing was
provided, the vectorizer would only vectorize this loop with a runtime
memory-overlap check:
void foo(int *a, float *b) {
for (int i = 0; i < 1600; ++i)
a[i] = b[i];
}
This is suboptimal because the TBAA metadata already provides the information
necessary to show that this check unnecessary. Of course, the vectorizer has a
limit on the number of such checks it will insert, so in practice, ignoring
TBAA means not vectorizing more-complicated loops that we should.
This change causes the vectorizer to use an AliasSetTracker to keep track of
the pointers in the loop. The resulting alias sets are then used to partition
the space of dependency checks, and potential runtime checks; this results in
more-efficient vectorizations.
When pointer locations are added to the AliasSetTracker, two things are done:
1. The location size is set to UnknownSize (otherwise you'd not catch
inter-iteration dependencies)
2. For instructions in blocks that would need to be predicated, TBAA is
removed (because the metadata might have a control dependency on the condition
being speculated).
For non-predicated blocks, you can leave the TBAA metadata. This is safe
because you can't have an iteration dependency on the TBAA metadata (if you
did, and you unrolled sufficiently, you'd end up with the same pointer value
used by two accesses that TBAA says should not alias, and that would yield
undefined behavior).
llvm-svn: 213486
There are some kinds of metadata that are safe to propagate from the scalar
instructions to the vector instructions (fpmath and tbaa currently).
Regarding TBAA, one might worry about propagating it on if-converted loads and
stores, because the metadata might have had a control dependency on the
condition, and thus actually aliased with some other non-speculated memory
access when the condition was false. However, this would be caught by the
runtime overlap checks.
llvm-svn: 213452
When we have a parameter (or call site return) with a dereferenceable
attribute, it can specify the size of an array pointed to by that parameter. If
we have a value for which we can accumulate a constant offset to such a
parameter, then we can use that offset in a direct comparison with the size
specified by the dereferenceable attribute.
This enables us to handle cases like this:
int foo(int a[static 3]) {
return a[2]; /* this is always dereferenceable */
}
llvm-svn: 213447
Merges equivalent loads on both sides of a hammock/diamond
and hoists into into the header.
Merges equivalent stores on both sides of a hammock/diamond
and sinks it to the footer.
Can enable if conversion and tolerate better load misses
and store operand latencies.
llvm-svn: 213396
This attribute indicates that the parameter or return pointer is
dereferenceable. Practically speaking, loads from such a pointer within the
associated byte range are safe to speculatively execute. Such pointer
parameters are common in source languages (C++ references, for example).
llvm-svn: 213385
Refactor code, no functionality change, test case moved from instcombine to instsimplify.
Differential Revision: http://reviews.llvm.org/D4102
llvm-svn: 213231
This reverts, "r213024 - Revert r212572 "improve BasicAA CS-CS queries", it
causes PR20303." with a fix for the bug in pr20303. As it turned out, the
relevant code was both wrong and over-conservative (because, as with the code
it replaced, it would return the overall ModRef mask even if just Ref had been
implied by the argument aliasing results). Hopefully, this correctly fixes both
problems.
Thanks to Nick Lewycky for reducing the test case for pr20303 (which I've
cleaned up a little and added in DSE's test directory). The BasicAA test has
also been updated to check for this error.
Original commit message:
BasicAA contains knowledge of certain intrinsics, such as memcpy and memset,
and uses that information to form more-accurate answers to CallSite vs. Loc
ModRef queries. Unfortunately, it did not use this information when answering
CallSite vs. CallSite queries.
Generically, when an intrinsic takes one or more pointers and the intrinsic is
marked only to read/write from its arguments, the offset/size is unknown. As a
result, the generic code that answers CallSite vs. CallSite (and CallSite vs.
Loc) queries in AA uses UnknownSize when forming Locs from an intrinsic's
arguments. While BasicAA's CallSite vs. Loc override could use more-accurate
size information for some intrinsics, it did not do the same for CallSite vs.
CallSite queries.
This change refactors the intrinsic-specific logic in BasicAA into a generic AA
query function: getArgLocation, which is overridden by BasicAA to supply the
intrinsic-specific knowledge, and used by AA's generic implementation. This
allows the intrinsic-specific knowledge to be used by both CallSite vs. Loc and
CallSite vs. CallSite queries, and simplifies the BasicAA implementation.
Currently, only one function, Mac's memset_pattern16, is handled by BasicAA
(all the rest are intrinsics). As a side-effect of this refactoring, BasicAA's
getModRefBehavior override now also returns OnlyAccessesArgumentPointees for
this function (which is an improvement).
llvm-svn: 213219
Summary:
Converting outermost zext(a) to sext(a) causes worse code when the
computation of zext(a) could be reused. For example, after converting
... = array[zext(a)]
... = array[zext(a) + 1]
to
... = array[sext(a)]
... = array[zext(a) + 1],
the program computes sext(a), which is actually unnecessary. I added one
test in split-gep-and-gvn.ll to illustrate this scenario.
Also, with r211281 and r211084, we annotate more "nuw" tags to
computation involving CUDA intrinsics such as threadIdx.x. These
annotations help with splitting GEP a lot, rendering the benefit we get
from this reverted optimization only marginal.
Test Plan: make check-all
Reviewers: eliben, meheff
Reviewed By: meheff
Subscribers: jholewinski, llvm-commits
Differential Revision: http://reviews.llvm.org/D4542
llvm-svn: 213209
This patch modifies the existing DiagnosticInfo system to create a generic base
class that is inherited to produce diagnostic-based warnings. This is used by
the loop vectorizer to trigger a warning when vectorization is forced and
fails. Several tests have been added to verify this behavior.
Reviewed by: Arnold Schwaighofer
llvm-svn: 213110
Determining the bounds of x/ -1 would start off with us dividing it by
INT_MIN. Suffice to say, this would not work very well.
Instead, handle it upfront by checking for -1 and mapping it to the
range: [INT_MIN + 1, INT_MAX. This means that the result of our
division can be any value other than INT_MIN.
llvm-svn: 212981
Summary:
When calculating the upper bound of X / -8589934592, we would perform
the following calculation: Floor[INT_MAX / 8589934592]
However, flooring the result would make us wrongly come to the
conclusion that 1073741824 was not in the set of possible values.
Instead, use the ceiling of the result.
Reviewers: nicholas
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D4502
llvm-svn: 212976
Fix a crash in `InstCombiner::Descale()` when a multiply-by-zero gets
created as an argument to a GEP partway through an iteration, causing
-instcombine to optimize the GEP before the multiply.
rdar://problem/17615671
llvm-svn: 212742
isDereferenceablePointer should not give up upon encountering any bitcast. If
we're casting from a pointer to a larger type to a pointer to a small type, we
can continue by examining the bitcast's operand. This missing capability
was noted in a comment in the function.
In order for this to work, isDereferenceablePointer now takes an optional
DataLayout pointer (essentially all callers already had such a pointer
available). Most code uses isDereferenceablePointer though
isSafeToSpeculativelyExecute (which already took an optional DataLayout
pointer), and to enable the LICM test case, LICM needs to actually provide its DL
pointer to isSafeToSpeculativelyExecute (which it was not doing previously).
llvm-svn: 212686
This lets us experiment with 512-bit vectorization without passing
force-vector-width manually.
The code generated for a simple integer memset loop is properly vectorized.
Disassembly is still broken for it though :(.
llvm-svn: 212634
In PR20059 ( http://llvm.org/pr20059 ), instcombine eliminates shuffles that are necessary before performing an operation that can trap (srem).
This patch calls isSafeToSpeculativelyExecute() and bails out of the optimization in SimplifyVectorOp() if needed.
Differential Revision: http://reviews.llvm.org/D4424
llvm-svn: 212629
This reverts commit 5b55a47e94e28fbb56d0cd5d72c3db9105c15b4c.
A test case was found to crash after this was applied. I'll file a bug to track fixing this with the test case needed.
llvm-svn: 212550
This patch adds to an existing loop over phi nodes in SimplifyCondBranchToCondBranch() to check for trapping ops and bails out of the optimization if we find one of those.
The test cases verify that trapping ops are not hoisted and non-trapping ops are still optimized as expected.
llvm-svn: 212490
We've been performing the wrong operation on ARM for "atomicrmw nand" for
years, since "a NAND b" is "~(a & b)" rather than ARM's very tempting "a & ~b".
This bled over into the generic expansion pass.
So I assume no-one has ever actually tried to do an atomic nand in the real
world. Oh well.
llvm-svn: 212443
A GEP of a non-weak global variable will not be equivalent to another
non-weak global variable or a GEP of such a variable.
Differential Revision: http://reviews.llvm.org/D4238
llvm-svn: 212360
This is useful for functions that are not actually available externally but
referenced by a vtable of some kind. Clang emits functions like this for the MS
ABI.
PR20182.
llvm-svn: 212337
When INT_MIN is the numerator in a sdiv, we would not properly handle
overflow when calculating the bounds of possible values; abs(INT_MIN) is
not a meaningful number.
Instead, check and handle INT_MIN by reasoning that the largest value is
INT_MIN/-2 and the smallest value is INT_MIN.
This fixes PR20199.
llvm-svn: 212307
Matching behavior with DeadArgumentElimination (and leveraging some
now-common infrastructure), keep track of the function from debug info
metadata if arguments are promoted.
This may produce interesting debug info - since the arguments may be
missing or of different types... but at least backtraces, inlining, etc,
will be correct.
llvm-svn: 212128
There were transforms whose *intent* was to downgrade the linkage of
external objects to have internal linkage.
However, it fired on things with private linkage as well.
llvm-svn: 212104
Inlining functions with block addresses can cause many problem and requires a
rich infrastructure to support including escape analysis. At this point the
safest approach to address these problems is by blocking inlining from
happening.
Background:
There have been reports on Ruby segmentation faults triggered by inlining
functions with block addresses like
//Ruby code snippet
vm_exec_core() {
finish_insn_seq_0 = &&INSN_LABEL_finish;
INSN_LABEL_finish:
;
}
This kind of scenario can also happen when LLVM picks a subset of blocks for
inlining, which is the case with the actual code in the Ruby environment.
LLVM suppresses inlining for such functions when there is an indirect branch.
The attached patch does so even when there is no indirect branch. Note that
user code like above would not make much sense: using the global for jumping
across function boundaries would be illegal.
Why was there a segfault:
In the snipped above the block with the label is recognized as dead So it is
eliminated. Instead of a block address the cloner stores a constant (sic!) into
the global resulting in the segfault (when the global is used in a goto).
Why had it worked in the past then:
By luck. In older versions vm_exec_core was also inlined but the label address
used was the block label address in vm_exec_core. So the global jump ended up
in the original function rather than in the caller which accidentally happened
to work.
Test case ./tools/clang/test/CodeGen/indirect-goto.c will fail as a result
of this commit.
rdar://17245966
llvm-svn: 212077
This both improves basic debug info quality, but also fixes a larger
hole whenever we inline a call/invoke without a location (debug info for
the entire inlining is lost and other badness that the debug info
emission code is currently working around but shouldn't have to).
llvm-svn: 212065
This patch enables transforms for
(x + (~(y | c) + 1) --> x - (y | c) if c is odd
Differential Revision: http://reviews.llvm.org/D4210
llvm-svn: 211881
If both instructions to be replaced are marked invariant the resulting
instruction is invariant.
rdar://13358910
Fix by Erik Eckstein!
llvm-svn: 211801
This patch enables transforms for
(x + (~(y | c) + 1) --> x - (y | c) if c is even
Differential Revision: http://reviews.llvm.org/D4209
llvm-svn: 211765
Folding a reference to a thread_local variable into another global
variable's initializer is very problematic, there is no relocation that
exists to represent such an access.
llvm-svn: 211762
[LLVM part]
These patches rename the loop unrolling and loop vectorizer metadata
such that they have a common 'llvm.loop.' prefix. Metadata name
changes:
llvm.vectorizer.* => llvm.loop.vectorizer.*
llvm.loopunroll.* => llvm.loop.unroll.*
This was a suggestion from an earlier review
(http://reviews.llvm.org/D4090) which added the loop unrolling
metadata.
Patch by Mark Heffernan.
llvm-svn: 211710
Fixes exponential compilation complexity in PR19835, caused by
LICM::sink not handling the following pattern well:
f = op g
e = op f, g
d = op e
c = op d, e
b = op c
a = op b, c
When an instruction with N uses is sunk, each of its operands gets N
new uses (all of them - phi nodes). In the example above, if a had 1
use, c would have 2, e would have 4, and g would have 8.
llvm-svn: 211673
Summary:
This new debug emission kind supports emitting line location
information in all instructions, but stops code generation
from emitting debug info to the final output.
This mode is useful when the backend wants to track source
locations during code generation, but it does not want to
produce debug info. This is currently used by optimization
remarks (-pass-remarks, -pass-remarks-missed and
-pass-remarks-analysis).
To prevent debug info emission, DIBuilder never inserts the
annotation 'llvm.dbg.cu' when LocTrackingOnly is enabled.
Reviewers: echristo, dblaikie
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D4234
llvm-svn: 211609
Referencing a dllimport variable requires actually instructions, not
just a relocation. This fixes PR19955.
Differential Revision: http://reviews.llvm.org/D4249
llvm-svn: 211571
Summary:
Different range metadata can lead to different optimizations in later
passes, possibly breaking the semantics of the merged function. So range
metadata must be taken into consideration when comparing Load
instructions.
Thanks!
llvm-svn: 211391
This patch adds support to recognize patterns such as fadd,fsub,fadd,fsub.../add,sub,add,sub... and
vectorizes them as vector shuffles if they are profitable.
These patterns of vector shuffle can later be converted to instructions such as addsubpd etc on X86.
Thanks to Arnold and Hal for the reviews. http://reviews.llvm.org/D4015
llvm-svn: 211339
We would previously put dllimport variables in switch lookup tables, which
doesn't work because the address cannot be used in a constant initializer.
This is basically the same problem that we have in PR19955.
Putting TLS variables in switch tables also desn't work, because the
address of such a variable is not constant.
Differential Revision: http://reviews.llvm.org/D4220
llvm-svn: 211331
Summary:
With this patch, range metadata can be added to call/invoke including
IntrinsicInst. Previously, it could only be added to load.
Rename computeKnownBitsLoad to computeKnownBitsFromRangeMetadata because
range metadata is not only used by load.
Update the language reference to reflect this change.
Test Plan:
Add several tests in range-2.ll to confirm the verifier is happy with
having range metadata on call/invoke.
Add two tests in AddOverFlow.ll to confirm annotating range metadata to
call/invoke can benefit InstCombine.
Reviewers: meheff, nlewycky, reames, hfinkel, eliben
Reviewed By: eliben
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D4187
llvm-svn: 211281
This patch enables transforms for following patterns.
(x + (~(y & c) + 1) --> x - (y & c)
(x + (~((y >> z) & c) + 1) --> x - ((y>>z) & c)
Differential Revision: http://reviews.llvm.org/D3733
llvm-svn: 211266
* Find factorization opportunities using identity values.
* Find factorization opportunities by treating shl(X, C) as mul (X, shl(C))
* Keep NSW flag while simplifying instruction using factorization.
This fixes PR19263.
Differential Revision: http://reviews.llvm.org/D3799
llvm-svn: 211261
InstCombineMulDivRem has:
// Canonicalize (X+C1)*CI -> X*CI+C1*CI.
InstCombineAddSub has:
// W*X + Y*Z --> W * (X+Z) iff W == Y
These two transforms could fight with each other if C1*CI would not fold
away to something simpler than a ConstantExpr mul.
The InstCombineMulDivRem transform only acted on ConstantInts until
r199602 when it was changed to operate on all Constants in order to
let it fire on ConstantVectors.
To fix this, make this transform more careful by checking to see if we
actually folded away C1*CI.
This fixes PR20079.
llvm-svn: 211258
These will be used for custom lowering and for library
implementations of various math functions, so it's useful
to expose these as builtins.
llvm-svn: 211247
This patch add code to remove unreachable blocks from function
as they may cause jump threading to stuck in infinite loop.
Differential Revision: http://reviews.llvm.org/D3991
llvm-svn: 211103
Summary:
As a starting step, we only use one simple heuristic: if the sign bits
of both a and b are zero, we can prove "add a, b" do not unsigned
overflow, and thus convert it to "add nuw a, b".
Updated all affected tests and added two new tests (@zero_sign_bit and
@zero_sign_bit2) in AddOverflow.ll
Test Plan: make check-all
Reviewers: eliben, rafael, meheff, chandlerc
Reviewed By: chandlerc
Subscribers: chandlerc, llvm-commits
Differential Revision: http://reviews.llvm.org/D4144
llvm-svn: 211084
r199771 accidently broke the logic that makes sure that SROA only splits
load on byte boundaries. If such a split happens, some bits get lost
when reassembling loads of wider types, causing data corruption.
Move the width check up to reject such splits early, avoiding the
corruption. Fixes PR19250.
Patch by: Björn Steinbrink <bsteinbr@gmail.com>
llvm-svn: 211082
[This is resubmitting r210721, which was reverted due to suspected breakage
which turned out to be unrelated].
Some extra review comments were addressed. See D4090 and D4147 for more details.
The Clang change that produces this metadata was committed in r210667
Patch by Mark Heffernan.
llvm-svn: 211076
When LowerSwitch transforms a switch instruction into a tree of ifs it
is actually performing a binary search into the various case ranges, to
see if the current value falls into one cases range of values.
So, if we have a program with something like this:
switch (a) {
case 0:
do0();
break;
case 1:
do1();
break;
case 2:
do2();
break;
default:
break;
}
the code produced is something like this:
if (a < 1) {
if (a == 0) {
do0();
}
} else {
if (a < 2) {
if (a == 1) {
do1();
}
} else {
if (a == 2) {
do2();
}
}
}
This code is inefficient because the check (a == 1) to execute do1() is
not needed.
The reason is that because we already checked that (a >= 1) initially by
checking that also (a < 2) we basically already inferred that (a == 1)
without the need of an extra basic block spawned to check if actually (a
== 1).
The patch addresses this problem by keeping track of already
checked bounds in the LowerSwitch algorithm, so that when the time
arrives to produce a Leaf Block that checks the equality with the case
value / range the algorithm can decide if that block is really needed
depending on the already checked bounds .
For example, the above with "a = 1" would work like this:
the bounds start as LB: NONE , UB: NONE
as (a < 1) is emitted the bounds for the else path become LB: 1 UB:
NONE. This happens because by failing the test (a < 1) we know that the
value "a" cannot be smaller than 1 if we enter the else branch.
After the emitting the check (a < 2) the bounds in the if branch become
LB: 1 UB: 1. This is because by checking that "a" is smaller than 2 then
the upper bound becomes 2 - 1 = 1.
When it is time to emit the leaf block for "case 1:" we notice that 1
can be squeezed exactly in between the LB and UB, which means that if we
arrived to that block there is no need to emit a block that checks if (a
== 1).
Patch by: Marcello Maggioni <hayarms@gmail.com>
llvm-svn: 211038
As a follow-up to r210375 which canonicalizes addrspacecast
instructions, this patch canonicalizes addrspacecast constant
expressions.
Given clang uses ConstantExpr::getAddrSpaceCast to emit addrspacecast
cosntant expressions, this patch is also a step towards having the
frontend emit canonicalized addrspacecasts.
Piggyback a minor refactor in InstCombineCasts.cpp
Update three affected tests in addrspacecast-alias.ll,
access-non-generic.ll and constant-fold-gep.ll and added one new test in
constant-fold-address-space-pointer.ll
llvm-svn: 211004
This patch is to move GlobalMerge pass from Transform/Scalar
to CodeGen, because GlobalMerge depends on TargetMachine.
In the mean time, the macro INITIALIZE_TM_PASS is also moved
to CodeGen/Passes.h. With this fix we can avoid making
libScalarOpts depend on libCodeGen.
llvm-svn: 210951
This also simplifies the IR we create slightly: instead of working out
where success & failure should go manually, it turns out we can just
always jump to a success/failure block created for the purpose. Later
phases will sort out the mess without much difficulty.
llvm-svn: 210917
This has two benefits: it makes the result more suitable for direct
insertaion into the struct to emulate the new cmpxchg, and it means
the name we give the instruction matches its actual effect better.
llvm-svn: 210916
This commit adds a weak variant of the cmpxchg operation, as described
in C++11. A cmpxchg instruction with this modifier is permitted to
fail to store, even if the comparison indicated it should.
As a result, cmpxchg instructions must return a flag indicating
success in addition to their original iN value loaded. Thus, for
uniformity *all* cmpxchg instructions now return "{ iN, i1 }". The
second flag is 1 when the store succeeded.
At the DAG level, a new ATOMIC_CMP_SWAP_WITH_SUCCESS node has been
added as the natural representation for the new cmpxchg instructions.
It is a strong cmpxchg.
By default this gets Expanded to the existing ATOMIC_CMP_SWAP during
Legalization, so existing backends should see no change in behaviour.
If they wish to deal with the enhanced node instead, they can call
setOperationAction on it. Beware: as a node with 2 results, it cannot
be selected from TableGen.
Currently, no use is made of the extra information provided in this
patch. Test updates are almost entirely adapting the input IR to the
new scheme.
Summary for out of tree users:
------------------------------
+ Legacy Bitcode files are upgraded during read.
+ Legacy assembly IR files will be invalid.
+ Front-ends must adapt to different type for "cmpxchg".
+ Backends should be unaffected by default.
llvm-svn: 210903
Enable value forwarding for loads from `calloc()` without an intervening
store.
This change extends GVN to handle the following case:
%1 = tail call noalias i8* @calloc(i64 1, i64 4)
%2 = bitcast i8* %1 to i32*
; This load is trivially constant zero
%3 = load i32* %2, align 4
This is analogous to the handling for `malloc()` in the same places.
`malloc()` returns `undef`; `calloc()` returns a zero value. Note that
it is correct to return zero even for out of bounds GEPs since the
result of such a GEP would be undefined.
Patch by Philip Reames!
llvm-svn: 210828
See http://reviews.llvm.org/D4090 for more details.
The Clang change that produces this metadata was committed in r210667
Patch by Mark Heffernan.
llvm-svn: 210721
This commit is to improve global merge pass and support global symbol merge.
The global symbol merge is not enabled by default. For aarch64, we need some
more back-end fix to make it really benifit ADRP CSE.
llvm-svn: 210640
This improves the X86 cost model for small constants with large types. Before
this commit we would even hoist trivial constants such as i96 2.
This is related to <rdar://problem/17070936>
llvm-svn: 210504
Originally this similar was initiated by Björn Steinbrink here:
http://reviews.llvm.org/D3437
Bug itself has been fixed by principal changes in MergeFunctions. Though
special checks for functions merging are still actual. And the test has
been accepted with slight modifications.
llvm-svn: 210486
For each array index that is in the form of zext(a), convert it to sext(a)
if we can prove zext(a) <= max signed value of typeof(a). The conversion
helps to split zext(x + y) into sext(x) + sext(y).
Reviewed in http://reviews.llvm.org/D4060
llvm-svn: 210444
The messages were
"PR19753: Optimize comparisons with "ashr exact" of a constanst."
"Added support to optimize comparisons with "lshr exact" of a constant."
They were not correctly handling signed/unsigned operation differences,
causing pr19958.
llvm-svn: 210393
addrspacecast X addrspace(M)* to Y addrspace(N)*
-->
bitcast X addrspace(M)* to Y addrspace(M)*
addrspacecast Y addrspace(M)* to Y addrspace(N)*
Updat all affected tests and add several new tests in addrspacecast.ll.
This patch is based on http://reviews.llvm.org/D2186 (authored by Matt
Arsenault) with fixes and more tests.
llvm-svn: 210375
If we have common uses on separate paths in the tree; process the one with greater common depth first.
This makes sure that we do not assume we need to extract a load when it is actually going to be part of a vectorized tree.
Review: http://reviews.llvm.org/D3800
llvm-svn: 210310
Alias with unnamed_addr were in a strange state. It is stored in GlobalValue,
the language reference talks about "unnamed_addr aliases" but the verifier
was rejecting them.
It seems natural to allow unnamed_addr in aliases:
* It is a property of how it is accessed, not of the data itself.
* It is perfectly possible to write code that depends on the address
of an alias.
This patch then makes unname_addr legal for aliases. One side effect is that
the syntax changes for a corner case: In globals, unnamed_addr is now printed
before the address space.
llvm-svn: 210302
Most issues are on mishandling s/zext.
Fixes:
1. When rebuilding new indices, s/zext should be distributed to
sub-expressions. e.g., sext(a +nsw (b +nsw 5)) = sext(a) + sext(b) + 5 but not
sext(a + b) + 5. This also affects the logic of recursively looking for a
constant offset, we need to include s/zext into the context of the searching.
2. Function find should return the bitwidth of the constant offset instead of
always sign-extending it to i64.
3. Stop shortcutting zext'ed GEP indices. LLVM conceptually sign-extends GEP
indices to pointer-size before computing the address. Therefore, gep base,
zext(a + b) != gep base, a + b
Improvements:
1. Add an optimization for splitting sext(a + b): if a + b is proven
non-negative (e.g., used as an index of an inbound GEP) and one of a, b is
non-negative, sext(a + b) = sext(a) + sext(b)
2. Function Distributable checks whether both sext and zext can be distributed
to operands of a binary operator. This helps us split zext(sext(a + b)) to
zext(sext(a) + zext(sext(b)) when a + b does not signed or unsigned overflow.
Refactoring:
Merge some common logic of handling add/sub/or in find.
Testing:
Add many tests in split-gep.ll and split-gep-and-gvn.ll to verify the changes
we made.
llvm-svn: 210291
This patch implements two things:
1. If we know one number is positive and another is negative, we return true as
signed addition of two opposite signed numbers will never overflow.
2. Implemented TODO : If one of the operands only has one non-zero bit, and if
the other operand has a known-zero bit in a more significant place than it
(not including the sign bit) the ripple may go up to and fill the zero, but
won't change the sign. e.x - (x & ~4) + 1
We make sure that we are ignoring 0 at MSB.
Patch by Suyog Sarda.
llvm-svn: 210186
This patch changes GlobalAlias to point to an arbitrary ConstantExpr and it is
up to MC (or the system assembler) to decide if that expression is valid or not.
This reduces our ability to diagnose invalid uses and how early we can spot
them, but it also lets us do things like
@test5 = alias inttoptr(i32 sub (i32 ptrtoint (i32* @test2 to i32),
i32 ptrtoint (i32* @bar to i32)) to i32*)
An important implication of this patch is that the notion of aliased global
doesn't exist any more. The alias has to encode the information needed to
access it in its metadata (linkage, visibility, type, etc).
Another consequence to notice is that getSection has to return a "const char *".
It could return a NullTerminatedStringRef if there was such a thing, but when
that was proposed the decision was to just uses "const char*" for that.
llvm-svn: 210062
The code was actually correct. Sorry for the confusion. I have expanded the
comment saying why the analysis is valid to avoid me misunderstaning it
again in the future.
llvm-svn: 210052
if ((x & C) == 0) x |= C becomes x |= C
if ((x & C) != 0) x ^= C becomes x &= ~C
if ((x & C) == 0) x ^= C becomes x |= C
if ((x & C) != 0) x &= ~C becomes x &= ~C
if ((x & C) == 0) x &= ~C becomes nothing
Differential Revision: http://reviews.llvm.org/D3777
llvm-svn: 210006
Handle "X + ~X" -> "-1" in the function Value *Reassociate::OptimizeAdd(Instruction *I, SmallVectorImpl<ValueEntry> &Ops);
This patch implements:
TODO: We could handle "X + ~X" -> "-1" if we wanted, since "-X = ~X+1".
Patch by Rahul Jain!
Differential Revision: http://reviews.llvm.org/D3835
llvm-svn: 209973
The C and C++ semantics for compare_exchange require it to return a bool
indicating success. This gets mapped to LLVM IR which follows each cmpxchg with
an icmp of the value loaded against the desired value.
When lowered to ldxr/stxr loops, this extra comparison is redundant: its
results are implicit in the control-flow of the function.
This commit makes two changes: it replaces that icmp with appropriate PHI
nodes, and then makes sure earlyCSE is called after expansion to actually make
use of the opportunities revealed.
I've also added -{arm,aarch64}-enable-atomic-tidy options, so that
existing fragile tests aren't perturbed too much by the change. Many
of them either rely on undef/unreachable too pervasively to be
restored to something well-defined (particularly while making sure
they test the same obscure assert from many years ago), or depend on a
particular CFG shape, which is disrupted by SimplifyCFG.
rdar://problem/16227836
llvm-svn: 209883
This patch adds support to vectorize intrinsics such as powi, cttz and ctlz in Vectorizer. These intrinsics are different from other
intrinsics as second argument to these function must be same in order to vectorize them and it should be represented as a scalar.
Review: http://reviews.llvm.org/D3851#inline-32769 and http://reviews.llvm.org/D3937#inline-32857
llvm-svn: 209873
The loop vectorizer instantiates be-taken-count + 1 as the loop iteration count.
If this expression overflows the generated code was invalid.
In case of overflow the code now jumps to the scalar loop.
Fixes PR17288.
llvm-svn: 209854
Currently LLVM will generally merge GEPs. This allows backends to use more
complex addressing modes. In some cases this is not happening because there
is PHI inbetween the two GEPs:
GEP1--\
|-->PHI1-->GEP3
GEP2--/
This patch checks to see if GEP1 and GEP2 are similiar enough that they can be
cloned (GEP12) in GEP3's BB, allowing GEP->GEP merging (GEP123):
GEP1--\ --\ --\
|-->PHI1-->GEP3 ==> |-->PHI2->GEP12->GEP3 == > |-->PHI2->GEP123
GEP2--/ --/ --/
This also breaks certain use chains that are preventing GEP->GEP merges that the
the existing instcombine would merge otherwise.
Tests included.
llvm-svn: 209843
During loop-unroll, loop exits from the current loop may end up in in different
outer loop. This requires to re-form LCSSA recursively for one level down from
the outer most loop where loop exits are landed during unroll. This fixes PR18861.
Differential Revision: http://reviews.llvm.org/D2976
llvm-svn: 209796
Currently LLVM will generally merge GEPs. This allows backends to use more
complex addressing modes. In some cases this is not happening because there
is PHI inbetween the two GEPs:
GEP1--\
|-->PHI1-->GEP3
GEP2--/
This patch checks to see if GEP1 and GEP2 are similiar enough that they can be
cloned (GEP12) in GEP3's BB, allowing GEP->GEP merging (GEP123):
GEP1--\ --\ --\
|-->PHI1-->GEP3 ==> |-->PHI2->GEP12->GEP3 == > |-->PHI2->GEP123
GEP2--/ --/ --/
This also breaks certain use chains that are preventing GEP->GEP merges that the
the existing instcombine would merge otherwise.
Tests included.
llvm-svn: 209755
This patch implements two things:
1. If we know one number is positive and another is negative, we return true as
signed addition of two opposite signed numbers will never overflow.
2. Implemented TODO : If one of the operands only has one non-zero bit, and if
the other operand has a known-zero bit in a more significant place than it
(not including the sign bit) the ripple may go up to and fill the zero, but
won't change the sign. e.x - (x & ~4) + 1
We make sure that we are ignoring 0 at MSB.
Patch by Suyog Sarda.
llvm-svn: 209746
This is an enhancement to SeparateConstOffsetFromGEP. With this patch, we can
extract a constant offset from "s/zext and/or/xor A, B".
Added a new test @ext_or to verify this enhancement.
Refactoring the code, I also extracted some common logic to function
Distributable.
llvm-svn: 209670
Detected by Daniel Jasper, Ilia Filippov, and Andrea Di Biagio
Fixed the argument order to select (the mask semantics to blendv* are the
inverse of select) and fixed the tests
Added parenthesis to the assert condition
Ran clang-format
llvm-svn: 209667
Summary:
Implemented an InstCombine transformation that takes a blendv* intrinsic
call and translates it into an IR select, if the mask is constant.
This will eventually get lowered into blends with immediates if possible,
or pblendvb (with an option to further optimize if we can transform the
pblendvb into a blend+immediate instruction, depending on the selector).
It will also enable optimizations by the IR passes, which give up on
sight of the intrinsic.
Both the transformation and the lowering of its result to asm got shiny
new tests.
The transformation is a bit convoluted because of blendvp[sd]'s
definition:
Its mask is a floating point value! This forces us to convert it and get
the highest bit. I suppose this happened because the mask has type
__m128 in Intel's intrinsic and v4sf (for blendps) in gcc's builtin.
I will send an email to llvm-dev to discuss if we want to change this or
not.
Reviewers: grosbach, delena, nadav
Differential Revision: http://reviews.llvm.org/D3859
llvm-svn: 209643
This commit starts with a "git mv ARM64 AArch64" and continues out
from there, renaming the C++ classes, intrinsics, and other
target-local objects for consistency.
"ARM64" test directories are also moved, and tests that began their
life in ARM64 use an arm64 triple, those from AArch64 use an aarch64
triple. Both should be equivalent though.
This finishes the AArch64 merge, and everyone should feel free to
continue committing as normal now.
llvm-svn: 209577
I'm doing this in two phases for a better "git blame" record. This
commit removes the previous AArch64 backend and redirects all
functionality to ARM64. It also deduplicates test-lines and removes
orphaned AArch64 tests.
The next step will be "git mv ARM64 AArch64" and rewire most of the
tests.
Hopefully LLVM is still functional, though it would be even better if
no-one ever had to care because the rename happens straight
afterwards.
llvm-svn: 209576
Fixed a TODO in r207783.
Add the extracted constant offset using GEP instead of ugly
ptrtoint+add+inttoptr. Using GEP simplifies future optimizations and makes IR
easier to understand.
Updated all affected tests, and added a new test in split-gep.ll to cover a
corner case where emitting uglygep is necessary.
llvm-svn: 209537
ScalarEvolution::isKnownPredicate() can wrongly reduce a comparison
when both the LHS and RHS are SCEVAddRecExprs. This checks that both
LHS and RHS are guarded in the case when both are SCEVAddRecExprs.
The test case is against indvars because I could not find a way to
directly test SCEV.
Patch by Sanjay Patel!
llvm-svn: 209487
Summary:
This adds two new diagnostics: -pass-remarks-missed and
-pass-remarks-analysis. They take the same values as -pass-remarks but
are intended to be triggered in different contexts.
-pass-remarks-missed is used by LLVMContext::emitOptimizationRemarkMissed,
which passes call when they tried to apply a transformation but
couldn't.
-pass-remarks-analysis is used by LLVMContext::emitOptimizationRemarkAnalysis,
which passes call when they want to inform the user about analysis
results.
The patch also:
1- Adds support in the inliner for the two new remarks and a
test case.
2- Moves emitOptimizationRemark* functions to the llvm namespace.
3- Adds an LLVMContext argument instead of making them member functions
of LLVMContext.
Reviewers: qcolombet
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D3682
llvm-svn: 209442
Currently the X86 backend doesn't support types larger than i128 very well. For
example an i192 multiply will assert in codegen when the 2nd argument is a constant and the constant got hoisted.
This fix changes the cost model to never hoist constants for types larger than
i128. Once the codegen issues have been resolved, the cost model can be updated
to allow also larger types.
This is related to <rdar://problem/16954938>
llvm-svn: 209162
The cost model conservatively assumes that it will always get scalarized and
that's about as good as we can get with the generic TTI; reasoning whether a
shuffle with an efficient lowering is available is hard. We can override that
conservative estimate for some targets in the future.
llvm-svn: 209125
Currently LLVM will generally merge GEPs. This allows backends to use more
complex addressing modes. In some cases this is not happening because there
is PHI inbetween the two GEPs:
GEP1--\
|-->PHI1-->GEP3
GEP2--/
This patch checks to see if GEP1 and GEP2 are similiar enough that they can be
cloned (GEP12) in GEP3's BB, allowing GEP->GEP merging (GEP123):
GEP1--\ --\ --\
|-->PHI1-->GEP3 ==> |-->PHI2->GEP12->GEP3 == > |-->PHI2->GEP123
GEP2--/ --/ --/
This also breaks certain use chains that are preventing GEP->GEP merges that the
the existing instcombine would merge otherwise.
Tests included.
rdar://15547484
llvm-svn: 209049
This allows us to put dynamic initializers for weak data into the same
comdat group as the data being initialized. This is necessary for MSVC
ABI compatibility. Once we have comdats for guard variables, we can use
the combination to help GlobalOpt fire more often for weak data with
guarded initialization on other platforms.
Reviewers: nlewycky
Differential Revision: http://reviews.llvm.org/D3499
llvm-svn: 209015
This patch changes the design of GlobalAlias so that it doesn't take a
ConstantExpr anymore. It now points directly to a GlobalObject, but its type is
independent of the aliasee type.
To avoid changing all alias related tests in this patches, I kept the common
syntax
@foo = alias i32* @bar
to mean the same as now. The cases that used to use cast now use the more
general syntax
@foo = alias i16, i32* @bar.
Note that GlobalAlias now behaves a bit more like GlobalVariable. We
know that its type is always a pointer, so we omit the '*'.
For the bitcode, a nice surprise is that we were writing both identical types
already, so the format change is minimal. Auto upgrade is handled by looking
through the casts and no new fields are needed for now. New bitcode will
simply have different types for Alias and Aliasee.
One last interesting point in the patch is that replaceAllUsesWith becomes
smart enough to avoid putting a ConstantExpr in the aliasee. This seems better
than checking and updating every caller.
A followup patch will delete getAliasedGlobal now that it is redundant. Another
patch will add support for an explicit offset.
llvm-svn: 209007
Summary:
Analyze the range of values produced by ashr/lshr cst, %V when it is
being used in an icmp.
Reviewers: nicholas
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D3774
llvm-svn: 209000
Summary:
The dividend in an sdiv tells us the largest and smallest possible
results. Use this fact to optimize comparisons against an sdiv with a
constant dividend.
Reviewers: nicholas
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D3795
llvm-svn: 208999
This reverts commit r208934.
The patch depends on aliases to GEPs with non zero offsets. That is not
supported and fairly broken.
The good news is that GlobalAlias is being redesigned and will have support
for offsets, so this patch should be a nice match for it.
llvm-svn: 208978
This commit implements two command line switches -global-merge-on-external
and -global-merge-aligned, and both of them are false by default, so this
optimization is disabled by default for all targets.
For ARM64, some back-end behaviors need to be tuned to get this optimization
further enabled.
llvm-svn: 208934
The allocas going out of scope are immediately killed by the return
instruction.
This is a resend of r208912, which was committed accidentally.
Reviewers: chandlerc
Differential Revision: http://reviews.llvm.org/D3792
llvm-svn: 208920
The allocas going out of scope are immediately killed by the return
instruction.
Reviewers: chandlerc
Differential Revision: http://reviews.llvm.org/D3630
llvm-svn: 208912
The interesting case is what happens when you inline a musttail call
through a musttail call site. In this case, we can't break perfect
forwarding or allow any stack growth.
Instead of merging control flow from the inlined return instruction
after a musttail call into the body of the caller, leave the inlined
return instruction in the caller so that the musttail call stays in the
tail position.
More work is required in http://reviews.llvm.org/D3630 to handle the
case where the inlined function has dynamic allocas or byval arguments.
Reviewers: chandlerc
Differential Revision: http://reviews.llvm.org/D3491
llvm-svn: 208910
much more effectively when trying to constant fold a load of a constant.
Previously, we only handled bitcasts by trying to find a totally generic
byte representation of the constant and use that. Now, we look through
the bitcast to see what constant we might fold the load into, and then
try to form a constant expression cast of the found value that would be
equivalent to loading the value.
You might wonder why on earth this actually matters. Well, turns out
that the Itanium ABI causes us to create a single array for a vtable
where the first elements are virtual base offsets, followed by the
virtual function pointers. Because the array is homogenous the element
type is consistently i8* and we inttoptr the virtual base offsets into
the initial elements.
Then constructors bitcast these pointers to i64 pointers prior to
loading them. Boom, no more constant folding of virtual base offsets.
This is the first fix to LLVM to address the *insane* performance Eric
Niebler discovered with Clang on his range comprehensions[1]. There is
more to come though, this doesn't *really* fix the problem fully.
[1]: http://ericniebler.com/2014/04/27/range-comprehensions/
llvm-svn: 208856
if ((x & C) == 0) x |= C becomes x |= C
if ((x & C) != 0) x ^= C becomes x &= ~C
if ((x & C) == 0) x ^= C becomes x |= C
if ((x & C) != 0) x &= ~C becomes x &= ~C
if ((x & C) == 0) x &= ~C becomes nothing
Z3 Verifications code for above transform
http://rise4fun.com/Z3/Pmsh
Differential Revision: http://reviews.llvm.org/D3717
llvm-svn: 208848
Summary:
This gets rid of a sub instruction by moving the negation to the
constant when valid.
Reviewers: nicholas
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D3773
llvm-svn: 208827
Tested by comparing make check VERBOSE=1 before and after to make sure
no tests are missed. (VERBOSE=1 prints the list of tests.)
Only one test :( remains where .cpp is required:
tools/llvm-cov/range_based_for.cpp:// RUN: llvm-cov range_based_for.cpp | FileCheck %s --check-prefix=STDOUT
The topic was discussed in this thread:
http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20140428/214905.html
llvm-svn: 208621
In transformation:
BinOp(shuffle(v1,undef), shuffle(v2,undef)) -> shuffle(BinOp(v1, v2),undef)
type of the undef argument must be same as type of BinOp.
llvm-svn: 208531
Do not apply transformation:
BinOp(shuffle(v1), shuffle(v2)) -> shuffle(BinOp(v1, v2))
if operands v1 and v2 are of different size.
This change fixes PR19717, which was caused by r208488.
llvm-svn: 208518
This patch enables transformations:
BinOp(shuffle(v1), shuffle(v2)) -> shuffle(BinOp(v1, v2))
BinOp(shuffle(v1), const1) -> shuffle(BinOp, const2)
They allow to eliminate extra shuffles in some cases.
Differential Revision: http://reviews.llvm.org/D3525
llvm-svn: 208488
There is no total ordering if the CFG is disconnected. We don't care if we
catch all CSE opportunities in dead code either so just exclude ignore them in
the assert.
PR19646
llvm-svn: 208461
Since ExtractValue is not included in ComputeSpeculationCost CFGs containing
ExtractValueInsts cannot be simplified. In particular this interacts with
InstCombineCompare's tendency to insert add.with.overflow intrinsics for
certain idiomatic math operations, preventing optimization.
This patch adds ExtractValue to the ComputeSpeculationCost. Test case included
rdar://14853450
llvm-svn: 208434
The old method used by X86TTI to determine partial-unrolling thresholds was
messy (because it worked by testing target features), and also would not
correctly identify the target CPU if certain target features were disabled.
After some discussions on IRC with Chandler et al., it was decided that the
processor scheduling models were the right containers for this information
(because it is often tied to special uop dispatch-buffer sizes).
This does represent a small functionality change:
- For generic x86-64 (which uses the SB model and, thus, will get some
unrolling).
- For AMD cores (because they still currently use the SB scheduling model)
- For Haswell (based on benchmarking by Louis Gerbarg, it was decided to bump
the default threshold to 50; we're working on a test case for this).
Otherwise, nothing has changed for any other targets. The logic, however, has
been moved into BasicTTI, so other targets may now also opt-in to this
functionality simply by setting LoopMicroOpBufferSize in their processor
model definitions.
llvm-svn: 208289
Visibilities of `hidden` and `protected` are meaningless for symbols
with local linkage.
- Change the assembler to reject non-default visibility on symbols
with local linkage.
- Change the bitcode reader to auto-upgrade `hidden` and `protected`
to `default` when the linkage is local.
- Update LangRef.
<rdar://problem/16141113>
llvm-svn: 208263
The number of tail call to loop conversions remains the same (1618 by my count).
The new algorithm does a local scan over the use-def chains to identify local "alloca-derived" values, as well as points where the alloca could escape. Then, a visit over the CFG marks blocks as being before or after the allocas have escaped, and annotates the calls accordingly.
llvm-svn: 208017
Visibility is meaningless when the linkage is local. Change
`-internalize` to reset the visibility to `default`.
<rdar://problem/16141113>
llvm-svn: 207979
Otherwise we use the same threshold as for complete unrolling, which is
way too high. This made us unroll any loop smaller than 150 instructions
by 8 times, but only if someone specified -march=core2 or better,
which happens to be the default on darwin.
llvm-svn: 207940
When can't assume a vectorized tree is rooted in an instruction. The IRBuilder
could have constant folded it. When we rebuild the build_vector (the series of
InsertElement instructions) use the last original InsertElement instruction. The
vectorized tree root is guaranteed to be before it.
Also, we can't assume that the n-th InsertElement inserts the n-th element into
a vector.
This reverts r207746 which reverted the revert of the revert of r205018 or so.
Fixes the test case in PR19621.
llvm-svn: 207939
This moves most of GlobalOpt's constructor optimization
code out of GlobalOpt into Transforms/Utils/CDtorUtils.{h,cpp}. The
public interface is a single function OptimizeGlobalCtorsList() that
takes a predicate returning which constructors to remove.
GlobalOpt calls this with a function that statically evaluates all
constructors, just like it did before. This part of the change is
behavior-preserving.
Also add a call to this from GlobalDCE with a filter that removes global
constructors that contain a "ret" instruction and nothing else – this
fixes PR19590.
llvm-svn: 207856
address to AnalyzeLoadFromClobberingLoad. This fixes a bug in load-PRE where
PRE is applied to a load that is not partially redundant.
<rdar://problem/16638765>.
llvm-svn: 207853
This optimization merges the common part of a group of GEPs, so we can compute
each pointer address by adding a simple offset to the common part.
The optimization is currently only enabled for the NVPTX backend, where it has
a large payoff on some benchmarks.
Review: http://reviews.llvm.org/D3462
Patch by Jingyue Wu.
llvm-svn: 207783
=[
Turns out that this was the root cause of PR19621. We found a crasher
only recently (likely due to improvements elsewhere in the SLP
vectorizer) but the reduced test case failed all the way back to here.
I've confirmed that reverting this patch both fixes the reduced test
case in PR19621 and the actual source file that led to it, so it seems
to really be rooted here. I've replied to the commit thread with
discussion of my (feeble) attempts to debug this. Didn't make it very
far, so reverting now that we have a good test case so that things can
get back to healthy while the debugging carries on.
llvm-svn: 207746
There is no need to check if we want to hoist the immediate value of an
shift instruction. Simply return TCC_Free right away.
This change is like r206101, but for X86.
rdar://problem/16190769
llvm-svn: 207692
The instcomine logic to handle vpermilvar's pd and 256 variants was incorrect.
The _256 variants have indexes into the individual 128 bit lanes and in all
cases it also has to mask out unused bits.
llvm-svn: 207577
This patch changes the vectorization remarks to also inform when
vectorization is possible but not beneficial.
Added tests to exercise some loop remarks.
llvm-svn: 207574
clang directly from the LLVM test suite! That doesn't work. I've
followed up on the review thread to try and get a viable solution sorted
out, but trying to get the tree clean here.
llvm-svn: 207462
by avoiding inlining massive switches merely because they have no
instructions in them. These switches still show up where we fail to form
lookup tables, and in those cases they are actually going to cause
a very significant code size hit anyways, so inlining them is not the
right call. The right way to fix any performance regressions stemming
from this is to enhance the switch-to-lookup-table logic to fire in more
places.
This makes PR19499 about 5x less bad. It uncovers a second compile time
problem in that test case that is unrelated (surprisingly!).
llvm-svn: 207403
more than 1 instruction. The caller need to be aware of this
and adjust instruction iterators accordingly.
rdar://16679376
Repaired r207302.
llvm-svn: 207309
right intrinsics.
A packed logical shift right with a shift count bigger than or equal to the
element size always produces a zero vector. In all other cases, it can be
safely replaced by a 'lshr' instruction.
llvm-svn: 207299
Consider this use from the new testcase:
LSR Use: Kind=ICmpZero, Offsets={0}, widest fixup type: i32
reg({1000,+,-1}<nw><%for.body>)
-3003 + reg({3,+,3}<nw><%for.body>)
-1001 + reg({1,+,1}<nuw><nsw><%for.body>)
-1000 + reg({0,+,1}<nw><%for.body>)
-3000 + reg({0,+,3}<nuw><%for.body>)
reg({-1000,+,1}<nw><%for.body>)
reg({-3000,+,3}<nsw><%for.body>)
This is the last use we consider for a solution in SolveRecurse, so CurRegs is
a large set. (CurRegs is the set of registers that are needed by the
previously visited uses in the in-progress solution.)
ReqRegs is {
{3,+,3}<nw><%for.body>,
{1,+,1}<nuw><nsw><%for.body>
}
This is the intersection of the regs used by any of the formulas for the
current use and CurRegs.
Now, the code requires a formula to contain *all* these regs (the comment is
simply wrong), otherwise the formula is immediately disqualified. Obviously,
no formula for this use contains two regs so they will all get disqualified.
The fix modifies the check to allow the formula in this case. The idea is
that neither of these formulae is introducing any new registers which is the
point of this early pruning as far as I understand.
In terms of set arithmetic, we now allow formulas whose used regs are a subset
of the required regs not just the other way around.
There are few more loops in the test-suite that are now successfully LSRed. I
have benchmarked those and found very minimal change.
Fixes <rdar://problem/13965777>
llvm-svn: 207271
override the default cold threshold.
When we use command line argument to set the inline threshold, the default
cold threshold will not be used. This is in line with how we use
OptSizeThreshold. When we want a higher threshold for all functions, we
do not have to set both inline threshold and cold threshold.
llvm-svn: 207245
This excludes avx512 as I don't have hardware to verify. It excludes _dq
variants because they are represented in the IR as <{2,4} x i64> when it's
actually a byte shift of the entire i{128,265}.
This also excludes _dq_bs as they aren't at all supported by the backend.
There are also no corresponding instructions in the ISA. I have no idea why
they exist...
llvm-svn: 207058
Summary:
Since the upper 64 bits of the destination register are undefined when
performing this operation, we can substitute it and let the optimizer
figure out that only a copy is needed.
Also added range merging, if an instruction copies a range that can be
merged with a previous copied range.
Added test cases for both optimizations.
Reviewers: grosbach, nadav
CC: llvm-commits
Differential Revision: http://reviews.llvm.org/D3357
llvm-svn: 207055
Use -stats to see how many loops were analyzed for possible vectorization and how many of them were actually vectorized.
Patch by Zinovy Nis
Differential Revision: http://reviews.llvm.org/D3438
llvm-svn: 206956
In the case where the constant comes from a cloned cast instruction, the
materialization code has to go before the cloned cast instruction.
This commit fixes the method that finds the materialization insertion point
by making it aware of this case.
This fixes <rdar://problem/15532441>
llvm-svn: 206913
With a constant mask a vpermil* is just a shufflevector. This patch implements
that simplification. This allows us to produce denser code. It should also
allow more folding down the line.
llvm-svn: 206801
The -tailcallelim pass should be checking if byval or inalloca args can
be captured before marking calls as tail calls. This was the real root
cause of PR7272.
With a better fix in place, revert the inliner change from r105255. The
test case it introduced still passes and has been moved to
test/Transforms/Inline/byval-tail-call.ll.
Reviewers: chandlerc
Differential Revision: http://reviews.llvm.org/D3403
llvm-svn: 206789
Summary:
This prevents the discriminator generation pass from triggering if
the DWARF version being used in the module is prior to 4.
Reviewers: echristo, dblaikie
CC: llvm-commits
Differential Revision: http://reviews.llvm.org/D3413
llvm-svn: 206507
After some discussions the preferred semantics of
the always_inline attribute is
inline always when the compiler can determine
that it it safe to do so.
llvm-svn: 206487
Still only 32-bit ARM using it at this stage, but the promotion allows
direct testing via opt and is a reasonably self-contained patch on the
way to switching ARM64.
At this point, other targets should be able to make use of it without
too much difficulty if they want. (See ARM64 commit coming soon for an
example).
llvm-svn: 206485
is set even when it contains a indirect branch.
The attribute overrules correctness concerns
like the escape of a local block address.
This is for rdar://16501761
llvm-svn: 206429
Implements the various TTI functions to enable constant hoisting on PPC. The
only significant test-suite change is this:
MultiSource/Benchmarks/VersaBench/bmm/bmm - 20% speedup
(which essentially reverses the slowdown from r206120).
llvm-svn: 206141
If multiplication involves zero-extended arguments and the result is
compared as in the patterns:
%mul32 = trunc i64 %mul64 to i32
%zext = zext i32 %mul32 to i64
%overflow = icmp ne i64 %mul64, %zext
or
%overflow = icmp ugt i64 %mul64 , 0xffffffff
then the multiplication may be replaced by call to umul.with.overflow.
This change fixes PR4917 and PR4918.
Differential Revision: http://llvm-reviews.chandlerc.com/D2814
llvm-svn: 206137
Originally the cost model would give up for large constants and just return the
maximum cost. This is not what we want for constant hoisting, because some of
these constants are large in bitwidth, but are still cheap to materialize.
This commit fixes the cost model to either return TCC_Free if the cost cannot be
determined, or accurately calculate the cost even for large constants
(bitwidth > 128).
This fixes <rdar://problem/16591573>.
llvm-svn: 206100
The current memory-instruction optimization logic in CGP, which sinks parts of
the address computation that can be adsorbed by the addressing mode, does this
by explicitly converting the relevant part of the address computation into
IR-level integer operations (making use of ptrtoint and inttoptr). For most
targets this is currently not a problem, but for targets wishing to make use of
IR-level aliasing analysis during CodeGen, the use of ptrtoint/inttoptr is a
problem for two reasons:
1. BasicAA becomes less powerful in the face of the ptrtoint/inttoptr
2. In cases where type-punning was used, and BasicAA was used
to override TBAA, BasicAA may no longer do so. (this had forced us to disable
all use of TBAA in CodeGen; something which we can now enable again)
This (use of GEPs instead of ptrtoint/inttoptr) is not currently enabled by
default (except for those targets that use AA during CodeGen), and so aside
from some PowerPC subtargets and SystemZ, there should be no change in
behavior. We may be able to switch completely away from the ptrtoint/inttoptr
sinking on all targets, but further testing is required.
I've doubled-up on a number of existing tests that are sensitive to the
address sinking behavior (including some store-merging tests that are
sensitive to the order of the resulting ADD operations at the SDAG level).
llvm-svn: 206092
The immediate cost calculation code was hitting an assertion in the included
test case, because APInt was still internally 128-bits. Truncating it to 64-bits
fixed the issue.
Fixes <rdar://problem/16572521>.
llvm-svn: 205947
This implements the target-hooks for ARM64 to enable constant hoisting.
This fixes <rdar://problem/14774662> and <rdar://problem/16381500>.
llvm-svn: 205791
into a constant size alloca by inlining.
Ran a run over the testsuite, no results out of the noise, fixes
the testcase in the PR.
PR19115.
llvm-svn: 205710
Some Intrinsics are overloaded to the extent that return type equality (all
that's been checked up to now) does not guarantee that the arguments are the
same. In these cases SLP vectorizer should not recurse into the operands, which
can be achieved by comparing them as "Function *" rather than simply the ID.
llvm-svn: 205424
For the purpose of calculating the cost of the loop at various vectorization
factors, we need to count dependencies of consecutive pointers as uniforms
(which means that the VF = 1 cost is used for all overall VF values).
For example, the TSVC benchmark function s173 has:
...
%3 = add nsw i64 %indvars.iv, 16000
%arrayidx8 = getelementptr inbounds %struct.GlobalData* @global_data, i64 0, i32 0, i64 %3
...
and we must realize that the add will be a scalar in order to correctly deduce
it to be profitable to vectorize this on PowerPC with VSX enabled. In fact, all
dependencies of a consecutive pointer must be a scalar (uniform), and so we
simply need to add all consecutive pointers to the worklist that currently
detects collects uniforms.
Fixes PR19296.
llvm-svn: 205387
This provides an initial implementation of getUnrollingPreferences for x86.
getUnrollingPreferences is used by the generic (concatenation) unroller, which
is distinct from the unrolling done by the loop vectorizer. Many modern x86
cores have some kind of uop cache and loop-stream detector (LSD) used to
efficiently dispatch small loops, and taking full advantage of this requires
unrolling small loops (small here means 10s of uops).
These caches also have limits on the number of taken branches in the loop, and
so we also cap the loop unrolling factor based on the maximum "depth" of the
loop. This is currently calculated with a partial DFS traversal (partial
because it will stop early if the path length grows too much). This is still an
approximation, and one that is both conservative (because it does not account
for branches eliminated via block placement) and optimistic (because it is only
recording the maximum depth over minimum paths). Nevertheless, because the
loops that fit in these uop caches are so small, it is not clear how much the
details matter.
The original set of patches posted for review produced the following test-suite
performance results (from the TSVC benchmark) at that time:
ControlLoops-dbl - 13% speedup
ControlLoops-flt - 15% speedup
Reductions-dbl - 7.5% speedup
llvm-svn: 205348
The generic (concatenation) loop unroller is currently placed early in the
standard optimization pipeline. This is a good place to perform full unrolling,
but not the right place to perform partial/runtime unrolling. However, most
targets don't enable partial/runtime unrolling, so this never mattered.
However, even some x86 cores benefit from partial/runtime unrolling of very
small loops, and follow-up commits will enable this. First, we need to move
partial/runtime unrolling late in the optimization pipeline (importantly, this
is after SLP and loop vectorization, as vectorization can drastically change
the size of a loop), while keeping the full unrolling where it is now. This
change does just that.
llvm-svn: 205264
This reverts commit r205018.
Conflicts:
lib/Transforms/Vectorize/SLPVectorizer.cpp
test/Transforms/SLPVectorizer/X86/insert-element-build-vector.ll
This is breaking libclc build.
llvm-svn: 205260
There is no direct AVX instruction to convert to unsigned. I have some ideas
how we may be able to do this with three vector instructions but the current
backend just bails on this to get it scalarized.
See the comment why we need to adjust the cost returned by BasicTTI.
The test is a bit roundabout (and checks assembly rather than bit code) because
I'd like it to work even if at some point we could vectorize this conversion.
Fixes <rdar://problem/16371920>
llvm-svn: 205159
This adds a second implementation of the AArch64 architecture to LLVM,
accessible in parallel via the "arm64" triple. The plan over the
coming weeks & months is to merge the two into a single backend,
during which time thorough code review should naturally occur.
Everything will be easier with the target in-tree though, hence this
commit.
llvm-svn: 205090
This reverts commit r204912, and follow-up commit r204948.
This introduced a performance regression, and the fix is not completely
clear yet.
llvm-svn: 205010
This reverts commit r203553, and follow-up commits r203558 and r203574.
I will follow this up on the mailinglist to do it in a way that won't
cause subtle PRE bugs.
llvm-svn: 205009
Fixes a miscompile introduced in r204912. It would miscompile code like
(unsigned)(a + -49) <= 5U. The transform would turn this into
(unsigned)a < 55U, which would return true for values in [0, 49], when
it should not.
llvm-svn: 204948
This adds back r204781.
Original message:
Aliases are just another name for a position in a file. As such, the
regular symbol resolutions are not applied. For example, given
define void @my_func() {
ret void
}
@my_alias = alias weak void ()* @my_func
@my_alias2 = alias void ()* @my_alias
We produce without this patch:
.weak my_alias
my_alias = my_func
.globl my_alias2
my_alias2 = my_alias
That is, in the resulting ELF file my_alias, my_func and my_alias are
just 3 names pointing to offset 0 of .text. That is *not* the
semantics of IR linking. For example, linking in a
@my_alias = alias void ()* @other_func
would require the strong my_alias to override the weak one and
my_alias2 would end up pointing to other_func.
There is no way to represent that with aliases being just another
name, so the best solution seems to be to just disallow it, converting
a miscompile into an error.
llvm-svn: 204934
This reverts commit r204781.
I will follow up to with msan folks to see what is what they
were trying to do with aliases to weak aliases.
llvm-svn: 204784
Aliases are just another name for a position in a file. As such, the
regular symbol resolutions are not applied. For example, given
define void @my_func() {
ret void
}
@my_alias = alias weak void ()* @my_func
@my_alias2 = alias void ()* @my_alias
We produce without this patch:
.weak my_alias
my_alias = my_func
.globl my_alias2
my_alias2 = my_alias
That is, in the resulting ELF file my_alias, my_func and my_alias are
just 3 names pointing to offset 0 of .text. That is *not* the
semantics of IR linking. For example, linking in a
@my_alias = alias void ()* @other_func
would require the strong my_alias to override the weak one and
my_alias2 would end up pointing to other_func.
There is no way to represent that with aliases being just another
name, so the best solution seems to be to just disallow it, converting
a miscompile into an error.
llvm-svn: 204781
Summary:
Previously the code didn't check if the before and after types for the
store were pointers to different address spaces. This resulted in
instcombine using a bitcast to convert between pointers to different
address spaces, causing an assertion due to the invalid cast.
It is not be appropriate to use addrspacecast this case because it is
not guaranteed to be a no-op cast. Instead bail out and do not do the
transformation.
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D3117
llvm-svn: 204733
A PHI node usually has only one value/basic block pair per incoming basic block.
In the case of a switch statement it is possible that a following PHI node may
have more than one such pair per incoming basic block. E.g.:
%0 = phi i64 [ 123456, %case2 ], [ 654321, %Entry ], [ 654321, %Entry ]
This is valid and the verfier doesn't complain, because both values are the
same.
Constant hoisting materializes the constant for each operand separately and the
value is still the same, but the variable names have changed. As a result the
verfier can't recognize anymore that they are the same value and complains.
This fix adds special update code for PHI node in constant hoisting to prevent
this corner case.
This fixes <rdar://problem/16394449>
llvm-svn: 204537
Extend the target hook to take also the operand index into account when
calculating the cost of the constant materialization.
Related to <rdar://problem/16381500>
llvm-svn: 204435
Originally the algorithm would search for expensive constants and track their
users, which could be instructions and constant expressions. This change only
tracks the constants for instructions, but constant expressions are indirectly
covered too. If an operand is an constant expression, then we look through the
expression to find anny expensive constants.
The algorithm keep now track of the instruction and the operand index where the
constant is used. This allows more precise hoisting of constant materialization
code for PHI instructions, because we only hoist to the basic block of the
incoming operand. Before we had to find the idom of all PHI operands and hoist
the materialization code there.
This also makes updating of instructions easier. Before we had to keep track of
the original constant, find it in the instructions, and then replace it. Now we
can just simply update the operand.
Related to <rdar://problem/16381500>
llvm-svn: 204433
This commit extends the coverage of the constant hoisting pass, adds additonal
debug output and updates the function names according to the style guide.
Related to <rdar://problem/16381500>
llvm-svn: 204389
This option caused LowerInvoke to generate code using SJLJ-based
exception handling, but there is no code left that interprets the
jmp_buf stack that the resulting code maintained (llvm.sjljeh.jblist).
This option has been obsolete for a while, and replaced by
SjLjEHPrepare.
This leaves the default behaviour of LowerInvoke, which is to convert
invokes to calls.
Differential Revision: http://llvm-reviews.chandlerc.com/D3136
llvm-svn: 204388
None of the existing tests for LowerInvoke check LowerInvoke's output,
and all but one use "-enable-correct-eh-support", which is obsolete,
so those tests will be removed when that option is removed.
To make sure LowerInvoke will still have test coverage, this adds a
test for its default mode which converts invokes to calls.
Differential Revision: http://llvm-reviews.chandlerc.com/D3124
llvm-svn: 204344
The use_iterator redesign in r203364 introduced an increment past the
end of a range in -objc-arc-contract. Added an explicit check for the
end of the range.
<rdar://problem/16333235>
llvm-svn: 204195
Summary:
The compiler does not always generate linkage names. If a function
has been inlined and its body elided, its linkage name may not be
generated.
When the binary executes, the profiler will use its unmangled name
when attributing samples. This results in unmangled names in the
input profile.
We are currently failing hard when this happens. However, in this case
all that happens is that we fail to attribute samples to the inlined
function. While this means fewer optimization opportunities, it should
not cause a compilation failure.
This patch accepts all valid function names, regardless of whether
they were mangled or not.
Reviewers: chandlerc
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D3087
llvm-svn: 204142
When GlobalOpt has determined that a GlobalVariable only ever has two values,
it would convert the GlobalVariable to a boolean, and introduce SelectInsts
at every load, to choose between the two possible values. These SelectInsts
introduce overhead and other unpleasantness.
This patch makes GlobalOpt just add range metadata to loads from such
GlobalVariables instead. This enables the same main optimization (as seen in
test/Transforms/GlobalOpt/integer-bool.ll), without introducing selects.
The main downside is that it doesn't get the memory savings of shrinking such
GlobalVariables, but this is expected to be negligible.
llvm-svn: 204076
Summary:
The sample profiler pass emits several error messages. Instead of
just aborting the compiler with report_fatal_error, we can emit
better messages using DiagnosticInfo.
This adds a new sub-class of DiagnosticInfo to handle the sample
profiler.
Reviewers: chandlerc, qcolombet
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D3086
llvm-svn: 203976
These linkages were introduced some time ago, but it was never very
clear what exactly their semantics were or what they should be used
for. Some investigation found these uses:
* utf-16 strings in clang.
* non-unnamed_addr strings produced by the sanitizers.
It turns out they were just working around a more fundamental problem.
For some sections a MachO linker needs a symbol in order to split the
section into atoms, and llvm had no idea that was the case. I fixed
that in r201700 and it is now safe to use the private linkage. When
the object ends up in a section that requires symbols, llvm will use a
'l' prefix instead of a 'L' prefix and things just work.
With that, these linkages were already dead, but there was a potential
future user in the objc metadata information. I am still looking at
CGObjcMac.cpp, but at this point I am convinced that linker_private
and linker_private_weak are not what they need.
The objc uses are currently split in
* Regular symbols (no '\01' prefix). LLVM already directly provides
whatever semantics they need.
* Uses of a private name (start with "\01L" or "\01l") and private
linkage. We can drop the "\01L" and "\01l" prefixes as soon as llvm
agrees with clang on L being ok or not for a given section. I have two
patches in code review for this.
* Uses of private name and weak linkage.
The last case is the one that one could think would fit one of these
linkages. That is not the case. The semantics are
* the linker will merge these symbol by *name*.
* the linker will hide them in the final DSO.
Given that the merging is done by name, any of the private (or
internal) linkages would be a bad match. They allow llvm to rename the
symbols, and that is really not what we want. From the llvm point of
view, these objects should really be (linkonce|weak)(_odr)?.
For now, just keeping the "\01l" prefix is probably the best for these
symbols. If we one day want to have a more direct support in llvm,
IMHO what we should add is not a linkage, it is just a hidden_symbol
attribute. It would be applicable to multiple linkages. For example,
on weak it would produce the current behavior we have for objc
metadata. On internal, it would be equivalent to private (and we
should then remove private).
llvm-svn: 203866
Summary:
This helps the instruction selector to lower an i64 * i64 -> i128
multiplication into a single instruction on targets which support it.
This is an update of D2973 which was reverted because of a bug reported
as PR19084.
Reviewers: t.p.northover, chapuni
Reviewed By: t.p.northover
CC: llvm-commits, alex, chapuni
Differential Revision: http://llvm-reviews.chandlerc.com/D3021
llvm-svn: 203797
On ELF and COFF an alias is just another name for a position in the file.
There is no way to refer to a position in another file, so an alias to
undefined is meaningless.
MachO currently doesn't support aliases. The spec has a N_INDR, which when
implemented will have a different set of restrictions. Adding support for
it shouldn't be harder than any other IR extension.
For now, having the IR represent what is actually possible with current
tools makes it easier to fix the design of GlobalAlias.
llvm-svn: 203705
This allows us to generate table lookups for code such as:
unsigned test(unsigned x) {
switch (x) {
case 100: return 0;
case 101: return 1;
case 103: return 2;
case 105: return 3;
case 107: return 4;
case 109: return 5;
case 110: return 6;
default: return f(x);
}
}
Since cases 102, 104, etc. are not constants, the lookup table has holes
in those positions. We therefore guard the table lookup with a bitmask check.
Patch by Jasper Neumann!
llvm-svn: 203694
After r203553 overflow intrinsics and their non-intrinsic (normal)
instruction get hashed to the same value. This patch prevents PRE from
moving an instruction into a predecessor block, and trying to add a phi
node that gets two different types (the intrinsic result and the
non-intrinsic result), resulting in a failing assert.
llvm-svn: 203574
The syntax for "cmpxchg" should now look something like:
cmpxchg i32* %addr, i32 42, i32 3 acquire monotonic
where the second ordering argument gives the required semantics in the case
that no exchange takes place. It should be no stronger than the first ordering
constraint and cannot be either "release" or "acq_rel" (since no store will
have taken place).
rdar://problem/15996804
llvm-svn: 203559
When an overflow intrinsic is followed by a non-overflow instruction,
replace the latter with an extract. For example:
%sadd = tail call { i32, i1 } @llvm.sadd.with.overflow.i32(i32 %a, i32 %b)
%sadd3 = add i32 %a, %b
Here the add statement will be replaced by an extract.
When an overflow intrinsic follows a non-overflow instruction, a clone
of the intrinsic is inserted before the normal instruction, which makes
it the same as the previous case. Subsequent runs of GVN can then clean
up the duplicate instructions and insert the extract.
This fixes PR8817.
llvm-svn: 203553
Summary:
When the sample profiles include discriminator information,
use the discriminator values to distinguish instruction weights
in different basic blocks.
This modifies the BodySamples mapping to map <line, discriminator> pairs
to weights. Instructions on the same line but different blocks, will
use different discriminator values. This, in turn, means that the blocks
may have different weights.
Other changes in this patch:
- Add tests for positive values of line offset, discriminator and samples.
- Change data types from uint32_t to unsigned and int and do additional
validation.
Reviewers: chandlerc
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2857
llvm-svn: 203508
optimize a call to a llvm intrinsic to something that invovles a call to a C
library call, make sure it sets the right calling convention on the call.
e.g.
extern double pow(double, double);
double t(double x) {
return pow(10, x);
}
Compiles to something like this for AAPCS-VFP:
define arm_aapcs_vfpcc double @t(double %x) #0 {
entry:
%0 = call double @llvm.pow.f64(double 1.000000e+01, double %x)
ret double %0
}
declare double @llvm.pow.f64(double, double) #1
Simplify libcall (part of instcombine) will turn the above into:
define arm_aapcs_vfpcc double @t(double %x) #0 {
entry:
%__exp10 = call double @__exp10(double %x) #1
ret double %__exp10
}
declare double @__exp10(double)
The pre-instcombine code works because calls to LLVM builtins are special.
Instruction selection will chose the right calling convention for the call.
However, the code after instcombine is wrong. The call to __exp10 will use
the C calling convention.
I can think of 3 options to fix this.
1. Make "C" calling convention just work since the target should know what CC
is being used.
This doesn't work because each function can use different CC with the "pcs"
attribute.
2. Have Clang add the right CC keyword on the calls to LLVM builtin.
This will work but it doesn't match the LLVM IR specification which states
these are "Standard C Library Intrinsics".
3. Fix simplify libcall so the resulting calls to the C routines will have the
proper CC keyword. e.g.
%__exp10 = call arm_aapcs_vfpcc double @__exp10(double %x) #1
This works and is the solution I implemented here.
Both solutions #2 and #3 would work. After carefully considering the pros and
cons, I decided to implement #3 for the following reasons.
1. It doesn't change the "spec" of the intrinsics.
2. It's a self-contained fix.
There are a couple of potential downsides.
1. There could be other places in the optimizer that is broken in the same way
that's not addressed by this.
2. There could be other calling conventions that need to be propagated by
simplify-libcall that's not handled.
But for now, this is the fix that I'm most comfortable with.
llvm-svn: 203488
The grammar for LLVM IR is not well specified in any document but seems
to obey the following rules:
- Attributes which have parenthesized arguments are never preceded by
commas. This form of attribute is the only one which ever has
optional arguments. However, not all of these attributes support
optional arguments: 'thread_local' supports an optional argument but
'addrspace' does not. Interestingly, 'addrspace' is documented as
being a "qualifier". What constitutes a qualifier? I cannot find a
definition.
- Some attributes use a space between the keyword and the value.
Examples of this form are 'align' and 'section'. These are always
preceded by a comma.
- Otherwise, the attribute has no argument. These attributes do not
have a preceding comma.
Sometimes an attribute goes before the instruction, between the
instruction and it's type, or after it's type. 'atomicrmw' has
'volatile' between the instruction and the type while 'call' has 'tail'
preceding the instruction.
With all this in mind, it seems most consistent for 'inalloca' on an
'inalloca' instruction to occur before between the instruction and the
type. Unlike the current formulation, there would be no preceding
comma. The combination 'alloca inalloca' doesn't look particularly
appetizing, perhaps a better spelling of 'inalloca' is down the road.
llvm-svn: 203376
This helps the instruction selector to lower an i64 * i64 -> i128
multiplication into a single instruction on targets which support it.
Patch by Manuel Jacob.
llvm-svn: 203230
Sequences of insertelement/extractelements are sometimes used to build
vectorsr; this code tries to put them back together into shuffles, but
could only produce a completely uniform shuffle types (<N x T> from two
<N x T> sources).
This should allow shuffles with different numbers of elements on the
input and output sides as well.
llvm-svn: 203229
are operations that do not access memory but may be sensitive
to floating-point environment changes. LLVM does not attempt
to model FP environment changes, so this was unnecessarily conservative
and was getting on the way of some optimizations, in particular
SLP vectorization.
llvm-svn: 203037
DWARF discriminators are used to distinguish multiple control flow paths
on the same source location. When this happens, instructions across
basic block boundaries will share the same debug location.
This pass detects this situation and creates a new lexical scope to one
of the two instructions. This lexical scope is a child scope of the
original and contains a new discriminator value. This discriminator is
then picked up from MCObjectStreamer::EmitDwarfLocDirective to be
written on the object file.
This fixes http://llvm.org/bugs/show_bug.cgi?id=18270.
llvm-svn: 202752
and update everything accordingly. This can be used to conditionalize
the amount of output in the backend based on the amount of debug
requested/metadata emission scheme by a front end (e.g. clang).
Paired with a commit to clang.
llvm-svn: 202332
We should apply fastcc whenever profitable. We can expand this list,
but there are lots of conventions with performance implications that we
don't want to change.
Differential Revision: http://llvm-reviews.chandlerc.com/D2705
llvm-svn: 202293
the default.
Based on the patch by Matt Arsenault, D1764!
I switched one place to use the more direct pointer type to compute the
desired address space, and I reworked the memcpy rewriting section to
reflect significant refactorings that this patch helped inspire.
Thanks to several of the folks who helped review and improve the patch
as well.
llvm-svn: 202247
to work independently for the slice side and the other side.
This allows us to only compute the minimum of the two when we actually
rewrite to a memcpy that needs to take the minimum, and preserve higher
alignment for one side or the other when rewriting to loads and stores.
This fix was inspired by seeing the result of some refactoring that
makes addrspace handling better.
llvm-svn: 202242
checking in SROA.
The primary change is to just rely on uge for checking that the offset
is within the allocation size. This removes the explicit checks against
isNegative which were terribly error prone (including the reversed logic
that led to PR18615) and prevented us from supporting stack allocations
larger than half the address space.... Ok, so maybe the latter isn't
*common* but it's a silly restriction to have.
Also, we used to try to support a PHI node which loaded from before the
start of the allocation if any of the loaded bytes were within the
allocation. This doesn't make any sense, we have never really supported
loading or storing *before* the allocation starts. The simplified logic
just doesn't care.
We continue to allow loading past the end of the allocation in part to
support cases where there is a PHI and some loads are larger than others
and the larger ones reach past the end of the allocation. We could solve
this a different and more conservative way, but I'm still somewhat
paranoid about this.
llvm-svn: 202224
ordering.
The fundamental problem that we're hitting here is that the use-def
chain ordering is *itself* not a stable thing to be relying on in the
rewriting for SROA. Further, we use a non-stable sort over the slices to
arrange them based on the section of the alloca they're operating on.
With a debugging STL implementation (or different implementations in
stage2 and stage3) this can cause stage2 != stage3.
The specific aspect of this problem fixed in this commit deals with the
rewriting and load-speculation around PHIs and Selects. This, like many
other aspects of the use-rewriting in SROA, is really part of the
"strong SSA-formation" that is doen by SROA where it works very hard to
canonicalize loads and stores in *just* the right way to satisfy the
needs of mem2reg[1]. When we have a select (or a PHI) with 2 uses of the
same alloca, we test that loads downstream of the select are
speculatable around it twice. If only one of the operands to the select
needs to be rewritten, then if we get lucky we rewrite that one first
and the select is immediately speculatable. This can cause the order of
operand visitation, and thus the order of slices to be rewritten, to
change an alloca from promotable to non-promotable and vice versa.
The fix is to defer all of the speculation until *after* the rewrite
phase is done. Once we've rewritten everything, we can accurately test
for whether speculation will work (once, instead of twice!) and the
order ceases to matter.
This also happens to simplify the other subtlety of speculation -- we
need to *not* speculate anything unless the result of speculating will
make the alloca fully promotable by mem2reg. I had a previous attempt at
simplifying this, but it was still pretty horrible.
There is actually already a *really* nice test case for this in
basictest.ll, but on multiple STL implementations and inputs, we just
got "lucky". Fortunately, the test case is very small and we can
essentially build it in exactly the opposite way to get reasonable
coverage in both directions even from normal STL implementations.
llvm-svn: 202092
On x86, shifting a vector by a scalar is significantly cheaper than shifting a
vector by another fully general vector. Unfortunately, because SelectionDAG
operates on just one basic block at a time, the shufflevector instruction that
reveals whether the right-hand side of a shift *is* really a scalar is often
not visible to CodeGen when it's needed.
This adds another handler to CodeGenPrepare, to sink any useful shufflevector
instructions down to the basic block where they're used, predicated on a target
hook (since on other architectures, doing so will often just introduce extra
real work).
rdar://problem/16063505
llvm-svn: 201655
During LSR of one loop we can run into a situation where we have to expand the
start of a recurrence of a loop induction variable in this loop. This start
value is a value derived of the induction variable of a preceeding loop. SCEV
has cannonicalized this value to a different recurrence than the recurrence of
the preceeding loop's induction variable (the type and/or step direction) has
changed). When we come to instantiate this SCEV we created a second induction
variable in this preceeding loop. This patch tries to base such derived
induction variables of the preceeding loop's induction variable.
This helps twolf on arm and seems to help scimark2 on x86.
Reapply with a fix for the case of a value derived from a pointer.
radar://15970709
llvm-svn: 201496
During LSR of one loop we can run into a situation where we have to expand the
start of a recurrence of a loop induction variable in this loop. This start
value is a value derived of the induction variable of a preceeding loop. SCEV
has cannonicalized this value to a different recurrence than the recurrence of
the preceeding loop's induction variable (the type and/or step direction) has
changed). When we come to instantiate this SCEV we created a second induction
variable in this preceeding loop. This patch tries to base such derived
induction variables of the preceeding loop's induction variable.
This helps twolf on arm and seems to help scimark2 on x86.
radar://15970709
llvm-svn: 201465
Summary:
AsmPrinter::EmitInlineAsm() will no longer use the EmitRawText() call for
targets with mature MC support. Such targets will always parse the inline
assembly (even when emitting assembly). Targets without mature MC support
continue to use EmitRawText() for assembly output.
The hasRawTextSupport() check in AsmPrinter::EmitInlineAsm() has been replaced
with MCAsmInfo::UseIntegratedAs which when true, causes the integrated assembler
to parse inline assembly (even when emitting assembly output). UseIntegratedAs
is set to true for targets that consider any failure to parse valid assembly
to be a bug. Target specific subclasses generally enable the integrated
assembler in their constructor. The default value can be overridden with
-no-integrated-as.
All tests that rely on inline assembly supporting invalid assembly (for example,
those that use mnemonics such as 'foo' or 'hello world') have been updated to
disable the integrated assembler.
Changes since review (and last commit attempt):
- Fixed test failures that were missed due to configuration of local build.
(fixes crash.ll and a couple others).
- Fixed tests that happened to pass because the local build was on X86
(should fix 2007-12-17-InvokeAsm.ll)
- mature-mc-support.ll's should no longer require all targets to be compiled.
(should fix ARM and PPC buildbots)
- Object output (-filetype=obj and similar) now forces the integrated assembler
to be enabled regardless of default setting or -no-integrated-as.
(should fix SystemZ buildbots)
Reviewers: rafael
Reviewed By: rafael
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2686
llvm-svn: 201333
As defined in LangRef, aliases do not have sections. However, LLVM's
GlobalAlias class inherits from GlobalValue, which means we can read and
set its section. We should probably ban that as a separate change,
since it doesn't make much sense for an alias to have a section that
differs from its aliasee.
Fixes PR18757, where the section was being lost on the global in code
from Clang like:
extern "C" {
__attribute__((used, section("CUSTOM"))) static int in_custom_section;
}
Reviewers: rafael.espindola
Differential Revision: http://llvm-reviews.chandlerc.com/D2758
llvm-svn: 201286
logical operations on the i1's driving them. This is a bad idea for every
target I can think of (confirmed with micro tests on all of: x86-64, ARM,
AArch64, Mips, and PowerPC) because it forces the i1 to be materialized into
a general purpose register, whereas consuming it directly into a select generally
allows it to exist only transiently in a predicate or flags register.
Chandler ran a set of performance tests with this change, and reported no
measurable change on x86-64.
llvm-svn: 201275
Summary:
AsmPrinter::EmitInlineAsm() will no longer use the EmitRawText() call for targets with mature MC support. Such targets will always parse the inline assembly (even when emitting assembly). Targets without mature MC support continue to use EmitRawText() for assembly output.
The hasRawTextSupport() check in AsmPrinter::EmitInlineAsm() has been replaced with MCAsmInfo::UseIntegratedAs which when true, causes the integrated assembler to parse inline assembly (even when emitting assembly output). UseIntegratedAs is set to true for targets that consider any failure to parse valid assembly to be a bug. Target specific subclasses generally enable the integrated assembler in their constructor. The default value can be overridden with -no-integrated-as.
All tests that rely on inline assembly supporting invalid assembly (for example, those that use mnemonics such as 'foo' or 'hello world') have been updated to disable the integrated assembler.
Reviewers: rafael
Reviewed By: rafael
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2686
llvm-svn: 201237
Fixes PR18753 and PR18782.
This is necessary for LICM to preserve LCSSA correctly and efficiently.
There is still some active discussion about whether we should be using
LCSSA, but we can't just immediately stop using it and we *need* LICM to
preserve it while we are using it. We can restore the old SSAUpdater
driven code if and when there is a serious effort to remove the reliance
on LCSSA from all of the loop passes.
However, this also serves as a great example of why LCSSA is very nice
to have. This change significantly simplifies the process of sinking
instructions for LICM, and makes it quite a bit less expensive.
It wouldn't even be as complex as it is except that I had to start the
process of removing the big recursive LCSSA formation hammer in order to
switch even this much of the re-forming code to asserting that LCSSA was
preserved. I'll fully remove that next just to tidy things up until the
LCSSA debate settles one way or the other.
llvm-svn: 201148
Before conditional store vectorization/unrolling we had only one
vectorized/unrolled basic block. After adding support for conditional store
vectorization this will not only be one block but multiple basic blocks. The
last block would have the back-edge. I updated the code to use a vector of basic
blocks instead of a single basic block and fixed the users to use the last entry
in this vector. But, I forgot to add the basic blocks to this vector!
Fixes PR18724.
llvm-svn: 201028
The bitcast instruction during constant materialization was not placed correcly
in the presence of phi nodes. This commit fixes the insertion point to be in the
idom instead.
This fixes PR18768
llvm-svn: 201009
225 is the default value of inline-threshold. This change will make sure
we have the same inlining behavior as prior to r200886.
As Chandler points out, even though we don't have code in our testing
suite that uses cold attribute, there are larger applications that do
use cold attribute.
r200886 + this commit intend to keep the same behavior as prior to r200886.
We can later on tune the inlinecold-threshold.
The main purpose of r200886 is to help performance of instrumentation based
PGO before we actually hook up inliner with analysis passes such as BPI and BFI.
For instrumentation based PGO, we try to increase inlining of hot functions and
reduce inlining of cold functions by setting inlinecold-threshold.
Another option suggested by Chandler is to use a boolean flag that controls
if we should use OptSizeThreshold for cold functions. The default value
of the boolean flag should not change the current behavior. But it gives us
less freedom in controlling inlining of cold functions.
llvm-svn: 200898
Added command line option inlinecold-threshold to set threshold for inlining
functions with cold attribute. Listen to the cold attribute when it would
decrease the inline threshold.
llvm-svn: 200886
Add the missing transformation strchr(p, 0) -> p + strlen(p) to SimplifyLibCalls
and remove the ToDo comment.
Reviewer: Duncan P.N. Exan Smith
llvm-svn: 200736
It disturbs the layout of the parameters in memory and registers,
leading to problems in the backend.
The plan for optimizing internal inalloca functions going forward is to
essentially SROA the argument memory and demote any captured arguments
(things that aren't trivially written by a load or store) to an indirect
pointer to a static alloca.
llvm-svn: 200717
LowerExpectIntrinsic previously only understood the idiom of an expect
intrinsic followed by a comparison with zero. For llvm.expect.i1, the
comparison would be stripped by the early-cse pass.
Patch by Daniel Micay.
llvm-svn: 200664
unrolling heuristic per default
Benchmarking on x86_64 (thanks Chandler!) and ARM has shown those options speed
up some benchmarks while not causing any interesting regressions.
llvm-svn: 200621
LCSSA when we promote to SSA registers inside of LICM.
Currently, this is actually necessary. The promotion logic in LICM uses
SSAUpdater which doesn't understand how to place LCSSA PHI nodes.
Teaching it to do so would be a very significant undertaking. It may be
worthwhile and I've left a FIXME about this in the code as well as
starting a thread on llvmdev to try to figure out the right long-term
solution.
For now, the PR needs to be fixed. Short of using the promition
SSAUpdater to place both the LCSSA PHI nodes and the promoted PHI nodes,
I don't see a cleaner or cheaper way of achieving this. Fortunately,
LCSSA is relatively lazy and sparse -- it should only update
instructions which need it. We can also skip the recursive variant when
we don't promote to SSA values.
llvm-svn: 200612
cost so that they don't impact the vector bonus. Fundamentally, counting
unsimplified instructions is just *wrong*; it will continue to introduce
instability as things which do not generate code bizarrely impact
inlining. For example, sufficiently nested inlined functions could turn
off the vector bonus with lifetime markers just like the debug
intrinsics do. =/
This is a short-term tactical fix. Long term, I think we need to remove
the vector bonus entirely. That's a separate patch and discussion
though.
The patch to fix this provided by Dario Domizioli. I've added some
comments about the planned direction and used a heavily pruned form of
debug info intrinsics for the test case. While this debug info doesn't
work or "do" anything useful, it lets us easily test all manner of
interference easily, and I suspect this will not be the last time we
want to craft a pattern where debug info interferes with the inliner in
a problematic way.
llvm-svn: 200609
This reverts commit r200576. It broke 32-bit self-host builds by
vectorizing two calls to @llvm.bswap.i64, which we then fail to expand.
llvm-svn: 200602
transform accordingly. Based on similar code from Loop vectorization.
Subsequent commits will include vectorization of function calls to
vector intrinsics and form function calls to vector library calls.
Patch by Raul Silvera! (Much delayed due to my not running dcommit)
llvm-svn: 200576
loop vectorizer to not do so when runtime pointer checks are needed and
share code with the new (not yet enabled) load/store saturation runtime
unrolling. Also ensure that we only consider the runtime checks when the
loop hasn't already been vectorized. If it has, the runtime check cost
has already been paid.
I've fleshed out a test case to cover the scalar unrolling as well as
the vector unrolling and comment clearly why we are or aren't following
the pattern.
llvm-svn: 200530
preserve loop simplify of enclosing loops.
The problem here starts with LoopRotation which ends up cloning code out
of the latch into the new preheader it is buidling. This can create
a new edge from the preheader into the exit block of the loop which
breaks LoopSimplify form. The code tries to fix this by splitting the
critical edge between the latch and the exit block to get a new exit
block that only the latch dominates. This sadly isn't sufficient.
The exit block may be an exit block for multiple nested loops. When we
clone an edge from the latch of the inner loop to the new preheader
being built in the outer loop, we create an exiting edge from the outer
loop to this exit block. Despite breaking the LoopSimplify form for the
inner loop, this is fine for the outer loop. However, when we split the
edge from the inner loop to the exit block, we create a new block which
is in neither the inner nor outer loop as the new exit block. This is
a predecessor to the old exit block, and so the split itself takes the
outer loop out of LoopSimplify form. We need to split every edge
entering the exit block from inside a loop nested more deeply than the
exit block in order to preserve all of the loop simplify constraints.
Once we try to do that, a problem with splitting critical edges
surfaces. Previously, we tried a very brute force to update LoopSimplify
form by re-computing it for all exit blocks. We don't need to do this,
and doing this much will sometimes but not always overlap with the
LoopRotate bug fix. Instead, the code needs to specifically handle the
cases which can start to violate LoopSimplify -- they aren't that
common. We need to see if the destination of the split edge was a loop
exit block in simplified form for the loop of the source of the edge.
For this to be true, all the predecessors need to be in the exact same
loop as the source of the edge being split. If the dest block was
originally in this form, we have to split all of the deges back into
this loop to recover it. The old mechanism of doing this was
conservatively correct because at least *one* of the exiting blocks it
rewrote was the DestBB and so the DestBB's predecessors were fixed. But
this is a much more targeted way of doing it. Making it targeted is
important, because ballooning the set of edges touched prevents
LoopRotate from being able to split edges *it* needs to split to
preserve loop simplify in a coherent way -- the critical edge splitting
would sometimes find the other edges in need of splitting but not
others.
Many, *many* thanks for help from Nick reducing these test cases
mightily. And helping lots with the analysis here as this one was quite
tricky to track down.
llvm-svn: 200393
because of the inside-out run of LoopSimplify in the LoopPassManager and
the fact that LoopSimplify couldn't be "preserved" across two
independent LoopPassManagers.
Anyways, in that case, IndVars wasn't correctly preserving an LCSSA PHI
node because it thought it was rewriting (via SCEV) the incoming value
to a loop invariant value. While it may well be invariant for the
current loop, it may be rewritten in terms of an enclosing loop's
values. This in and of itself is fine, as the LCSSA PHI node in the
enclosing loop for the inner loop value we're rewriting will have its
own LCSSA PHI node if used outside of the enclosing loop. With me so
far?
Well, the current loop and the enclosing loop may share an exiting
block and exit block, and when they do they also share LCSSA PHI nodes.
In this case, its not valid to RAUW through the LCSSA PHI node.
Expected crazy test included.
llvm-svn: 200372
When simplifycfg moves an instruction, it must drop metadata it doesn't know
is still valid with the preconditions changes. In particular, it must drop
the range and tbaa metadata.
The patch implements this with an utility function to drop all metadata not
in a white list.
llvm-svn: 200322
vectorizer, placing it behind an off-by-default flag.
It turns out that block frequency isn't what we want at all, here or
elsewhere. This has been I think a nagging feeling for several of us
working with it, but Arnold has given some really nice simple examples
where the results are so comprehensively wrong that they aren't useful.
I'm planning to email the dev list with a summary of why its not really
useful and a couple of ideas about how to better structure these types
of heuristics.
llvm-svn: 200294
Summary:
I searched Transforms/ and Analysis/ for 'ByVal' and updated those call
sites to check for inalloca if appropriate.
I added tests for any change that would allow an optimization to fire on
inalloca.
Reviewers: nlewycky
Differential Revision: http://llvm-reviews.chandlerc.com/D2449
llvm-svn: 200281
The vectorizer takes a loop like this and widens all instructions except for the
store. The stores are scalarized/unrolled and hidden behind an "if" block.
for (i = 0; i < 128; ++i) {
if (a[i] < 10)
a[i] += val;
}
for (i = 0; i < 128; i+=2) {
v = a[i:i+1];
v0 = (extract v, 0) + 10;
v1 = (extract v, 1) + 10;
if (v0 < 10)
a[i] = v0;
if (v1 < 10)
a[i] = v1;
}
The vectorizer relies on subsequent optimizations to sink instructions into the
conditional block where they are anticipated.
The flag "vectorize-num-stores-pred" controls whether and how many stores to
handle this way. Vectorization of conditional stores is disabled per default for
now.
This patch also adds a change to the heuristic when the flag
"enable-loadstore-runtime-unroll" is enabled (off by default). It unrolls small
loops until load/store ports are saturated. This heuristic uses TTI's
getMaxUnrollFactor as a measure for load/store ports.
I also added a second flag -enable-cond-stores-vec. It will enable vectorization
of conditional stores. But there is no cost model for vectorization of
conditional stores in place yet so this will not do good at the moment.
rdar://15892953
Results for x86-64 -O3 -mavx +/- -mllvm -enable-loadstore-runtime-unroll
-vectorize-num-stores-pred=1 (before the BFI change):
Performance Regressions:
Benchmarks/Ptrdist/yacr2/yacr2 7.35% (maze3() is identical but 10% slower)
Applications/siod/siod 2.18%
Performance improvements:
mesa -4.42%
libquantum -4.15%
With a patch that slightly changes the register heuristics (by subtracting the
induction variable on both sides of the register pressure equation, as the
induction variable is probably not really unrolled):
Performance Regressions:
Benchmarks/Ptrdist/yacr2/yacr2 7.73%
Applications/siod/siod 1.97%
Performance Improvements:
libquantum -13.05% (we now also unroll quantum_toffoli)
mesa -4.27%
llvm-svn: 200270
uint32.
When folding branches to common destination, the updated branch weights
can exceed uint32 by more than factor of 2. We should keep halving the
weights until they can fit into uint32.
llvm-svn: 200262
cold loops as-if they were being optimized for size.
Nothing fancy here. Simply test case included. The nice thing is that we
can now incrementally build on top of this to drive other heuristics.
All of the infrastructure work is done to get the profile information
into this layer.
The remaining work necessary to make this a fully general purpose loop
unroller for very hot loops is to make it a fully general purpose loop
unroller. Things I know of but am not going to have time to benchmark
and fix in the immediate future:
1) Don't disable the entire pass when the target is lacking vector
registers. This really doesn't make any sense any more.
2) Teach the unroller at least and the vectorizer potentially to handle
non-if-converted loops. This is trivial for the unroller but hard for
the vectorizer.
3) Compute the relative hotness of the loop and thread that down to the
various places that make cost tradeoffs (very likely only the
unroller makes sense here, and then only when dealing with loops that
are small enough for unrolling to not completely blow out the LSD).
I'm still dubious how useful hotness information will be. So far, my
experiments show that if we can get the correct logic for determining
when unrolling actually helps performance, the code size impact is
completely unimportant and we can unroll in all cases. But at least
we'll no longer burn code size on cold code.
One somewhat unrelated idea that I've had forever but not had time to
implement: mark all functions which are only reachable via the global
constructors rigging in the module as optsize. This would also decrease
the impact of any more aggressive heuristics here on code size.
llvm-svn: 200219
powers of two. This is essentially always the correct thing given the
impact on alignment, scaling factors that can be used in addressing
modes, etc. Also, fix the management of the unroll vs. small loop cost
to more accurately model things with this world.
Enhance a test case to actually exercise more of the unroll machinery if
using synthetic constants rather than a specific target model. Before
this change, with the added flags this test will unroll 3 times instead
of either 2 or 4 (the two sensible answers).
While I don't expect this to make a huge difference, if there are lots
of loops sitting right on the edge of hitting the 'small unroll' factor,
they might change behavior. However, I've benchmarked moving the small
loop cost up and down in many various ways and by a huge factor (2x)
without seeing more than 0.2% code size growth. Small adjustments such
as the series that led up here have led to about 1% improvement on some
benchmarks, but it is very close to the noise floor so I mostly checked
that nothing regressed. Let me know if you see bad behavior on other
targets but I don't expect this to be a sufficiently dramatic change to
trigger anything.
llvm-svn: 200213
the loops in a function, and teach LICM to work in the presance of
LCSSA.
Previously, LCSSA was a loop pass. That made passes requiring it also be
loop passes and unable to depend on function analysis passes easily. It
also caused outer loops to have a different "canonical" form from inner
loops during analysis. Instead, we go into LCSSA form and preserve it
through the loop pass manager run.
Note that this has the same problem as LoopSimplify that prevents
enabling its verification -- loop passes which run at the end of the loop
pass manager and don't preserve these are valid, but the subsequent loop
pass runs of outer loops that do preserve this pass trigger too much
verification and fail because the inner loop no longer verifies.
The other problem this exposed is that LICM was completely unable to
handle LCSSA form. It didn't preserve it and it actually would give up
on moving instructions in many cases when they were used by an LCSSA phi
node. I've taught LICM to support detecting LCSSA-form PHI nodes and to
hoist and sink around them. This may actually let LICM fire
significantly more because we put everything into LCSSA form to rotate
the loop before running LICM. =/ Now LICM should handle that fine and
preserve it correctly. The down side is that LICM has to require LCSSA
in order to preserve it. This is just a fact of life for LCSSA. It's
entirely possible we should completely remove LCSSA from the optimizer.
The test updates are essentially accomodating LCSSA phi nodes in the
output of LICM, and the fact that we now completely sink every
instruction in ashr-crash below the loop bodies prior to unrolling.
With this change, LCSSA is computed only three times in the pass
pipeline. One of them could be removed (and potentially a SCEV run and
a separate LoopPassManager entirely!) if we had a LoopPass variant of
InstCombine that ran InstCombine on the loop body but refused to combine
away LCSSA PHI nodes. Currently, this also prevents loop unrolling from
being in the same loop pass manager is rotate, LICM, and unswitch.
There is one thing that I *really* don't like -- preserving LCSSA in
LICM is quite expensive. We end up having to re-run LCSSA twice for some
loops after LICM runs because LICM can undo LCSSA both in the current
loop and the parent loop. I don't really see good solutions to this
other than to completely move away from LCSSA and using tools like
SSAUpdater instead.
llvm-svn: 200067
Argument promotion can replace an argument of a call with an alloca. This
requires clearing the tail marker as it is very likely that the callee is now
using an alloca in the caller.
This fixes pr14710.
llvm-svn: 199909
function and a FunctionPass.
This has many benefits. The motivating use case was to be able to
compute function analysis passes *after* running LoopSimplify (to avoid
invalidating them) and then to run other passes which require
LoopSimplify. Specifically passes like unrolling and vectorization are
critical to wire up to BranchProbabilityInfo and BlockFrequencyInfo so
that they can be profile aware. For the LoopVectorize pass the only
things in the way are LoopSimplify and LCSSA. This fixes LoopSimplify
and LCSSA is next on my list.
There are also a bunch of other benefits of doing this:
- It is now very feasible to make more passes *preserve* LoopSimplify
because they can simply run it after changing a loop. Because
subsequence passes can assume LoopSimplify is preserved we can reduce
the runs of this pass to the times when we actually mutate a loop
structure.
- The new pass manager should be able to more easily support loop passes
factored in this way.
- We can at long, long last observe that LoopSimplify is preserved
across SCEV. This *halves* the number of times we run LoopSimplify!!!
Now, getting here wasn't trivial. First off, the interfaces used by
LoopSimplify are all over the map regarding how analysis are updated. We
end up with weird "pass" parameters as a consequence. I'll try to clean
at least some of this up later -- I'll have to have it all clean for the
new pass manager.
Next up I discovered a really frustrating bug. LoopUnroll *claims* to
preserve LoopSimplify. That's actually a lie. But the way the
LoopPassManager ends up running the passes, it always ran LoopSimplify
on the unrolled-into loop, rectifying this oversight before any
verification could kick in and point out that in fact nothing was
preserved. So I've added code to the unroller to *actually* simplify the
surrounding loop when it succeeds at unrolling.
The only functional change in the test suite is that we now catch a case
that was previously missed because SCEV and other loop transforms see
their containing loops as simplified and thus don't miss some
opportunities. One test case has been converted to check that we catch
this case rather than checking that we miss it but at least don't get
the wrong answer.
Note that I have #if-ed out all of the verification logic in
LoopSimplify! This is a temporary workaround while extracting these bits
from the LoopPassManager. Currently, there is no way to have a pass in
the LoopPassManager which preserves LoopSimplify along with one which
does not. The LPM will try to verify on each loop in the nest that
LoopSimplify holds but the now-Function-pass cannot distinguish what
loop is being verified and so must try to verify all of them. The inner
most loop is clearly no longer simplified as there is a pass which
didn't even *attempt* to preserve it. =/ Once I get LCSSA out (and maybe
LoopVectorize and some other fixes) I'll be able to re-enable this check
and catch any places where we are still failing to preserve
LoopSimplify. If this causes problems I can back this out and try to
commit *all* of this at once, but so far this seems to work and allow
much more incremental progress.
llvm-svn: 199884
This logic hadn't been updated to handle FastMathFlags, and it took me a while to detect it because it doesn't show up in a simple search for CreateFAdd.
llvm-svn: 199629
intrinsics.
Reported on the list by Evan with a couple of attempts to fix, but it
took a while to dig down to the root cause. There are two overlapping
bugs here, both centering around the circumstance of discovering
a memcpy operand which is known to be completely outside the bounds of
the alloca.
First, we need to kill the *other* side of the memcpy if it was added to
this alloca. Otherwise we'll factor it into our slicing and try to
rewrite it even though we know for a fact that it is dead. This is made
more tricky because we can visit the sides in either order. So we have
to both kill the other side and skip instructions marked as dead. The
latter really should be goodness in every case, but here is a matter of
correctness.
Second, we need to actually remove the *uses* of the alloca by the
memcpy when queuing it for later deletion. Otherwise it may still be
using the alloca when we go to promote it (if the rewrite re-uses the
existing alloca instruction). Do this by factoring out the
use-clobbering used when for nixing a Phi argument and re-using it
across the operands of a to-be-deleted instruction.
llvm-svn: 199590
a reduction.
Really. Under certain circumstances (the use list of an instruction has to be
set up right - hence the extra pass in the test case) we would not recognize
when a value in a potential reduction cycle was used multiple times by the
reduction cycle.
Fixes PR18526.
radar://15851149
llvm-svn: 199570
There has been an old FIXME to find the right cut-off for when it's worth
analyzing and potentially transforming a switch to a lookup table.
The switches always have two or more cases. I could not measure any speed-up
by transforming a switch with two cases. A switch with three cases gets a nice
speed-up, and I couldn't measure any compile-time regression, so I think this
is the right threshold.
In a Clang self-host, this causes 480 new switches to be transformed,
and reduces the final binary size with 8 KB.
llvm-svn: 199294
case when the lookup table doesn't have any holes.
This means we can build a lookup table for switches like this:
switch (x) {
case 0: return 1;
case 1: return 2;
case 2: return 3;
case 3: return 4;
default: exit(1);
}
The default case doesn't yield a constant result here, but that doesn't matter,
since a default result is only necessary for filling holes in the lookup table,
and this table doesn't have any holes.
This makes us transform 505 more switches in a clang bootstrap, and shaves 164 KB
off the resulting clang binary.
llvm-svn: 199025
1- Use the line_iterator class to read profile files.
2- Allow comments in profile file. Lines starting with '#'
are completely ignored while reading the profile.
3- Add parsing support for discriminators and indirect call samples.
Our external profiler can emit more profile information that we are
currently not handling. This patch does not add new functionality to
support this information, but it allows profile files to provide it.
I will add actual support later on (for at least one of these
features, I need support for DWARF discriminators in Clang).
A sample line may contain the following additional information:
Discriminator. This is used if the sampled program was compiled with
DWARF discriminator support
(http://wiki.dwarfstd.org/index.php?title=Path_Discriminators). This
is currently only emitted by GCC and we just ignore it.
Potential call targets and samples. If present, this line contains a
call instruction. This models both direct and indirect calls. Each
called target is listed together with the number of samples. For
example,
130: 7 foo:3 bar:2 baz:7
The above means that at relative line offset 130 there is a call
instruction that calls one of foo(), bar() and baz(). With baz()
being the relatively more frequent call target.
Differential Revision: http://llvm-reviews.chandlerc.com/D2355
4- Simplify format of profile input file.
This implements earlier suggestions to simplify the format of the
sample profile file. The symbol table is not necessary and function
profiles do not need to know the number of samples in advance.
Differential Revision: http://llvm-reviews.chandlerc.com/D2419
llvm-svn: 198973
This adds a propagation heuristic to convert instruction samples
into branch weights. It implements a similar heuristic to the one
implemented by Dehao Chen on GCC.
The propagation proceeds in 3 phases:
1- Assignment of block weights. All the basic blocks in the function
are initial assigned the same weight as their most frequently
executed instruction.
2- Creation of equivalence classes. Since samples may be missing from
blocks, we can fill in the gaps by setting the weights of all the
blocks in the same equivalence class to the same weight. To compute
the concept of equivalence, we use dominance and loop information.
Two blocks B1 and B2 are in the same equivalence class if B1
dominates B2, B2 post-dominates B1 and both are in the same loop.
3- Propagation of block weights into edges. This uses a simple
propagation heuristic. The following rules are applied to every
block B in the CFG:
- If B has a single predecessor/successor, then the weight
of that edge is the weight of the block.
- If all the edges are known except one, and the weight of the
block is already known, the weight of the unknown edge will
be the weight of the block minus the sum of all the known
edges. If the sum of all the known edges is larger than B's weight,
we set the unknown edge weight to zero.
- If there is a self-referential edge, and the weight of the block is
known, the weight for that edge is set to the weight of the block
minus the weight of the other incoming edges to that block (if
known).
Since this propagation is not guaranteed to finalize for every CFG, we
only allow it to proceed for a limited number of iterations (controlled
by -sample-profile-max-propagate-iterations). It currently uses the same
GCC default of 100.
Before propagation starts, the pass builds (for each block) a list of
unique predecessors and successors. This is necessary to handle
identical edges in multiway branches. Since we visit all blocks and all
edges of the CFG, it is cleaner to build these lists once at the start
of the pass.
Finally, the patch fixes the computation of relative line locations.
The profiler emits lines relative to the function header. To discover
it, we traverse the compilation unit looking for the subprogram
corresponding to the function. The line number of that subprogram is the
line where the function begins. That becomes line zero for all the
relative locations.
llvm-svn: 198972
for (i = 0; i < N; ++i)
A[i * Stride1] += B[i * Stride2];
We take loops like this and check that the symbolic strides 'Strided1/2' are one
and drop to the scalar loop if they are not.
This is currently disabled by default and hidden behind the flag
'enable-mem-access-versioning'.
radar://13075509
llvm-svn: 198950
This doesn't seem to have actually broken anything. It was paranoia
on my part. Trying again now that bots are more stable.
This is a follow up of the r198338 commit that added truncates for
lcssa phi nodes. Sinking the truncates below the phis cleans up the
loop and simplifies subsequent analysis within the indvars pass.
llvm-svn: 198678
This is a follow up of the r198338 commit that added truncates for
lcssa phi nodes. Sinking the truncates below the phis cleans up the
loop and simplifies subsequent analysis within the indvars pass.
llvm-svn: 198654
Now with a fix for PR18384: ValueHandleBase::ValueIsDeleted.
We need to invalidate SCEV's loop info when we delete a block, even if no values are hoisted.
llvm-svn: 198631
This commit was the source of crasher PR18384:
While deleting: label %for.cond127
An asserting value handle still pointed to this value!
UNREACHABLE executed at llvm/lib/IR/Value.cpp:671!
Reverting to get the builders green, feel free to re-land after fixing up.
(Renato has a handy isolated repro if you need it.)
This reverts commit r198478.
llvm-svn: 198503
getSCEV for an ashr instruction creates an intermediate zext
expression when it truncates its operand.
The operand is initially inside the loop, so the narrow zext
expression has a non-loop-invariant loop disposition.
LoopSimplify then runs on an outer loop, hoists the ashr operand, and
properly invalidate the SCEVs that are mapped to value.
The SCEV expression for the ashr is now an AddRec with the hoisted
value as the now loop-invariant start value.
The LoopDisposition of this wide value was properly invalidated during
LoopSimplify.
However, if we later get the ashr SCEV again, we again try to create
the intermediate zext expression. We get the same SCEV that we did
earlier, and it is still cached because it was never mapped to a
Value. When we try to create a new AddRec we abort because we're using
the old non-loop-invariant LoopDisposition.
I don't have a solution for this other than to clear LoopDisposition
when LoopSimplify hoists things.
I think the long-term strategy should be to perform LoopSimplify on
all loops before computing SCEV and before running any loop opts on
individual loops. It's possible we may want to rerun LoopSimplify on
individual loops, but it should rarely do anything, so rarely require
invalidating SCEV.
llvm-svn: 198478
The loop rerolling pass was failing with an assertion failure from a
failed cast on loops like this:
void foo(int *A, int *B, int m, int n) {
for (int i = m; i < n; i+=4) {
A[i+0] = B[i+0] * 4;
A[i+1] = B[i+1] * 4;
A[i+2] = B[i+2] * 4;
A[i+3] = B[i+3] * 4;
}
}
The code was casting the SCEV-expanded code for the new
induction variable to a phi-node. When the loop had a non-constant
lower bound, the SCEV expander would end the code expansion with an
add insted of a phi node and the cast would fail.
It looks like the cast to a phi node was only needed to get the
induction variable value coming from the backedge to compute the end
of loop condition. This patch changes the loop reroller to compare
the induction variable to the number of times the backedge is taken
instead of the iteration count of the loop. In other words, we stop
the loop when the current value of the induction variable ==
IterationCount-1. Previously, the comparison was comparing the
induction variable value from the next iteration == IterationCount.
This problem only seems to occur on 32-bit targets. For some reason,
the loop is not rerolled on 64-bit targets.
PR18290
llvm-svn: 198425
cycles
This allows the value equality check to work even if we don't have a dominator
tree. Also add some more comments.
I was worried about compile time impacts and did not implement reachability but
used the dominance check in the initial patch. The trade-off was that the
dominator tree was required.
The llvm utility function isPotentiallyReachable cuts off the recursive search
after 32 visits. Testing did not show any compile time regressions showing my
worries unjustfied.
No compile time or performance regressions at O3 -flto -mavx on test-suite +
externals.
Addresses review comments from r198290.
llvm-svn: 198400
When widening an IV to remove s/zext, we generally try to eliminate
the original narrow IV. However, LCSSA phi nodes outside the loop were
still using the original IV. Clean this up more aggressively to avoid
redundancy in generated code.
llvm-svn: 198338
When there are cycles in the value graph we have to be careful interpreting
"Value*" identity as "value" equivalence. We interpret the value of a phi node
as the value of its operands.
When we check for value equivalence now we make sure that the "Value*" dominates
all cycles (phis).
%0 = phi [%noaliasval, %addr2]
%l = load %ptr
%addr1 = gep @a, 0, %l
%addr2 = gep @a, 0, (%l + 1)
store %ptr ...
Before this patch we would return NoAlias for (%0, %addr1) which is wrong
because the value of the load is from different iterations of the loop.
Tested on x86_64 -mavx at O3 and O3 -flto with no performance or compile time
regressions.
PR18068
radar://15653794
llvm-svn: 198290
widespread glibc bugs.
The glibc implementation of exp10 has a very serious precision bug in
version 2.15 (and older versions). This is still very widely used (the
current Ubuntu LTS for example uses it) and so it isn't reasonable to
make transforms that produce these functions. This fixes many
miscompiles introduced when we started transforming pow(10.0, ...) into
exp10, and it may have fixed other latent miscompiles where exp10
provided sufficient precision but exp10f did not.
This is all really horrible. The primary bug has been fixed for over
a year and glibc 2.18 works correctly for the test cases I have, but it
will be 2017 before the LTS using 2.15 is no longer supported by Ubuntu
(and thus reasonable for folks to be relying on). =[ We're either going
to need to live without these optimizations, or find a way to switch
behavior more dynamically than using simply the fact that the OS is
"Linux".
To make matters worse, there appears to be significant testing and
fixing of numerous other bugs in the exp10 family of functions right now
in glibc. While those haven't been causing problems I've seen in the
wild, it gives me concerns that we may need to wait until an even later
release of glibc before we can reliably transform code into exp10.
llvm-svn: 198093
Split sadd.with.overflow into add + sadd.with.overflow to allow
analysis and optimization. This should ideally be done after
InstCombine, which can perform code motion (eventually indvars should
run after all canonical instcombines). We want ISEL to recombine the
add and the check, at least on x86.
This is currently under an option for reducing live induction
variables: -liv-reduce. The next step is reducing liveness of IVs that
are live out of the overflow check paths. Once the related
optimizations are fully developed, reviewed and tested, I do expect
this to become default.
llvm-svn: 197926
If the Scalarizer scalarized a vector PHI but could not scalarize
all uses of it, it would insert a series of insertelements to reconstruct
the vector PHI value from the scalar ones. The problem was that it would
emit these insertelements immediately after the PHI, even if there were
other PHIs after it.
llvm-svn: 197909
If we happen to eliminate every case in a switch that has branch
weights, we currently try to create metadata for the one remaining
branch, triggering an assert. Instead, we need to check that the
metadata we're trying to create is sensible.
llvm-svn: 197791
A phi node operand or an instruction operand could be a constant expression that
can trap (division). Check that we don't vectorize such cases.
PR16729
radar://15653590
llvm-svn: 197449
through an invoke instruction.
The original patch for this was written by Mark Seaborn, but I've
reworked his test case into the existing returns_twice test case and
implemented the fix by the prior refactoring to actually run the cost
analysis over invoke instructions, and then here fixing our detection of
the returns_twice attribute to work for both calls and invokes. We never
noticed because we never saw an invoke. =[
llvm-svn: 197216
handles terminator instructions.
The inline cost analysis inheritted some pretty rough handling of
terminator insts from the original cost analysis, and then made it much,
much worse by factoring all of the important analyses into a separate
instruction visitor. That instruction visitor never visited the
terminator.
This works fine for things like conditional branches, but for many other
things we simply computed The Wrong Value. First example are
unconditional branches, which should be free but were counted as full
cost. This is most significant for conditional branches where the
condition simplifies and folds during inlining. We paid a 1 instruction
tax on every branch in a straight line specialized path. =[
Oh, we also claimed that the unreachable instruction had cost.
But it gets worse. Let's consider invoke. We never applied the call
penalty. We never accounted for the cost of the arguments. Nope. Worse
still, we didn't handle the *correctness* constraints of not inlining
recursive invokes, or exception throwing returns_twice functions. Oops.
See PR18206. Sadly, PR18206 requires yet another fix, but this
refactoring is at least a huge step in that direction.
llvm-svn: 197215
GlobalOpt's CleanupConstantGlobalUsers function uses a worklist array to manage
constant users to be visited. The pointers in this array need to be weak
handles because when we delete a constant array, we may also be holding a
pointer to one of its elements (or an element of one of its elements if we're
dealing with an array of arrays) in the worklist.
Fixes PR17347.
llvm-svn: 197178
This avoids creating branch weight metadata of length one when we fold
cases into the default of a switch instruction, which was triggering
an assert.
llvm-svn: 196845
Before this change, inlining one "invoke" into an outer "invoke" call
site can lead to the outer landingpad's catch/filter clauses being
copied multiple times into the resulting landingpad. This happens:
* when the inlined function contains multiple "resume" instructions,
because forwardResume() copies the clauses but is called multiple
times;
* when the inlined function contains a "resume" and a "call", because
HandleCallsInBlockInlinedThroughInvoke() copies the clauses but is
redundant with forwardResume().
Fix this by deduplicating the code.
This problem doesn't lead to any incorrect execution; it's only
untidy.
This change will make fixing PR17872 a little easier.
llvm-svn: 196710
The intended behaviour is to force vectorization on the presence
of the flag (either turn on or off), and to continue the behaviour
as expected in its absence. Tests were added to make sure the all
cases are covered in opt. No tests were added in other tools with
the assumption that they should use the PassManagerBuilder in the
same way.
This patch also removes the outdated -late-vectorize flag, which was
on by default and not helping much.
The pragma metadata is being attached to the same place as other loop
metadata, but nothing forbids one from attaching it to a function
(to enable #pragma optimize) or basic blocks (to hint the basic-block
vectorizers), etc. The logic should be the same all around.
Patches to Clang to produce the metadata will be produced after the
initial implementation is agreed upon and committed. Patches to other
vectorizers (such as SLP and BB) will be added once we're happy with
the pass manager changes.
llvm-svn: 196537
We were creating external uses for scalar values in MustGather entries that also
had a ScalarToTreeEntry (they also are present in a vectorized tuple). This
meant we would keep a value 'alive' as a scalar and vectorized causing havoc.
This is not necessary because when we create a MustGather vector we explicitly
create external uses entries for the insertelement instructions of the
MustGather vector elements.
Fixes PR18129.
radar://15582184
llvm-svn: 196508
clang enables vectorization at optimization levels > 1 and size level < 2. opt
should behave similarily.
Loop vectorization and SLP vectorization can be disabled with the flags
-disable-(loop/slp)-vectorization.
llvm-svn: 196294
The profile file parser needed some tests for its parsing actions.
This adds tests for each of the error messages emitted by the parser.
llvm-svn: 196106
In signed arithmetic we could end up with an i64 trip count for an i32 phi.
Because it is signed arithmetic we know that this is only defined if the i32
does not wrap. It is therefore safe to truncate the i64 trip count to a i32
value.
Fixes PR18049.
llvm-svn: 195787
we generate PHI nodes with multiple entries from the same basic block but
with different values. Enabling CSE on ExtractElement instructions make sure
that all of the RAUWed instructions are the same.
llvm-svn: 195773
Short description.
This issue is about case of treating pointers as integers.
We treat pointers as different if they references different address space.
At the same time, we treat pointers equal to integers (with machine address
width). It was a point of false-positive. Consider next case on 32bit machine:
void foo0(i32 addrespace(1)* %p)
void foo1(i32 addrespace(2)* %p)
void foo2(i32 %p)
foo0 != foo1, while
foo1 == foo2 and foo0 == foo2.
As you can see it breaks transitivity. That means that result depends on order
of how functions are presented in module. Next order causes merging of foo0
and foo1: foo2, foo0, foo1
First foo0 will be merged with foo2, foo0 will be erased. Second foo1 will be
merged with foo2.
Depending on order, things could be merged we don't expect to.
The fix:
Forbid to treat any pointer as integer, except for those, who belong to address space 0.
llvm-svn: 195769
We are going to drop debug info without a version number or with a different
version number, to make sure we don't crash when we see bitcode files with
different debug info metadata format.
Make tests more robust by removing hard-coded metadata numbers in CHECK lines.
llvm-svn: 195535
We are going to drop debug info without a version number or with a different
version number, to make sure we don't crash when we see bitcode files with
different debug info metadata format.
llvm-svn: 195504
If the beginning of the loop was also the entry block
of the function, branches were inserted to the entry block
which isn't allowed. If this occurs, create a new dummy
function entry block that branches to the start of the loop.
llvm-svn: 195493
Instead of permanently outputting "MVLL" as the file checksum, clang
will create gcno and gcda checksums by hashing the destination block
numbers of every arc. This allows for llvm-cov to check if the two gcov
files are synchronized.
Regenerated the test files so they contain the checksum. Also added
negative test to ensure error when the checksums don't match.
llvm-svn: 195191
We are slicing an array of Value pointers and process those slices in a loop.
The problem is that we might invalidate a later slice by vectorizing a former
slice.
Use a WeakVH to track the pointer. If the pointer is deleted or RAUW'ed we can
tell.
The test case will only fail when running with libgmalloc.
radar://15498655
llvm-svn: 195162
order of slices of the alloca which have exactly the same size and other
properties. This was found by a perniciously unstable sort
implementation used to flush out buggy uses of the algorithm.
The fundamental idea is that findCommonType should return the best
common type it can find across all of the slices in the range. There
were two bugs here previously:
1) We would accept an integer type smaller than a byte-width multiple,
and if there were different bit-width integer types, we would accept
the first one. This caused an actual failure in the testcase updated
here when the sort order changed.
2) If we found a bad combination of types or a non-load, non-store use
before an integer typed load or store we would bail, but if we found
the integere typed load or store, we would use it. The correct
behavior is to always use an integer typed operation which covers the
partition if one exists.
While a clever debugging sort algorithm found problem #1 in our existing
test cases, I have no useful test case ideas for #2. I spotted in by
inspection when looking at this code.
llvm-svn: 195118
(except functions marked always_inline).
Functions with 'optnone' must also have 'noinline' so they don't get
inlined into any other function.
Based on work by Andrea Di Biagio.
llvm-svn: 195046
In some case the loop exit count computation can overflow. Extend the type to
prevent most of those cases.
The problem is loops like:
int main ()
{
int a = 1;
char b = 0;
lbl:
a &= 4;
b--;
if (b) goto lbl;
return a;
}
The backedge count is 255. The induction variable type is i8. If we add one to
255 to get the exit count we overflow to zero.
To work around this issue we extend the type of the induction variable to i32 in
the case of i8 and i16.
PR17532
llvm-svn: 195008
Generally speaking, control flow paths with error reporting calls are cold.
So far, error reporting calls are calls to perror and calls to fprintf,
fwrite, etc. with stderr as the stream. This can be extended in the future.
The primary motivation is to improve block placement (the cold attribute
affects the static branch prediction heuristics).
llvm-svn: 194943
This adds a loop rerolling pass: the opposite of (partial) loop unrolling. The
transformation aims to take loops like this:
for (int i = 0; i < 3200; i += 5) {
a[i] += alpha * b[i];
a[i + 1] += alpha * b[i + 1];
a[i + 2] += alpha * b[i + 2];
a[i + 3] += alpha * b[i + 3];
a[i + 4] += alpha * b[i + 4];
}
and turn them into this:
for (int i = 0; i < 3200; ++i) {
a[i] += alpha * b[i];
}
and loops like this:
for (int i = 0; i < 500; ++i) {
x[3*i] = foo(0);
x[3*i+1] = foo(0);
x[3*i+2] = foo(0);
}
and turn them into this:
for (int i = 0; i < 1500; ++i) {
x[i] = foo(0);
}
There are two motivations for this transformation:
1. Code-size reduction (especially relevant, obviously, when compiling for
code size).
2. Providing greater choice to the loop vectorizer (and generic unroller) to
choose the unrolling factor (and a better ability to vectorize). The loop
vectorizer can take vector lengths and register pressure into account when
choosing an unrolling factor, for example, and a pre-unrolled loop limits that
choice. This is especially problematic if the manual unrolling was optimized
for a machine different from the current target.
The current implementation is limited to single basic-block loops only. The
rerolling recognition should work regardless of how the loop iterations are
intermixed within the loop body (subject to dependency and side-effect
constraints), but the significant restriction is that the order of the
instructions in each iteration must be identical. This seems sufficient to
capture all current use cases.
This pass is not currently enabled by default at any optimization level.
llvm-svn: 194939
InstCombine, in visitFPTrunc, applies the following optimization to sqrt calls:
(fptrunc (sqrt (fpext x))) -> (sqrtf x)
but does not apply the same optimization to llvm.sqrt. This is a problem
because, to enable vectorization, Clang generates llvm.sqrt instead of sqrt in
fast-math mode, and because this optimization is being applied to sqrt and not
applied to llvm.sqrt, sometimes the fast-math code is slower.
This change makes InstCombine apply this optimization to llvm.sqrt as well.
This fixes the specific problem in PR17758, although the same underlying issue
(optimizations applied to libcalls are not applied to intrinsics) exists for
other optimizations in SimplifyLibCalls.
llvm-svn: 194935
When we vectorize a scalar access with no alignment specified, we have to set
the target's abi alignment of the scalar access on the vectorized access.
Using the same alignment of zero would be wrong because most targets will have a
bigger abi alignment for vector types.
This probably fixes PR17878.
llvm-svn: 194876
We used to use std::map<IndicesVector, LoadInst*> for OriginalLoads, and when we
try to promote two arguments, they will both write to OriginalLoads causing
created loads for the two arguments to have the same original load. And the same
tbaa tag and alignment will be put to the created loads for the two arguments.
The fix is to use std::map<std::pair<Argument*, IndicesVector>, LoadInst*>
for OriginalLoads, so each Argument will write to different parts of the map.
PR17906
llvm-svn: 194846
with and without -g.
Adding a test case to make sure that the threshold used in the memory
dependence analysis is respected. The test case also checks that debug
intrinsics are not counted towards this threshold.
Differential Revision: http://llvm-reviews.chandlerc.com/D2141
llvm-svn: 194646
This adds a new scalar pass that reads a file with samples generated
by 'perf' during runtime. The samples read from the profile are
incorporated and emmited as IR metadata reflecting that profile.
The profile file is assumed to have been generated by an external
profile source. The profile information is converted into IR metadata,
which is later used by the analysis routines to estimate block
frequencies, edge weights and other related data.
External profile information files have no fixed format, each profiler
is free to define its own. This includes both the on-disk representation
of the profile and the kind of profile information stored in the file.
A common kind of profile is based on sampling (e.g., perf), which
essentially counts how many times each line of the program has been
executed during the run.
The SampleProfileLoader pass is organized as a scalar transformation.
On startup, it reads the file given in -sample-profile-file to
determine what kind of profile it contains. This file is assumed to
contain profile information for the whole application. The profile
data in the file is read and incorporated into the internal state of
the corresponding profiler.
To facilitate testing, I've organized the profilers to support two file
formats: text and native. The native format is whatever on-disk
representation the profiler wants to support, I think this will mostly
be bitcode files, but it could be anything the profiler wants to
support. To do this, every profiler must implement the
SampleProfile::loadNative() function.
The text format is mostly meant for debugging. Records are separated by
newlines, but each profiler is free to interpret records as it sees fit.
Profilers must implement the SampleProfile::loadText() function.
Finally, the pass will call SampleProfile::emitAnnotations() for each
function in the current translation unit. This function needs to
translate the loaded profile into IR metadata, which the analyzer will
later be able to use.
This patch implements the first steps towards the above design. I've
implemented a sample-based flat profiler. The format of the profile is
fairly simplistic. Each sampled function contains a list of relative
line locations (from the start of the function) together with a count
representing how many samples were collected at that line during
execution. I generate this profile using perf and a separate converter
tool.
Currently, I have only implemented a text format for these profiles. I
am interested in initial feedback to the whole approach before I send
the other parts of the implementation for review.
This patch implements:
- The SampleProfileLoader pass.
- The base ExternalProfile class with the core interface.
- A SampleProfile sub-class using the above interface. The profiler
generates branch weight metadata on every branch instructions that
matches the profiles.
- A text loader class to assist the implementation of
SampleProfile::loadText().
- Basic unit tests for the pass.
Additionally, the patch uses profile information to compute branch
weights based on instruction samples.
This patch converts instruction samples into branch weights. It
does a fairly simplistic conversion:
Given a multi-way branch instruction, it calculates the weight of
each branch based on the maximum sample count gathered from each
target basic block.
Note that this assignment of branch weights is somewhat lossy and can be
misleading. If a basic block has more than one incoming branch, all the
incoming branches will get the same weight. In reality, it may be that
only one of them is the most heavily taken branch.
I will adjust this assignment in subsequent patches.
llvm-svn: 194566
Constant merge can merge a constant with implicit alignment with one that has
explicit alignment. Before this change it was assuming that the explicit
alignment was higher than the implicit one, causing the result to be under
aligned in some cases.
Fixes pr17815.
Patch by Chris Smowton!
llvm-svn: 194506
The symptom is that an assertion is triggered. The assertion was added by
me to detect the situation when value is propagated from dead blocks.
(We can certainly get rid of assertion; it is safe to do so, because propagating
value from dead block to alive join node is certainly ok.)
The root cause of this bug is : edge-splitting is conducted on the fly,
the edge being split could be a dead edge, therefore the block that
split the critial edge needs to be flagged "dead" as well.
There are 3 ways to fix this bug:
1) Get rid of the assertion as I mentioned eariler
2) When an dead edge is split, flag the inserted block "dead".
3) proactively split the critical edges connecting dead and live blocks when
new dead blocks are revealed.
This fix go for 3) with additional 2 LOC.
Testing case was added by Rafael the other day.
llvm-svn: 194424
Summary:
Consider a GEP of:
i8* getelementptr ({ [2 x i8], i32, i8, [3 x i8] }* @main.c, i32 0, i32 0, i64 0)
If we proceeded to GEP the aforementioned object by 8, would form a GEP of:
i8* getelementptr ({ [2 x i8], i32, i8, [3 x i8] }* @main.c, i32 0, i32 0, i64 8)
Note that we would go through the first array member, causing an
out-of-bounds accesses. This is problematic because we might get fooled
if we are trying to evaluate loads using this GEP, for example, based
off of an object with a constant initializer where the array is zero.
This fixes PR17732.
Reviewers: nicholas, chandlerc, void
Reviewed By: void
CC: llvm-commits, echristo, void, aemerson
Differential Revision: http://llvm-reviews.chandlerc.com/D2093
llvm-svn: 194220
Patch by Michele Scandale!
Rewrite of the functions used to compute the backedge taken count of a
loop on LT and GT comparisons.
I decided to split the handling of LT and GT cases becasue the trick
"a > b == -a < -b" in some cases prevents the trip count computation
due to the multiplication by -1 on the two operands of the
comparison. This issue comes from the conservative computation of
value range of SCEVs: taking the negative SCEV of an expression that
have a small positive range (e.g. [0,31]), we would have a SCEV with a
fullset as value range.
Indeed, in the new rewritten function I tried to better handle the
maximum backedge taken count computation when MAX/MIN expression are
used to handle the cases where no entry guard is found.
Some test have been modified in order to check the new value correctly
(I manually check them and reasoning on possible overflow the new
values seem correct).
I finally added a new test case related to the multiplication by -1
issue on GT comparisons.
llvm-svn: 194116
Due to the previously added overflow checks, we can have a retain/release
relation that is one directional. This occurs specifically when we run into an
additive overflow causing us to drop state in only one direction. If that
occurs, we should bail and not optimize that retain/release instead of
asserting.
Apologies for the size of the testcase. It is necessary to cause the additive
cfg overflow to trigger.
rdar://15377890
llvm-svn: 194083
When the elements are extracted from a select on vectors
or a vector select, do the select on the extracted scalars
from the input if there is only one use.
llvm-svn: 194013
This adds an SimplifyLibCalls case which converts the special __sinpi and
__cospi (float & double variants) into a __sincospi_stret where appropriate to
remove duplicated work.
Patch by Tim Northover
llvm-svn: 193943
When the loop vectorizer was part of the SCC inliner pass manager gvn would
run after the loop vectorizer followed by instcombine. This way redundancy
(multiple uses) were removed and instcombine could perform scalarization on the
induction variables. Having moved the loop vectorizer to later we no longer run
any form of redundancy elimination before we perform instcombine. This caused
vectorized induction variables to survive that did not before.
On a recent iMac this helps linpack back from 6000Mflops to 7000Mflops.
This should also help lpbench and paq8p.
I ran a Release (without Asserts) build over the test-suite and did not see any
negative impact on compile time.
radar://15339680
llvm-svn: 193891
When a dependence check fails we can still try to vectorize loops with runtime
array bounds checks.
This helps linpack to vectorize a loop in dgefa. And we are back to 2x of the
scalar performance on a corei7-avx.
radar://15339680
llvm-svn: 193853
Given that backend does not handle "invoke asm" correctly ("invoke asm" will be
handled by SelectionDAGBuilder::visitInlineAsm, which does not have the right
setup for LPadToCallSiteMap) and we already made the assumption that inline asm
does not throw in InstCombiner::visitCallSite, we are going to make the same
assumption in Inliner to make sure we don't convert "call asm" to "invoke asm".
If it becomes necessary to add support for "invoke asm" later on, we will need
to modify the backend as well as remove the assumptions that inline asm does
not throw.
Fix rdar://15317907
llvm-svn: 193808
There are two ways one could implement hiding of linkonce_odr symbols in LTO:
* LLVM tells the linker which symbols can be hidden if not used from native
files.
* The linker tells LLVM which symbols are not used from other object files,
but will be put in the dso symbol table if present.
GOLD's API is the second option. It was implemented almost 1:1 in llvm by
passing the list down to internalize.
LLVM already had partial support for the first option. It is also very similar
to how ld64 handles hiding these symbols when *not* doing LTO.
This patch then
* removes the APIs for the DSO list.
* marks LTO_SYMBOL_SCOPE_DEFAULT_CAN_BE_HIDDEN all linkonce_odr unnamed_addr
global values and other linkonce_odr whose address is not used.
* makes the gold plugin responsible for handling the API mismatch.
llvm-svn: 193800
By vectorizing a series of srl, or, ... instructions we have obfuscated the
intention so much that the backend does not know how to fold this code away.
radar://15336950
llvm-svn: 193573
Partial fix for PR17459: wrong code at -O3 on x86_64-linux-gnu
(affecting trunk and 3.3)
When SCEV expands a recurrence outside of a loop it attempts to scale
by the stride of the recurrence. Chained recurrences don't work that
way. We could compute binomial coefficients, but would hve to
guarantee that the chained AddRec's are in a perfectly reduced form.
llvm-svn: 193438
Partial fix for PR17459: wrong code at -O3 on x86_64-linux-gnu
(affecting trunk and 3.3)
ScalarEvolutionNormalization was attempting to normalize by adding and
subtracting strides. Chained recurrences don't work that way.
llvm-svn: 193437
This patch teaches GlobalStatus to analyze a call that uses the global value as
a callee, not as an argument.
With this change internalize call handle the common use of linkonce_odr
functions. This reduces the number of linkonce_odr functions in a LTO build of
clang (checked with the emit-llvm gold plugin option) from 1730 to 60.
llvm-svn: 193436
The loop vectorizer does not currently understand how to vectorize
extractelement instructions. The existing check, which excluded all
vector-valued instructions, did not catch extractelement instructions because
it checked only the return value. As a result, vectorization would proceed,
producing illegal instructions like this:
%58 = extractelement <2 x i32> %15, i32 0
%59 = extractelement i32 %58, i32 0
where the second extractelement is illegal because its first operand is not a vector.
llvm-svn: 193434
Make sure we mark all loops (scalar and vector) when vectorizing,
so that we don't try to vectorize them anymore. Also, set unroll
to 1, since this is what we check for on early exit.
llvm-svn: 193349
Major steps include:
1). introduces a not-addr-taken bit-field in GlobalVariable
2). GlobalOpt pass sets "not-address-taken" if it proves a global varirable
dosen't have its address taken.
3). AA use this info for disambiguation.
llvm-svn: 193251
When a linkonce_odr value that is on the dso list is not unnamed_addr
we can still look to see if anything is actually using its address. If
not, it is safe to hide it.
This patch implements that by moving GlobalStatus to Transforms/Utils
and using it in Internalize.
llvm-svn: 193090
A landing pad can be jumped to only by the unwind edge of an invoke
instruction. If we eliminate a partially redundant load in a landing pad, it
will create a basic block that violates this constraint. It then leads to other
problems down the line if it tries to merge that basic block with the landing
pad. Avoid this by not eliminating the load in a landing pad.
PR17621
llvm-svn: 193064
One optimization simplify-cfg performs is the converting of switches to
lookup tables if the switch has > 4 cases. This is done by:
1. Finding the max/min case value and calculating the switch case range.
2. Create a lookup table basic block.
3. Perform a check in the switch's BB to see if the input value is in
the switch's case range. If the input value satisfies said predicate
branch to the lookup table BB, otherwise branch to the switch's default
destination BB using the default value as the result.
The conditional check consists of subtracting the min case value of the
table from any input iN value and then ensuring that said value is
unsigned less than the size of the lookup table represented as an iN
value.
If the lookup table is a covered lookup table, the size of the table will be N
which is 0 as an iN value. Thus the comparison will be an `icmp ult` of an iN
value against 0 which is always false yielding the incorrect result.
This patch fixes this problem by recognizing if we have a covered lookup table
and if we do, unconditionally jumps to the lookup table BB since the covering
property of the lookup table implies no input values could not be handled by
said BB.
rdar://15268442
llvm-svn: 193045
If the predecessor's being spliced into a landing pad, then we need the PHIs to
come first and the rest of the predecessor's code to come *after* the landing
pad instruction.
llvm-svn: 193035
Before this patch we relied on the order of phi nodes when we looked for phi
nodes of the same type. This could prevent vectorization of cases where there
was a phi node of a second type in between phi nodes of some type.
This is important for vectorization of an internal graphics kernel. On the test
suite + external on x86_64 (and on a run on armv7s) it showed no impact on
either performance or compile time.
radar://15024459
llvm-svn: 192537
If a function seen at compile time is not necessarily the one linked to
the binary being built, it is illegal to change the actual arguments
passing to it.
e.g.
--------------------------
void foo(int lol) {
// foo() has linkage satisifying isWeakForLinker()
// "lol" is not used at all.
}
void bar(int lo2) {
// xform to foo(undef) is illegal, as compiler dose not know which
// instance of foo() will be linked to the the binary being built.
foo(lol2);
}
-----------------------------
Such functions can be captured by isWeakForLinker(). NOTE that
mayBeOverridden() is insufficient for this purpose as it dosen't include
linkage types like AvailableExternallyLinkage and LinkOnceODRLinkage.
Take link_odr* as an example, it indicates a set of *EQUIVALENT* globals
that can be merged at link-time. However, the semantic of
*EQUIVALENT*-functions includes parameters. Changing parameters breaks
the assumption.
Thank John McCall for help, especially for the explanation of subtle
difference between linkage types.
rdar://11546243
llvm-svn: 192302
is updated to use DITypeRef.
Move isUnsignedDIType and getOriginalTypeSize from DebugInfo.h to be static
helper functions in DwarfCompileUnit. We already have a static helper function
"isTypeSigned" in DwarfCompileUnit, and a pointer to DwarfDebug is added to
resolve the derived-from field. All three functions need to go across link
for derived-from fields, so we need to get hold of a type identifier map.
A pointer to DwarfDebug is also added to DbgVariable in order to resolve the
derived-from field.
Debug info verifier is updated to check a derived-from field is a TypeRef.
Verifier will not go across link for derived-from fields, in debug info finder,
we go across the link to add derived-from fields to types.
Function getDICompositeType is only used by dragonegg and since dragonegg does
not generate identifier for types, we use an empty map to resolve the
derived-from field.
When printing a derived-from field, we use DITypeRef::getName to either return
the type identifier or getName of the DIType.
A paired commit at clang is required due to changes to DIBuilder.
llvm-svn: 192018
UpdatePHINodes has an optimization to reuse an existing PHI node, where it
first deletes all of its entries and then replaces them. Unfortunately, in the
case where we had duplicate predecessors (which are allowed so long as the
associated PHI entries have the same value), the loop removing the existing PHI
entries from the to-be-reused PHI would assert (if that PHI was not the one
which had the duplicates).
llvm-svn: 192001
Sort the operands of the other entries in the current vectorization root
according to the first entry's operands opcodes.
%conv0 = uitofp ...
%load0 = load float ...
= fmul %conv0, %load0
= fmul %load0, %conv1
= fmul %load0, %conv2
Make sure that we recursively vectorize <%conv0, %conv1, %conv2> and <%load0,
%load0, %load0>.
This makes it more likely to obtain vectorizable trees. We have to be careful
when we sort that we don't destroy 'good' existing ordering implied by source
order.
radar://15080067
llvm-svn: 191977
Generalize the API so we can distinguish symbols that are needed just for a DSO
symbol table from those that are used from some native .o.
The symbols that are only wanted for the dso symbol table can be dropped if
llvm can prove every other dso has a copy (linkonce_odr) and the address is not
important (unnamed_addr).
llvm-svn: 191922
Don't vectorize with a runtime check if it requires a
comparison between pointers with different address spaces.
The values can't be assumed to be directly comparable.
Previously it would create an illegal bitcast.
llvm-svn: 191862
This recursively strips all GEPs like the existing code. It also handles bitcasts and
other operations that do not change the pointer value.
llvm-svn: 191847
Switch instructions were crashing the StructurizeCFG pass, and it's
probably easier anyway if we don't need to handle them in this pass.
Reviewed-by: Christian König <christian.koenig@amd.com>
llvm-svn: 191841
is updated to use DITypeRef.
Move isUnsignedDIType and getOriginalTypeSize from DebugInfo.h to be static
helper functions in DwarfCompileUnit. We already have a static helper function
"isTypeSigned" in DwarfCompileUnit, and a pointer to DwarfDebug is added to
resolve the derived-from field. All three functions need to go across link
for derived-from fields, so we need to get hold of a type identifier map.
A pointer to DwarfDebug is also added to DbgVariable in order to resolve the
derived-from field.
Debug info verifier is updated to check a derived-from field is a TypeRef.
Verifier will not go across link for derived-from fields, in debug info finder,
we go across the link to add derived-from fields to types.
Function getDICompositeType is only used by dragonegg and since dragonegg does
not generate identifier for types, we use an empty map to resolve the
derived-from field.
When printing a derived-from field, we use DITypeRef::getName to either return
the type identifier or getName of the DIType.
A paired commit at clang is required due to changes to DIBuilder.
llvm-svn: 191800
Inspired by the object from the SLPVectorizer. This found a minor bug in the
debug loc restoration in the vectorizer where the location of a following
instruction was attached instead of the location from the original instruction.
llvm-svn: 191673
when it was actually a Constant*.
There are quite a few other casts to Instruction that might have the same problem,
but this is the only one I have a test case for.
llvm-svn: 191668
Currently foldSelectICmpAndOr asserts if the "or" involves a vector
containing several of the same power of two. We can easily avoid this by
only performing the fold on integer types, like foldSelectICmpAnd does.
Fixes <rdar://problem/15012516>
llvm-svn: 191552
Remove the command line argument "struct-path-tbaa" since we should not depend
on command line argument to decide which format the IR file is using. Instead,
we check the first operand of the tbaa tag node, if it is a MDNode, we treat
it as struct-path aware TBAA format, otherwise, we treat it as scalar TBAA
format.
When clang starts to use struct-path aware TBAA format no matter whether
struct-path-tbaa is no, and we can auto-upgrade existing bc files, the support
for scalar TBAA format can be dropped.
Existing testing cases are updated to use the struct-path aware TBAA format.
llvm-svn: 191538
We were previously using getFirstInsertionPt to insert PHI
instructions when vectorizing, but getFirstInsertionPt also skips past
landingpads, causing this to generate invalid IR.
We can avoid this issue by using getFirstNonPHI instead.
llvm-svn: 191526
Put them under a separate flag for experimentation. They are more likely to
interfere with loop vectorization which happens later in the pass pipeline.
llvm-svn: 191371
Some supplemental information for r191314: We would like to make sure SLP Vectorizer will not try to vectorize tiny trees even with a negative threshold so we set the cost to INT_MAX.
llvm-svn: 191327
This is safe per C++11 18.6.1.1p3: [operator new returns] a non-null pointer to
suitably aligned storage (3.7.4), or else throw a bad_alloc exception. This
requirement is binding on a replacement version of this function.
Brings us a tiny bit closer to eliminating more vector push_backs.
llvm-svn: 191310
Revert 191122 - with extra checks we are allowed to vectorize math library
function calls.
Standard library indentifiers are reserved names so functions with external
linkage must not overrided them. However, functions with internal linkage can.
Therefore, we can vectorize calls to math library functions with a check for
external linkage and matching signature. This matches what we do during
SelectionDAG building.
llvm-svn: 191206
Overflow doesn't affect the correctness of equalities. Computing this is cheap,
we just reuse the computation for the inbounds case and try to peel of more
non-inbounds GEPs. This pattern is unlikely to ever appear in code generated by
Clang, but SCEV occasionally produces it.
llvm-svn: 191200
SROA wants to convert any types of equivalent widths but it's not possible to
convert vectors of pointers to an integer scalar with a single cast. As a
workaround we add a bitcast to the corresponding int ptr type first. This type
of cast used to be an edge case but has become common with SLP vectorization.
Fixes PR17271.
llvm-svn: 191143
Reapply r191108 with a fix for a memory corruption error I introduced. Of
course, we can't reference the scalars that we replace by vectorizing and then
call their eraseFromParent method. I only 'needed' the scalars to get the
DebugLoc. Just store the DebugLoc before actually vectorizing instead. As a nice
side effect, this also simplifies the interface between BoUpSLP and the
HorizontalReduction class to returning a value pointer (the vectorized tree
root).
radar://14607682
llvm-svn: 191123
The problem of r191017 is that when GVN fabricate a val-number for a dead instruction (in order
to make following expr-PRE happy), it forget to fabricate a leader-table entry for it as well.
llvm-svn: 191118
Match reductions starting at binary operation feeding into a phi. The code
handles trees like
r += v1 + v2 + v3 ...
and
r += v1
r += v2
...
and
r *= v1 + v2 + ...
We currently only handle associative operations (add, fadd fast).
The code can now also handle reductions feeding into stores.
a[i] = v1 + v2 + v3 + ...
The code is currently disabled behind the flag "-slp-vectorize-hor". The cost
model for most architectures is not there yet.
I found one opportunity of a horizontal reduction feeding a phi in TSVC
(LoopRerolling-flt) and there are several opportunities where reductions feed
into stores.
radar://14607682
llvm-svn: 191108
The GEP pattern is what SCEV expander emits for "ugly geps". The latter is what
you get for pointer subtraction in C code. The rest of instcombine already
knows how to deal with that so just canonicalize on that.
llvm-svn: 191090
If "C1/X" were having multiple uses, the only benefit of this
transformation is to potentially shorten critical path. But it is at the
cost of instroducing additional div.
The additional div may or may not incur cost depending on how div is
implemented. If it is implemented using Newton–Raphson iteration, it dosen't
seem to incur any cost (FIXME). However, if the div blocks the entire
pipeline, that sounds to be pretty expensive. Let CodeGen to take care
this transformation.
This patch sees 6% on a benchmark.
rdar://15032743
llvm-svn: 191037
This is how it ignores the dead code:
1) When a dead branch target, say block B, is identified, all the
blocks dominated by B is dead as well.
2) The PHIs of those blocks in dominance-frontier(B) is updated such
that the operands corresponding to dead predecessors are replaced
by "UndefVal".
Using lattice's jargon, the "UndefVal" is the "Top" in essence.
Phi node like this "phi(v1 bb1, undef xx)" will be optimized into
"v1" if v1 is constant, or v1 is an instruction which dominate this
PHI node.
3) When analyzing the availability of a load L, all dead mem-ops which
L depends on disguise as a load which evaluate exactly same value as L.
4) The dead mem-ops will be materialized as "UndefVal" during code motion.
llvm-svn: 191017
XCore target: Add XCoreTargetTransformInfo
This is where getNumberOfRegisters() resides, which in turn returns the
number of vector registers (=0).
llvm-svn: 190936
Some of this code is no longer necessary since int<->ptr casts are no
longer occur as of r187444.
This also fixes handling vectors of pointers, and adds a bunch of new
testcases for vectors and address spaces.
llvm-svn: 190885
We can't insert an insertelement after an invoke. We would have to split a
critical edge. So when we see a phi node that uses an invoke we just give up.
radar://14990770
llvm-svn: 190871
other in memory.
The motivation was to get rid of truncate and shift right instructions that get
in the way of paired load or floating point load.
E.g.,
Consider the following example:
struct Complex {
float real;
float imm;
};
When accessing a complex, llvm was generating a 64-bits load and the imm field
was obtained by a trunc(lshr) sequence, resulting in poor code generation, at
least for x86.
The idea is to declare that two load instructions is the canonical form for
loading two arithmetic type, which are next to each other in memory.
Two scalar loads at a constant offset from each other are pretty
easy to detect for the sorts of passes that like to mess with loads.
<rdar://problem/14477220>
llvm-svn: 190870
We would have to compute the pre increment value, either by computing it on
every loop iteration or by splitting the edge out of the loop and inserting a
computation for it there.
For now, just give up vectorizing such loops.
Fixes PR17179.
llvm-svn: 190790
This pass was based on the previous (essentially unused) profiling
infrastructure and the assumption that by ordering the basic blocks at
the IR level in a particular way, the correct layout would happen in the
end. This sometimes worked, and mostly didn't. It also was a really
naive implementation of the classical paper that dates from when branch
predictors were primarily directional and when loop structure wasn't
commonly available. It also didn't factor into the equation
non-fallthrough branches and other machine level details.
Anyways, for all of these reasons and more, I wrote
MachineBlockPlacement, which completely supercedes this pass. It both
uses modern profile information infrastructure, and actually works. =]
llvm-svn: 190748
The PowerPC A2 core greatly benefits from aggressive concatenation unrolling;
use the new getUnrollingPreferences to enable this by default when targeting
the PPC A2 core.
llvm-svn: 190549
LLVM IR doesn't currently allow atomic bool load/store operations, and the
transformation is dubious anyway because it isn't profitable on all platforms.
PR17163.
llvm-svn: 190357
Several architectures use the same instruction to perform both a comparison and
a subtract. The instruction selection framework does not allow to consider
different basic blocks to expose such fusion opportunities.
Therefore, these instructions are “merged” by CSE at MI IR level.
To increase the likelihood of CSE to apply in such situation, we reorder the
operands of the comparison, when they have the same complexity, so that they
matches the order of the most frequent subtract.
E.g.,
icmp A, B
...
sub B, A
<rdar://problem/14514580>
llvm-svn: 190352
The work on this project was left in an unfinished and inconsistent state.
Hopefully someone will eventually get a chance to implement this feature, but
in the meantime, it is better to put things back the way the were. I have
left support in the bitcode reader to handle the case-range bitcode format,
so that we do not lose bitcode compatibility with the llvm 3.3 release.
This reverts the following commits: 155464, 156374, 156377, 156613, 156704,
156757, 156804 156808, 156985, 157046, 157112, 157183, 157315, 157384, 157575,
157576, 157586, 157612, 157810, 157814, 157815, 157880, 157881, 157882, 157884,
157887, 157901, 158979, 157987, 157989, 158986, 158997, 159076, 159101, 159100,
159200, 159201, 159207, 159527, 159532, 159540, 159583, 159618, 159658, 159659,
159660, 159661, 159703, 159704, 160076, 167356, 172025, 186736
llvm-svn: 190328
Field 2 of DIType (Context), field 9 of DIDerivedType (TypeDerivedFrom),
field 12 of DICompositeType (ContainingType), fields 2, 7, 12 of DISubprogram
(Context, Type, ContainingType).
llvm-svn: 190205
This reverts commit r189886.
I found a corner case where this optimization is not valid:
Say we have a "linkonce_odr unnamed_addr" in two translation units:
* In TU 1 this optimization kicks in and makes it hidden.
* In TU 2 it gets const merged with a constant that is *not* unnamed_addr,
resulting in a non unnamed_addr constant with default visibility.
* The static linker rules for combining visibility them produce a hidden
symbol, which is incorrect from the point of view of the non unnamed_addr
constant.
The one place we can do this is when we know that the symbol is not used from
another TU in the same shared object, i.e., during LTO. I will move it there.
llvm-svn: 189954
"(icmp op i8 A, B)" is equivalent to "(icmp op i8 (A & 0xff), B)" as a
degenerate case. Allowing this as a "masked" comparison when analysing "(icmp)
&/| (icmp)" allows us to combine them in more cases.
rdar://problem/7625728
llvm-svn: 189931
Even in cases which aren't universally optimisable like "(A & B) != 0 && (A &
C) != 0", the masks can make one of the comparisons completely redundant. In
this case, since we've gone to the effort of spotting masked comparisons we
should combine them.
rdar://problem/7625728
llvm-svn: 189930
Original message:
If a constant or a function has linkonce_odr linkage and unnamed_addr, mark
hidden. Being linkonce_odr guarantees that it is available in every dso that
needs it. Being a constant/function with unnamed_addr guarantees that the
copies don't have to be merged.
llvm-svn: 189886
The reason that I am turning off this optimization is that there is an
additional case where a block can escape that has come up. Specifically, this
occurs when a block is used in a scope outside of its current scope.
This can cause a captured retainable object pointer whose life is preserved by
the objc_retainBlock to be deallocated before the block is invoked.
An example of the code needed to trigger the bug is:
----
\#import <Foundation/Foundation.h>
int main(int argc, const char * argv[]) {
void (^somethingToDoLater)();
{
NSObject *obj = [NSObject new];
somethingToDoLater = ^{
[obj self]; // Crashes here
};
}
NSLog(@"test.");
somethingToDoLater();
return 0;
}
----
In the next commit, I remove all the dead code that results from this.
Once I put in the fixing commit I will bring back the tests that I deleted in
this commit.
rdar://14802782.
rdar://14868830.
llvm-svn: 189869
1) If the width of vectorization list candidate is bigger than vector reg width, we will break it down to fit the vector reg.
2) We do not vectorize the width which is not power of two.
The performance result shows it will help some spec benchmarks. mesa improved 6.97% and ammp improved 1.54%.
llvm-svn: 189830
The existing code missed some edge cases when e.g. we're going to emit sqrtf but
only the availability of sqrt was checked. This happens on odd platforms like
windows.
llvm-svn: 189724
PR17026. Also avoid undefined shifts and shift amounts larger than 64 bits
(those are always undef because we can't represent integer types that large).
llvm-svn: 189672
When unrolling is disabled in the pass manager, the loop vectorizer should also
not unroll loops. This will allow the -fno-unroll-loops option in Clang to
behave as expected (even for vectorizable loops). The loop vectorizer's
-force-vector-unroll option will (continue to) override the pass-manager
setting (including -force-vector-unroll=0 to force use of the internal
auto-selection logic).
In order to test this, I added a flag to opt (-disable-loop-unrolling) to force
disable unrolling through opt (the analog of -fno-unroll-loops in Clang). Also,
this fixes a small bug in opt where the loop vectorizer was enabled only after
the pass manager populated the queue of passes (the global_alias.ll test needed
a slight update to the RUN line as a result of this fix).
llvm-svn: 189499
The builder inserts from before the insert point,
not after, so this would insert before the last
instruction in the bundle instead of after it.
I'm not sure if this can actually be a problem
with any of the current insertions.
llvm-svn: 189285
DICompositeType will have an identifier field at position 14. For now, the
field is set to null in DIBuilder.
For DICompositeTypes where the template argument field (the 13th field)
was optional, modify DIBuilder to make sure the template argument field is set.
Now DICompositeType has 15 fields.
Update DIBuilder to use NULL instead of "i32 0" for null value of a MDNode.
Update verifier to check that DICompositeType has 15 fields and the last
field is null or a MDString.
Update testing cases to include an extra field for DICompositeType.
The identifier field will be used by type uniquing so a front end can
genearte a DICompositeType with a unique identifer.
llvm-svn: 189282
This patch enables unrolling of loops when vectorization is legal but not profitable.
We add a new class InnerLoopUnroller, that extends InnerLoopVectorizer and replaces some of the vector-specific logic with scalars.
This patch does not introduce any runtime regressions and improves the following workloads:
SingleSource/Benchmarks/Shootout/matrix -22.64%
SingleSource/Benchmarks/Shootout-C++/matrix -13.06%
External/SPEC/CINT2006/464_h264ref/464_h264ref -3.99%
SingleSource/Benchmarks/Adobe-C++/simple_types_constant_folding -1.95%
llvm-svn: 189281
The current version of StripDeadDebugInfo became stale and no longer actually
worked since it was expecting an older version of debug info.
This patch updates it to use DebugInfoFinder and the modern DebugInfo classes as
much as possible to make it more redundent to such changes. Additionally, the
only place where that was avoided (the code where we replace the old sets with
the new), I call verify on the DIContextUnit implying that if the format changes
and my live set changes no longer make sense an assert will be hit. In order to
ensure that that occurs I have included a test case.
The actual stripping of the dead debug info follows the same strategy as was
used before in this class: find the live set and replace the old set in the
given compile unit (which may contain dead global variables/functions) with the
new live one.
llvm-svn: 189078
A single metadata will not span multiple lines. This also helps me with
my script to automatic update the testing cases.
A debug info testing case should have a llvm.dbg.cu.
Do not use hard-coded id for debug nodes.
llvm-svn: 189033
using GEPs. Previously, it used a number of different heuristics for
analyzing the GEPs. Several of these were conservatively correct, but
failed to fall back to SCEV even when SCEV might have given a reasonable
answer. One was simply incorrect in how it was formulated.
There was good code already to recursively evaluate the constant offsets
in GEPs, look through pointer casts, etc. I gathered this into a form
code like the SLP code can use in a previous commit, which allows all of
this code to become quite simple.
There is some performance (compile time) concern here at first glance as
we're directly attempting to walk both pointers constant GEP chains.
However, a couple of thoughts:
1) The very common cases where there is a dynamic pointer, and a second
pointer at a constant offset (usually a stride) from it, this code
will actually not do any unnecessary work.
2) InstCombine and other passes work very hard to collapse constant
GEPs, so it will be rare that we iterate here for a long time.
That said, if there remain performance problems here, there are some
obvious things that can improve the situation immensely. Doing
a vectorizer-pass-wide memoizer for each individual layer of pointer
values, their base values, and the constant offset is likely to be able
to completely remove redundant work and strictly limit the scaling of
the work to scrape these GEPs. Since this optimization was not done on
the prior version (which would still benefit from it), I've not done it
here. But if folks have benchmarks that slow down it should be straight
forward for them to add.
I've added a test case, but I'm not really confident of the amount of
testing done for different access patterns, strides, and pointer
manipulation.
llvm-svn: 189007
Update iterator when the SLP vectorizer changes the instructions in the basic
block by restarting the traversal of the basic block.
Patch by Yi Jiang!
Fixes PR 16899.
llvm-svn: 188832
This adds a llvm.copysign intrinsic; We already have Libfunc recognition for
copysign (which is turned into the FCOPYSIGN SDAG node). In order to
autovectorize calls to copysign in the loop vectorizer, we need a corresponding
intrinsic as well.
In addition to the expected changes to the language reference, the loop
vectorizer, BasicTTI, and the SDAG builder (the intrinsic is transformed into
an FCOPYSIGN node, just like the function call), this also adds FCOPYSIGN to a
few lists in LegalizeVector{Ops,Types} so that vector copysigns can be
expanded.
In TargetLoweringBase::initActions, I've made the default action for FCOPYSIGN
be Expand for vector types. This seems correct for all in-tree targets, and I
think is the right thing to do because, previously, there was no way to generate
vector-values FCOPYSIGN nodes (and most targets don't specify an action for
vector-typed FCOPYSIGN).
llvm-svn: 188728
- Instead of setting the suffixes in a bunch of places, just set one master
list in the top-level config. We now only modify the suffix list in a few
suites that have one particular unique suffix (.ml, .mc, .yaml, .td, .py).
- Aside from removing the need for a bunch of lit.local.cfg files, this enables
4 tests that were inadvertently being skipped (one in
Transforms/BranchFolding, a .s file each in DebugInfo/AArch64 and
CodeGen/PowerPC, and one in CodeGen/SI which is now failing and has been
XFAILED).
- This commit also fixes a bunch of config files to use config.root instead of
older copy-pasted code.
llvm-svn: 188513
When both constants are positive or both constants are negative,
InstCombine already simplifies comparisons like this, but when
it's exactly zero and -1, the operand sorting ends up reversed
and the pattern fails to match. Handle that special case.
Follow up for rdar://14689217
llvm-svn: 188512
Use the pointer size if datalayout is available.
Use i64 if it's not, which is consistent with what other
places do when the pointer size is unknown.
The test doesn't really test this in a useful way
since it will be transformed to that later anyway,
but this now tests it for non-zero arrays and when
datalayout isn't available. The cases in
visitGetElementPtrInst should save an extra re-visit to
the newly created GEP since it won't need to cleanup after
itself.
llvm-svn: 188339
When computing the use set of a store, we need to add the store to the write
set prior to iterating over later instructions. Otherwise, if there is a later
aliasing load of that store, that load will not be tagged as a use, and bad
things will happen.
trackUsesOfI still adds later dependent stores of an instruction to that
instruction's write set, but it never sees the original instruction, and so
when tracking uses of a store, the store must be added to the write set by the
caller.
Fixes PR16834.
llvm-svn: 188329
However, opt -O2 doesn't run mem2reg directly so nobody noticed until r188146
when SROA started sending more things directly down the PromoteMemToReg path.
In order to revert r187191, I also revert dependent revisions r187296, r187322
and r188146. Fixes PR16867. Does not add the testcases from that PR, but both
of them should get added for both mem2reg and sroa when this revert gets
unreverted.
llvm-svn: 188327
Do not generate new vector values for the same entries because we know that the incoming values
from the same block must be identical.
llvm-svn: 188185
Various tests had sprung up over the years which had --check-prefix=ABC on the
RUN line, but "CHECK-ABC:" later on. This happened to work before, but was
strictly incorrect. FileCheck is getting stricter soon though.
Patch by Ron Ofir.
llvm-svn: 188173
These functions used to assume that the lsb of an integer corresponds
to vector element 0, whereas for big-endian it's the other way around:
the msb is in the first element and the lsb is in the last element.
Fixes MultiSource/Benchmarks/mediabench/gsm/toast for z.
llvm-svn: 188155
I fixed the aforementioned problems that came up on some of the linux boxes.
Major thanks to Nick Lewycky for his help debugging!
rdar://14590914
llvm-svn: 188122
This reverts commit r187941.
The commit was passing on my os x box, but it is failing on some non-osx
platforms. I do not have time to look into it now, so I am reverting and will
recommit after I figure this out.
llvm-svn: 187946
All libm floating-point rounding functions, except for round(), had their own
ISD nodes. Recent PowerPC cores have an instruction for round(), and so here I'm
adding ISD::FROUND so that round() can be custom lowered as well.
For the most part, this is straightforward. I've added an intrinsic
and a matching ISD node just like those for nearbyint() and friends. The
SelectionDAG pattern I've named frnd (because ISD::FP_ROUND has already claimed
fround).
This will be used by the PowerPC backend in a follow-up commit.
llvm-svn: 187926
Also remove checking of llvm.dbg.sp since it is not used in generating dwarf.
Current state of Finder:
DebugInfoFinder tries to list all debug info MDNodes used in a module. To
list debug info MDNodes used by an instruction, DebugInfoFinder provides
processDeclare, processValue and processLocation to handle DbgDeclareInst,
DbgValueInst and DbgLoc attached to instructions. processModule will go
through all DICompileUnits in llvm.dbg.cu and list debug info MDNodes
used by the CUs.
TODO:
1> Finder has a list of CUs, SPs, Types, Scopes and global variables. We
need to add a list of variables that are used by DbgDeclareInst and
DbgValueInst.
2> MDString fields should be null or isa<MDString> and MDNode fields should be
null or isa<MDNode>. We currently use empty string or int 0 to represent null.
3> Go though Verify functions and make sure that they check field types.
4> Clean up existing testing cases to remove llvm.dbg.sp and make sure each
testing case has a llvm.dbg.cu.
Re-apply r187609 with fix to pass ocaml binding. vmcore.ml generates a debug
location with scope being metadata !{}, in verifier we treat this as a null
scope.
llvm-svn: 187812
Function attributes are the future! So just query whether we want to realign the
stack directly from the function instead of through a random target options
structure.
llvm-svn: 187618
Also remove checking of llvm.dbg.sp since it is not used in generating dwarf.
Current state of Finder:
DebugInfoFinder tries to list all debug info MDNodes used in a module. To
list debug info MDNodes used by an instruction, DebugInfoFinder provides
processDeclare, processValue and processLocation to handle DbgDeclareInst,
DbgValueInst and DbgLoc attached to instructions. processModule will go
through all DICompileUnits in llvm.dbg.cu and list debug info MDNodes
used by the CUs.
TODO:
1> Finder has a list of CUs, SPs, Types, Scopes and global variables. We
need to add a list of variables that are used by DbgDeclareInst and
DbgValueInst.
2> MDString fields should be null or isa<MDString> and MDNode fields should be
null or isa<MDNode>. We currently use empty string or int 0 to represent null.
3> Go though Verify functions and make sure that they check field types.
4> Clean up existing testing cases to remove llvm.dbg.sp and make sure each
testing case has a llvm.dbg.cu.
llvm-svn: 187609
Call into ComputeMaskedBits to figure out which bits are set on both add
operands and determine if the value is a power-of-two-or-zero or not.
llvm-svn: 187445
It will now only convert the arguments / return value and call
the underlying function if the types are able to be bitcasted.
This avoids using fp<->int conversions that would occur before.
llvm-svn: 187444
Also always add DIType, DISubprogram and DIGlobalVariable to the list
in DebugInfoFinder without checking them, so we can verify them later
on.
llvm-svn: 187285
We used to call Verify before adding DICompileUnit to the list, and now we
remove the check and always add DICompileUnit to the list in DebugInfoFinder,
so we can verify them later on.
llvm-svn: 187237
robust. It now uses an InstVisitor and worklist to actually walk the
uses of the Alloca transitively and detect the pattern which we can
directly promote: loads & stores of the whole alloca and instructions we
can completely ignore.
Also, with this new implementation teach both the predicate for testing
whether we can promote and the promotion engine itself to use the same
code so we no longer have strange divergence between the two code paths.
I've added some silly test cases to demonstrate that we can handle
slightly more degenerate code patterns now. See the below for why this
is even interesting.
Performance impact: roughly 1% regression in the performance of SROA or
ScalarRepl on a large C++-ish test case where most of the allocas are
basically ready for promotion. The reason is because of silly redundant
work that I've left FIXMEs for and which I'll address in the next
commit. I wanted to separate this commit as it changes the behavior.
Once the redundant work in removing the dead uses of the alloca is
fixed, this code appears to be faster than the old version. =]
So why is this useful? Because the previous requirement for promotion
required a *specific* visit pattern of the uses of the alloca to verify:
we *had* to look for no more than 1 intervening use. The end goal is to
have SROA automatically detect when an alloca is already promotable and
directly hand it to the mem2reg machinery rather than trying to
partition and rewrite it. This is a 25% or more performance improvement
for SROA, and a significant chunk of the delta between it and
ScalarRepl. To get there, we need to make mem2reg actually capable of
promoting allocas which *look* promotable to SROA without have SROA do
tons of work to massage the code into just the right form.
This is actually the tip of the iceberg. There are tremendous potential
savings we can realize here by de-duplicating work between mem2reg and
SROA.
llvm-svn: 187191
Also avoid locals evicting locals just because they want a cheaper register.
Problem: MI Sched knows exactly how many registers we have and assumes
they can be colored. In cases where we have large blocks, usually from
unrolled loops, greedy coloring fails. This is a source of
"regressions" from the MI Scheduler on x86. I noticed this issue on
x86 where we have long chains of two-address defs in the same live
range. It's easy to see this in matrix multiplication benchmarks like
IRSmk and even the unit test misched-matmul.ll.
A fundamental difference between the LLVM register allocator and
conventional graph coloring is that in our model a live range can't
discover its neighbors, it can only verify its neighbors. That's why
we initially went for greedy coloring and added eviction to deal with
the hard cases. However, for singly defined and two-address live
ranges, we can optimally color without visiting neighbors simply by
processing the live ranges in instruction order.
Other beneficial side effects:
It is much easier to understand and debug regalloc for large blocks
when the live ranges are allocated in order. Yes, global allocation is
still very confusing, but it's nice to be able to comprehend what
happened locally.
Heuristics could be added to bias register assignment based on
instruction locality (think late register pairing, banks...).
Intuituvely this will make some test cases that are on the threshold
of register pressure more stable.
llvm-svn: 187139
Make sure the context and type fields are MDNodes. We will generate
verification errors if those fields are non-empty strings.
Fix testing cases to make them pass the verifier.
llvm-svn: 187106
The language reference says that:
"If a symbol appears in the @llvm.used list, then the compiler,
assembler, and linker are required to treat the symbol as if there is
a reference to the symbol that it cannot see"
Since even the linker cannot see the reference, we must assume that
the reference can be using the symbol table. For example, a user can add
__attribute__((used)) to a debug helper function like dump and use it from
a debugger.
llvm-svn: 187103
schedule an alloca for another iteration in SROA. This only showed up
with a mixture of promotable and unpromotable selects and phis. Added
a test case for this.
llvm-svn: 187031
pending speculation for a phi node. The problem here is that we were
using growth of the specluation set as an indicator of whether
speculation would occur, and if the phi node is already in the set we
don't see it grow. This is a symptom of the fact that this signal is
a total hack.
Unfortunately, I couldn't really come up with a non-hacky way of
signaling that promotion remains valid *after* speculation occurs, such
that we only speculate when all else looks good for promotion. In the
end, I went with at least a much more explicit approach of doing the
work of queuing inside the phi and select processing and setting
a preposterously named flag to convey that we're in the special state of
requiring speculating before promotion.
Thanks to Richard Trieu and Nick Lewycky for the excellent work reducing
a testcase for this from a pretty giant, nasty assert in a big
application. =] The testcase was excellent.
llvm-svn: 187029
MDNodes used by DbgDeclareInst and DbgValueInst.
Another 16 testing cases failed and they are disabled with
-disable-debug-info-verifier.
A total of 34 cases are disabled with -disable-debug-info-verifier and will be
corrected.
llvm-svn: 186902
GlobalOpt simplifies llvm.compiler.used by removing any members that are also
in the more strict llvm.used. Handle the special case where llvm.compiler.used
becomes empty.
llvm-svn: 186778
We were incorrectly using compiler_used instead of compiler.used. Unfortunately
the passes using the broken name had tests also using the broken name.
llvm-svn: 186705
SROA.
The crux of the issue is that now we track uses of a partition of the
alloca in two places: the iterators over the partitioning uses and the
previously collected split uses vector. We weren't accounting for the
fact that the split uses might invalidate integer widening in ways other
than due to their width (in this case due to being volatile).
Further reduced testcase added to the tests.
llvm-svn: 186655
end of a vector. This was found with ASan. I've had one other report of
a crasher, but thus far been unable to reproduce the crash. It may well
be fixed with this version, and if not I'd like to get more information
from the build bots about what is happening.
See r186316 for the full commit log for the new implementation of the
SROA algorithm.
llvm-svn: 186565
Duncan pointed out a mistake in my fix in r186425 when only one of the allocas
being compared had the target-default alignment. This is essentially his
suggested solution. Thanks!
llvm-svn: 186510
For safety, the inliner cannot decrease the allignment on an alloca when
merging it with another.
I've included two variants of the test case for this: one with DataLayout
available, and one without. When DataLayout is not available, if only one of
the allocas uses the default alignment (getAlignment() == 0), then they cannot
be safely merged.
llvm-svn: 186425
a bot.
This reverts the commit which introduced a new implementation of the
fancy SROA pass designed to reduce its overhead. I'll skip the huge
commit log here, refer to r186316 if you're looking for how this all
works and why it works that way.
llvm-svn: 186332
different core implementation strategy.
Previously, SROA would build a relatively elaborate partitioning of an
alloca, associate uses with each partition, and then rewrite the uses of
each partition in an attempt to break apart the alloca into chunks that
could be promoted. This was very wasteful in terms of memory and compile
time because regardless of how complex the alloca or how much we're able
to do in breaking it up, all of the datastructure work to analyze the
partitioning was done up front.
The new implementation attempts to form partitions of the alloca lazily
and on the fly, rewriting the uses that make up that partition as it
goes. This has a few significant effects:
1) Much simpler data structures are used throughout.
2) No more double walk of the recursive use graph of the alloca, only
walk it once.
3) No more complex algorithms for associating a particular use with
a particular partition.
4) PHI and Select speculation is simplified and happens lazily.
5) More precise information is available about a specific use of the
alloca, removing the need for some side datastructures.
Ultimately, I think this is a much better implementation. It removes
about 300 lines of code, but arguably removes more like 500 considering
that some code grew in the process of being factored apart and cleaned
up for this all to work.
I've re-used as much of the old implementation as possible, which
includes the lion's share of code in the form of the rewriting logic.
The interesting new logic centers around how the uses of a partition are
sorted, and split into actual partitions.
Each instruction using a pointer derived from the alloca gets
a 'Partition' entry. This name is totally wrong, but I'll do a rename in
a follow-up commit as there is already enough churn here. The entry
describes the offset range accessed and the nature of the access. Once
we have all of these entries we sort them in a very specific way:
increasing order of begin offset, followed by whether they are
splittable uses (memcpy, etc), followed by the end offset or whatever.
Sorting by splittability is important as it simplifies the collection of
uses into a partition.
Once we have these uses sorted, we walk from the beginning to the end
building up a range of uses that form a partition of the alloca.
Overlapping unsplittable uses are merged into a single partition while
splittable uses are broken apart and carried from one partition to the
next. A partition is also introduced to bridge splittable uses between
the unsplittable regions when necessary.
I've looked at the performance PRs fairly closely. PR15471 no longer
will even load (the module is invalid). Not sure what is up there.
PR15412 improves by between 5% and 10%, however it is nearly impossible
to know what is holding it up as SROA (the entire pass) takes less time
than reading the IR for that test case. The analysis takes the same time
as running mem2reg on the final allocas. I suspect (without much
evidence) that the new implementation will scale much better however,
and it is just the small nature of the test cases that makes the changes
small and noisy. Either way, it is still simpler and cleaner I think.
llvm-svn: 186316
This conversion was done with the following bash script:
find test/Transforms -name "*.ll" | \
while read NAME; do
echo "$NAME"
if ! grep -q "^; *RUN: *llc" $NAME; then
TEMP=`mktemp -t temp`
cp $NAME $TEMP
sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \
while read FUNC; do
sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)define\([^@]*\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3define\4@$FUNC(/g" $TEMP
done
mv $TEMP $NAME
fi
done
llvm-svn: 186269
This update was done with the following bash script:
find test/Transforms -name "*.ll" | \
while read NAME; do
echo "$NAME"
if ! grep -q "^; *RUN: *llc" $NAME; then
TEMP=`mktemp -t temp`
cp $NAME $TEMP
sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \
while read FUNC; do
sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP
done
mv $TEMP $NAME
fi
done
llvm-svn: 186268
If an outside loop user of the reduction value uses the header phi node we
cannot just reduce the vectorized phi value in the vector code epilog because
we would loose VF-1 reductions.
lp:
p = phi (0, lv)
lv = lv + 1
...
brcond , lp, outside
outside:
usr = add 0, p
(Say the loop iterates two times, the value of p coming out of the loop is one).
We cannot just transform this to:
vlp:
p = phi (<0,0>, lv)
lv = lv + <1,1>
..
brcond , lp, outside
outside:
p_reduced = p[0] + [1];
usr = add 0, p_reduced
(Because the original loop iterated two times the vectorized loop would iterate
one time, but p_reduced ends up being zero instead of one).
We would have to execute VF-1 iterations in the scalar remainder loop in such
cases. For now, just disable vectorization.
PR16522
llvm-svn: 186256
In general, one should always complete CFG modifications first, update
CFG-based analyses, like Dominatores and LoopInfo, then generate
instruction sequences.
LoopVectorizer was creating a new loop, calling SCEVExpander to
generate checks, then updating LoopInfo. I just changed the order.
llvm-svn: 186241
Fixes a 35% degradation compared to unvectorized code in
MiBench/automotive-susan and an equally serious regression on a private
image processing benchmark.
radar://14351991
llvm-svn: 186188
against a constant."
This reverts commit r186107. It didn't handle wrapping arithmetic in the
loop correctly and thus caused the following C program to count from
0 to UINT64_MAX instead of from 0 to 255 as intended:
#include <stdio.h>
int main() {
unsigned char first = 0, last = 255;
do { printf("%d\n", first); } while (first++ != last);
}
Full test case and instructions to reproduce with just the -indvars pass
sent to the original review thread rather than to r186107's commit.
llvm-svn: 186152
Before we could vectorize PHINodes scanning successors was a good way of finding candidates. Now we can vectorize the phinodes which is simpler.
llvm-svn: 186139
Patch by Michele Scandale!
Adds a special handling of the case where, during the loop exit
condition rewriting, the exit value is a constant of bitwidth lower
than the type of the induction variable: instead of introducing a
trunc operation in order to match correctly the operand types, it
allows to convert the constant value to an equivalent constant,
depending on the initial value of the induction variable and the trip
count, in order have an equivalent comparison between the induction
variable and the new constant.
llvm-svn: 186107
We can vectorize them because in the case where we wrap in the address space the
unvectorized code would have had to access a pointer value of zero which is
undefined behavior in address space zero according to the LLVM IR semantics.
(Thank you Duncan, for pointing this out to me).
Fixes PR16592.
llvm-svn: 186088
predecessors of the two blocks it is attempting to merge supply the
same incoming values to any phi in the successor block. This change
allows merging in the case where there is one or more incoming values
that are undef. The undef values are rewritten to match the non-undef
value that flows from the other edge. Patch by Mark Lacey.
llvm-svn: 186069
Without the changes introduced into this patch, if TRE saw any allocas at all,
TRE would not perform TRE *or* mark callsites with the tail marker.
Because TRE runs after mem2reg, this inadequacy is not a death sentence. But
given a callsite A without escaping alloca argument, A may not be able to have
the tail marker placed on it due to a separate callsite B having a write-back
parameter passed in via an argument with the nocapture attribute.
Assume that B is the only other callsite besides A and B only has nocapture
escaping alloca arguments (*NOTE* B may have other arguments that are not passed
allocas). In this case not marking A with the tail marker is unnecessarily
conservative since:
1. By assumption A has no escaping alloca arguments itself so it can not
access the caller's stack via its arguments.
2. Since all of B's escaping alloca arguments are passed as parameters with
the nocapture attribute, we know that B does not stash said escaping
allocas in a manner that outlives B itself and thus could be accessed
indirectly by A.
With the changes introduced by this patch:
1. If we see any escaping allocas passed as a capturing argument, we do
nothing and bail early.
2. If we do not see any escaping allocas passed as captured arguments but we
do see escaping allocas passed as nocapture arguments:
i. We do not perform TRE to avoid PR962 since the code generator produces
significantly worse code for the dynamic allocas that would be created
by the TRE algorithm.
ii. If we do not return twice, mark call sites without escaping allocas
with the tail marker. *NOTE* This excludes functions with escaping
nocapture allocas.
3. If we do not see any escaping allocas at all (whether captured or not):
i. If we do not have usage of setjmp, mark all callsites with the tail
marker.
ii. If there are no dynamic/variable sized allocas in the function,
attempt to perform TRE on all callsites in the function.
Based off of a patch by Nick Lewycky.
rdar://14324281.
llvm-svn: 186057
(add nsw x, (and x, y)) isn't a power of two if x is zero, it's zero
(add nsw x, (xor x, y)) isn't a power of two if y has bits set that aren't set in x
llvm-svn: 185954
The following transforms are valid if -C is a power of 2:
(icmp ugt (xor X, C), ~C) -> (icmp ult X, C)
(icmp ult (xor X, C), -C) -> (icmp uge X, C)
These are nice, they get rid of the xor.
llvm-svn: 185915
Back in r179493 we determined that two transforms collided with each
other. The fix back then was to reorder the transforms so that the
preferred transform would give it a try and then we would try the
secondary transform. However, it was noted that the best approach would
canonicalize one transform into the other, removing the collision and
allowing us to optimize IR given to us in that form.
llvm-svn: 185808
This is a complete re-write if the bottom-up vectorization class.
Before this commit we scanned the instruction tree 3 times. First in search of merge points for the trees. Second, for estimating the cost. And finally for vectorization.
There was a lot of code duplication and adding the DCE exposed bugs. The new design is simpler and DCE was a part of the design.
In this implementation we build the tree once. After that we estimate the cost by scanning the different entries in the constructed tree (in any order). The vectorization phase also works on the built tree.
llvm-svn: 185774
This transform was originally added in r185257 but later removed in
r185415. The original transform would create instructions speculatively
and then discard them if the speculation was proved incorrect. This has
been replaced with a scheme that splits the transform into two parts:
preflight and fold. While we preflight, we build up fold actions that
inform the folding stage on how to act.
llvm-svn: 185667
Chandler Carruth