A common way to implement nearbyint is by fiddling with the floating
point environment and calling rint. This is used at least by the BSD
libm and musl. As such, canonicalizing the latter to the former will
create infinite loops for libm and generally pessimize performance, at
least when the generic C versions are used.
This change preserves the rint in the libcall translation and also
handles the domain truncation logic, so that rint with float argument
will be reduced to rintf etc.
llvm-svn: 299247
Summary: Currently the VP metadata was dropped when InstCombine converts a call to direct call. This patch converts the VP metadata to branch_weights so that its hotness is recorded.
Reviewers: eraman, davidxl
Reviewed By: davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D31344
llvm-svn: 299228
Summary:
Triggered by commit r298620: "[LV] Vectorize GEPs".
If we encounter a vector GEP with scalar arguments, we splat the scalar
into a vector of appropriate size before we scatter the argument.
Reviewers: arsenm, mehdi_amini, bkramer
Reviewed By: arsenm
Subscribers: bjope, mssimpso, wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D31416
llvm-svn: 299186
Since there is no sdiv in SCEV, an 'udiv' is a better canonical form than an 'sdiv' as the user of induction variable
Differential Revision: https://reviews.llvm.org/D31488
llvm-svn: 299118
Some of the GEP combines (e.g., descaling) can't handle vector GEPs. We have an
existing check that attempts to bail out if given a vector GEP. However, the
check only tests the GEP's pointer operand. A GEP results in a vector of
pointers if at least one of its operands is vector-typed (e.g., its pointer
operand could be a scalar, but its index could be a vector). We should just
check the type of the GEP itself. This should fix PR32414.
Reference: https://bugs.llvm.org/show_bug.cgi?id=32414
Differential Revision: https://reviews.llvm.org/D31470
llvm-svn: 299017
The vectorizer tries to replace truncations of induction variables with new
induction variables having the smaller type. After r295063, this optimization
was applied to all integer induction variables, including non-primary ones.
When optimizing the truncation of a non-primary induction variable, we still
need to transform the new induction so that it has the correct start value.
This should fix PR32419.
Reference: https://bugs.llvm.org/show_bug.cgi?id=32419
llvm-svn: 298882
Summary:
We are incorrectly folding selects into phi nodes when the incoming value of a phi
node is a constant vector. This optimization is done in `FoldOpIntoPhi` when the
select condition is a phi node with constant incoming values.
Without the fix, we are miscompiling (i.e. incorrectly folding the
select into the phi node) when the vector contains non-zero
elements.
This patch fixes the miscompile and we will correctly fold based on the
select vector operand (see added test cases).
Reviewers: majnemer, sanjoy, spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D31189
llvm-svn: 298845
The first variant contains all current transformations except
transforming switches into lookup tables. The second variant
contains all current transformations.
The switch-to-lookup-table conversion results in code that is more
difficult to analyze and optimize by other passes. Most importantly,
it can inhibit Dead Code Elimination. As such it is often beneficial to
only apply this transformation very late. A common example is inlining,
which can often result in range restrictions for the switch expression.
Changes in execution time according to LNT:
SingleSource/Benchmarks/Misc/fp-convert +3.03%
MultiSource/Benchmarks/ASC_Sequoia/CrystalMk/CrystalMk -11.20%
MultiSource/Benchmarks/Olden/perimeter/perimeter -10.43%
and a couple of smaller changes. For perimeter it also results 2.6%
a smaller binary.
Differential Revision: https://reviews.llvm.org/D30333
llvm-svn: 298799
This moves it to the iterator facade utilities giving it full random
access semantics, etc. It can also now be used with standard algorithms
like std::all_of and std::any_of and range adaptors like llvm::reverse.
Also make the semantics of iterating match what every other iterator
uses and forbid decrementing past the begin iterator. This was used as
a hacky way to work around iterator invalidation. However, every
instance trying to do this failed to actually avoid touching invalid
iterators despite the clear documentation that the removed and all
subsequent iterators become invalid including the end iterator. So I've
added a return of the next iterator to removeCase and rewritten the
loops that were doing this to correctly follow the iterator pattern of
either incremneting or removing and assigning fresh values to the
iterator and the end.
In one case we were trying to go backwards to make this cleaner but it
doesn't actually work. I've made that code match the code we use
everywhere else to remove cases as we iterate. This changes the order of
cases in one test output and I moved that test to CHECK-DAG so it
wouldn't care -- the order isn't semantically meaningful anyways.
llvm-svn: 298791
Reason: breaks linking Chromium with LLD + ThinLTO (a pass crashes)
LLVM bug: https://bugs.llvm.org//show_bug.cgi?id=32413
Original change description:
[LV] Vectorize GEPs
This patch adds support for vectorizing GEPs. Previously, we only generated
vector GEPs on-demand when creating gather or scatter operations. All GEPs from
the original loop were scalarized by default, and if a pointer was to be stored
to memory, we would have to build up the pointer vector with insertelement
instructions.
With this patch, we will vectorize all GEPs that haven't already been marked
for scalarization.
The patch refines collectLoopScalars to more exactly identify the scalar GEPs.
The function now more closely resembles collectLoopUniforms. And the patch
moves vector GEP creation out of vectorizeMemoryInstruction and into the main
vectorization loop. The vector GEPs needed for gather and scatter operations
will have already been generated before vectoring the memory accesses.
Original Differential Revision: https://reviews.llvm.org/D30710
llvm-svn: 298735
Summary: Declarations need to be filtered out when counting functions.
Reviewers: eraman
Subscribers: Prazek, llvm-commits
Differential Revision: https://reviews.llvm.org/D31336
llvm-svn: 298720
Summary: In DeadArgumentElimination, the call instructions will be replaced. We also need to set the prof weights so that function inlining can find the correct profile.
Reviewers: eraman
Reviewed By: eraman
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D31143
llvm-svn: 298660
Library functions can have specific semantics that affect the behavior of
certain passes. DSE, for instance, gives special treatment to malloc-ed pointers
but not to pointers returned from an equivalently typed (but differently named)
function.
MetaRenamer ought not to alter program semantics, so library functions must
remain untouched.
Reviewers: mehdi_amini, majnemer, chandlerc, davide
Reviewed By: davide
Subscribers: davide, llvm-commits
Differential Revision: https://reviews.llvm.org/D31304
llvm-svn: 298659
Summary: The current prefix based function layout algorithm only looks at function's entry count, which is not sufficient. A function should be grouped together if its entry count or any call edge count is hot.
Reviewers: davidxl, eraman
Reviewed By: eraman
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D31225
llvm-svn: 298656
The new test asserts that scalarized memory operations get memcheck metadata
added even if the loop is only unrolled.
Differential Revision: https://reviews.llvm.org/D30972
llvm-svn: 298641
Summary:
The cumulative size of the bitcode files for a very large application
can be huge, particularly with -g. In a distributed build environment,
all of these files must be sent to the remote build node that performs
the thin link step, and this can exceed size limits.
The thin link actually only needs the summary along with a bitcode
symbol table. Until we have a proper bitcode symbol table, simply
stripping the debug metadata results in significant size reduction.
Add support for an option to additionally emit minimized bitcode
modules, just for use in the thin link step, which for now just strips
all debug metadata. I plan to add a cc1 option so this can be invoked
easily during the compile step.
However, care must be taken to ensure that these minimized thin link
bitcode files produce the same index as with the original bitcode files,
as these original bitcode files will be used in the backends.
Specifically:
1) The module hash used for caching is typically produced by hashing the
written bitcode, and we want to include the hash that would correspond
to the original bitcode file. This is because we want to ensure that
changes in the stripped portions affect caching. Added plumbing to emit
the same module hash in the minimized thin link bitcode file.
2) The module paths in the index are constructed from the module ID of
each thin linked bitcode, and typically is automatically generated from
the input file path. This is the path used for finding the modules to
import from, and obviously we need this to point to the original bitcode
files. Added gold-plugin support to take a suffix replacement during the
thin link that is used to override the identifier on the MemoryBufferRef
constructed from the loaded thin link bitcode file. The assumption is
that the build system can specify that the minimized bitcode file has a
name that is similar but uses a different suffix (e.g. out.thinlink.bc
instead of out.o).
Added various tests to ensure that we get identical index files out of
the thin link step.
Reviewers: mehdi_amini, pcc
Subscribers: Prazek, llvm-commits
Differential Revision: https://reviews.llvm.org/D31027
llvm-svn: 298638
This patch adds support for vectorizing GEPs. Previously, we only generated
vector GEPs on-demand when creating gather or scatter operations. All GEPs from
the original loop were scalarized by default, and if a pointer was to be stored
to memory, we would have to build up the pointer vector with insertelement
instructions.
With this patch, we will vectorize all GEPs that haven't already been marked
for scalarization.
The patch refines collectLoopScalars to more exactly identify the scalar GEPs.
The function now more closely resembles collectLoopUniforms. And the patch
moves vector GEP creation out of vectorizeMemoryInstruction and into the main
vectorization loop. The vector GEPs needed for gather and scatter operations
will have already been generated before vectoring the memory accesses.
Differential Revision: https://reviews.llvm.org/D30710
llvm-svn: 298620
The code for generating scalar base pointers in vectorizeMemoryInstruction is
not needed. We currently scalarize all GEPs and maintain the scalarized values
in VectorLoopValueMap. The GEP cloning in this unneeded code is the same as
that in scalarizeInstruction. The test cases that changed as a result of this
patch changed because we were able to reuse the scalarized GEP that we
previously generated instead of cloning a new one.
Differential Revision: https://reviews.llvm.org/D30587
llvm-svn: 298615
Summary: ThinLTO will annotate the CFG twice. If the branch weight is set by the first annotation, we should not set the branch weight again in the second annotation because the first annotation is more accurate as there is less optimization that could affect debug info accuracy.
Reviewers: tejohnson, davidxl
Reviewed By: tejohnson
Subscribers: mehdi_amini, aprantl, llvm-commits
Differential Revision: https://reviews.llvm.org/D31228
llvm-svn: 298602
insertelement (insertelement X, Y, IdxC1), ScalarC, IdxC2 -->
insertelement (insertelement X, ScalarC, IdxC2), Y, IdxC1
As noted in the code comment and seen in the test changes, the motivation is that by pulling
constant insertion up, we may be able to constant fold some insertelement instructions.
Differential Revision: https://reviews.llvm.org/D31196
llvm-svn: 298520
Summary: Subtracts can have constants on the left side, but we don't shrink them based on demanded bits. This patch fixes that to match the right hand side.
Reviewers: davide, majnemer, spatel, sanjoy, hfinkel
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D31119
llvm-svn: 298478
Currently the default C calling convention functions are treated
the same as compute kernels. Make this explicit so the default
calling convention can be changed to a non-kernel.
Converted with perl -pi -e 's/define void/define amdgpu_kernel void/'
on the relevant test directories (and undoing in one place that actually
wanted a non-kernel).
llvm-svn: 298444
This adds a parameter to @llvm.objectsize that makes it return
conservative values if it's given null.
This fixes PR23277.
Differential Revision: https://reviews.llvm.org/D28494
llvm-svn: 298430
Summary: Inliner should update the branch_weights annotation to scale it to proper value.
Reviewers: davidxl, eraman
Reviewed By: eraman
Subscribers: zzheng, llvm-commits
Differential Revision: https://reviews.llvm.org/D30767
llvm-svn: 298270
Summary:
In case we are loading on a phi-load in SimplifyPartiallyRedundantLoad.
Try to phi translate it into incoming values in the predecessors before
we search for available loads.
This needs https://reviews.llvm.org/D30524
Reviewers: davide, sanjoy, efriedma, dberlin, rengolin
Reviewed By: dberlin
Subscribers: junbuml, llvm-commits
Differential Revision: https://reviews.llvm.org/D30543
llvm-svn: 298217
Summary:
The reverse of an artbitrary bitpattern is also an arbitrary
bitpattern.
Reviewers: trentxintong, arsenm, majnemer
Reviewed By: majnemer
Subscribers: majnemer, wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D31118
llvm-svn: 298201
Loop unswitching can be extremely harmful for a SIMT target. In case
if hoisted condition is not uniform a SIMT machine will execute both
clones of a loop sequentially. Therefor LoopUnswitch checks if the
condition is non-divergent.
Since DivergenceAnalysis adds an expensive PostDominatorTree analysis
not needed for non-SIMT targets a new option is added to avoid unneded
analysis initialization. The method getAnalysisUsage is called when
TargetTransformInfo is not yet available and we cannot use it here.
For that reason a new field DivergentTarget is added to PassManagerBuilder
to control the behavior and set this field from a target.
Differential Revision: https://reviews.llvm.org/D30796
llvm-svn: 298104
We were not handling getelemenptr instructions of vector type before.
Since getelemenptr instructions for vector types follow the same rule as
getelementptr instructions for non-vector types, we can just handle them
in the same way.
llvm-svn: 298028
[Reapplies r297971 and punting on finding a better API for findDbgValues()]
This patch improves debug info quality in InstCombine by looking at
values that are about to be deleted, checking whether there are any
dbg.value instrinsics referring to them, and potentially encoding the
semantics of the deleted instruction into the dbg.value's
DIExpression.
In the example in the testcase (which was extracted from XNU) there is a sequence of
%4 = load %struct.entry*, %struct.entry** %next2, align 8, !dbg !41
%5 = bitcast %struct.entry* %4 to i8*, !dbg !42
%add.ptr4 = getelementptr inbounds i8, i8* %5, i64 -8, !dbg !43
%6 = bitcast i8* %add.ptr4 to %struct.entry*, !dbg !44
call void @llvm.dbg.value(metadata %struct.entry* %6, i64 0, metadata !20, metadata !21), !dbg 34
When these instructions are eliminated by instcombine one after
another, we can still salvage the otherwise dead debug info:
- Bitcasts have no effect, so have the dbg.value point to operand(0)
- Loads can be expressed via a DW_OP_deref
- Constant gep instructions can be replaced by DWARF expression arithmetic
The API introduced by this patch is not specific to instcombine and
can be useful in other places, too.
rdar://problem/30725338
Differential Revision: https://reviews.llvm.org/D30919
llvm-svn: 297994
As the related tests show, we're not canonicalizing to this form for scalars or vectors yet,
but this solves the immediate problem in:
https://bugs.llvm.org/show_bug.cgi?id=32306
llvm-svn: 297989
This patch improves debug info quality in InstCombine by looking at
values that are about to be deleted, checking whether there are any
dbg.value instrinsics referring to them, and potentially encoding the
semantics of the deleted instruction into the dbg.value's
DIExpression.
In the example in the testcase (which was extracted from XNU) there is a sequence of
%4 = load %struct.entry*, %struct.entry** %next2, align 8, !dbg !41
%5 = bitcast %struct.entry* %4 to i8*, !dbg !42
%add.ptr4 = getelementptr inbounds i8, i8* %5, i64 -8, !dbg !43
%6 = bitcast i8* %add.ptr4 to %struct.entry*, !dbg !44
call void @llvm.dbg.value(metadata %struct.entry* %6, i64 0, metadata !20, metadata !21), !dbg 34
When these instructions are eliminated by instcombine one after
another, we can still salvage the otherwise dead debug info:
- Bitcasts have no effect, so have the dbg.value point to operand(0)
- Loads can be expressed via a DW_OP_deref
- Constant gep instructions can be replaced by DWARF expression arithmetic
The API introduced by this patch is not specific to instcombine and
can be useful in other places, too.
rdar://problem/30725338
Differential Revision: https://reviews.llvm.org/D30919
llvm-svn: 297971
in r297374.
I've extracted a small version of this from the C++ metaprogram Richard
came up with to exercise these kinds of issues and written comments to
describe both how to reproduce a fresh version of the test case and what
likely failure modes are.
The test case is still a bit brittle as it depends on the particular
inline cost modeling and SCC visitation order, but it definitely would
have caught the bug right away when developing things so it seems
a really valuable test case to have.
llvm-svn: 297935
This patch adds the value profile support to profile the size parameter of
memory intrinsic calls: memcpy, memcmp, and memmov.
Differential Revision: http://reviews.llvm.org/D28965
llvm-svn: 297897
Summary:
In SamplePGO, if the profile is collected from non-LTO binary, and used to drive ThinLTO, the indirect call promotion may fail because ThinLTO adjusts local function names to avoid conflicts. There are two places of where the mismatch can happen:
1. thin-link prepends SourceFileName to front of FuncName to build the GUID (GlobalValue::getGlobalIdentifier). Unlike instrumentation FDO, SamplePGO does not use the PGOFuncName scheme and therefore the indirect call target profile data contains a hash of the OriginalName.
2. backend compiler promotes some local functions to global and appends .llvm.{$ModuleHash} to the end of the FuncName to derive PromotedFunctionName
This patch tries at the best effort to find the GUID from the original local function name (in profile), and use that in ICP promotion, and in SamplePGO matching that happens in the backend after importing/inlining:
1. in thin-link, it builds the map from OriginalName to GUID so that when thin-link reads in indirect call target profile (represented by OriginalName), it knows which GUID to import.
2. in backend compiler, if sample profile reader cannot find a profile match for PromotedFunctionName, it will try to find if there is a match for OriginalFunctionName.
3. in backend compiler, we build symbol table entry for OriginalFunctionName and pointer to the same symbol of PromotedFunctionName, so that ICP can find the correct target to promote.
Reviewers: mehdi_amini, tejohnson
Reviewed By: tejohnson
Subscribers: llvm-commits, Prazek
Differential Revision: https://reviews.llvm.org/D30754
llvm-svn: 297757
If it is possible for the RHS of a shift operation to be greater than or equal
to the bit-width, then the result might be undef, and we can't report any known
bits.
In some cases, this was allowing a transformation in instcombine which widened
an undef value from i1 to i32, increasing the range of values that a function
could return.
Differential revision: https://reviews.llvm.org/D30781
llvm-svn: 297724
getIntrinsicInstrCost() used to only compute scalarization cost based on types.
This patch improves this so that the actual arguments are checked when they are
available, in order to handle only unique non-constant operands.
Tests updates:
Analysis/CostModel/X86/arith-fp.ll
Transforms/LoopVectorize/AArch64/interleaved_cost.ll
Transforms/LoopVectorize/ARM/interleaved_cost.ll
The improvement in getOperandsScalarizationOverhead() to differentiate on
constants made it necessary to update the interleaved_cost.ll tests even
though they do not relate to intrinsics.
Review: Hal Finkel
https://reviews.llvm.org/D29540
llvm-svn: 297705
Summary:
This change solves the same problem as D30726, except that this only
throws out the bathwater.
AST was not correctly tracking and deleting UnknownInstructions via
handles. The existing code only tracks "pointers" in its
`ASTCallbackVH`, so an UnknownInstruction (that isn't also def'ing a
pointer used by another memory instruction) never gets a
`ASTCallbackVH`.
There are two other ways to solve this problem:
- Use the `PointerRec` scheme for both known and unknown instructions.
- Use a `CallbackVH` that erases the offending Instruction from the
UnknownInstruction list.
Both of the above changes seemed to be significantly (and unnecessarily
IMO) more complex than this.
Reviewers: chandlerc, dberlin, hfinkel, reames
Subscribers: mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D30849
llvm-svn: 297539
This reverts r293386, r294027, r294029 and r296411.
Turns out the SLP tree isn't actually a "tree" and we don't handle
accessing the same packet of loads in several different orders well,
causing miscompiles.
Revert until we can fix this properly.
llvm-svn: 297493
entire SCC before iterating on newly-introduced call edges resulting
from any inlined function bodies.
This more closely matches the behavior of the old PM's inliner. While it
wasn't really clear to me initially, this behavior is actually essential
to the inliner behaving reasonably in its current design.
Because the inliner is fundamentally a bottom-up inliner and all of its
cost modeling is designed around that it often runs into trouble within
an SCC where we don't have any meaningful bottom-up ordering to use. In
addition to potentially cyclic, infinite inlining that we block with the
inline history mechanism, it can also take seemingly simple call graph
patterns within an SCC and turn them into *insanely* large functions by
accidentally working top-down across the SCC without any of the
threshold limitations that traditional top-down inliners use.
Consider this diabolical monster.cpp file that Richard Smith came up
with to help demonstrate this issue:
```
template <int N> extern const char *str;
void g(const char *);
template <bool K, int N> void f(bool *B, bool *E) {
if (K)
g(str<N>);
if (B == E)
return;
if (*B)
f<true, N + 1>(B + 1, E);
else
f<false, N + 1>(B + 1, E);
}
template <> void f<false, MAX>(bool *B, bool *E) { return f<false, 0>(B, E); }
template <> void f<true, MAX>(bool *B, bool *E) { return f<true, 0>(B, E); }
extern bool *arr, *end;
void test() { f<false, 0>(arr, end); }
```
When compiled with '-DMAX=N' for various values of N, this will create an SCC
with a reasonably large number of functions. Previously, the inliner would try
to exhaust the inlining candidates in a single function before moving on. This,
unfortunately, turns it into a top-down inliner within the SCC. Because our
thresholds were never built for that, we will incrementally decide that it is
always worth inlining and proceed to flatten the entire SCC into that one
function.
What's worse, we'll then proceed to the next function, and do the exact same
thing except we'll skip the first function, and so on. And at each step, we'll
also make some of the constant factors larger, which is awesome.
The fix in this patch is the obvious one which makes the new PM's inliner use
the same technique used by the old PM: consider all the call edges across the
entire SCC before beginning to process call edges introduced by inlining. The
result of this is essentially to distribute the inlining across the SCC so that
every function incrementally grows toward the inline thresholds rather than
allowing the inliner to grow one of the functions vastly beyond the threshold.
The code for this is a bit awkward, but it works out OK.
We could consider in the future doing something more powerful here such as
prioritized order (via lowest cost and/or profile info) and/or a code-growth
budget per SCC. However, both of those would require really substantial work
both to design the system in a way that wouldn't break really useful
abstraction decomposition properties of the current inliner and to be tuned
across a reasonably diverse set of code and workloads. It also seems really
risky in many ways. I have only found a single real-world file that triggers
the bad behavior here and it is generated code that has a pretty pathological
pattern. I'm not worried about the inliner not doing an *awesome* job here as
long as it does *ok*. On the other hand, the cases that will be tricky to get
right in a prioritized scheme with a budget will be more common and idiomatic
for at least some frontends (C++ and Rust at least). So while these approaches
are still really interesting, I'm not in a huge rush to go after them. Staying
even closer to the existing PM's behavior, especially when this easy to do,
seems like the right short to medium term approach.
I don't really have a test case that makes sense yet... I'll try to find a
variant of the IR produced by the monster template metaprogram that is both
small enough to be sane and large enough to clearly show when we get this wrong
in the future. But I'm not confident this exists. And the behavior change here
*should* be unobservable without snooping on debug logging. So there isn't
really much to test.
The test case updates come from two incidental changes:
1) We now visit functions in an SCC in the opposite order. I don't think there
really is a "right" order here, so I just update the test cases.
2) We no longer compute some analyses when an SCC has no call instructions that
we consider for inlining.
llvm-svn: 297374
!type metadata can not be dropped. An alternative to this is adding
!type metadata from the replaced globals to the replacement, but that
may weaken type tests and make them slower at the same time.
The merged global gets !dbg metadata from replaced globals, and can
end up with multiple debug locations.
llvm-svn: 297327
Because IRBuilder performs constant-folding, it's not guaranteed that an
instruction in the original loop map to an instruction in the vector loop. It
could map to a constant vector instead. The handling of first-order recurrences
was incorrectly making this assumption when setting the IRBuilder's insert
point.
llvm-svn: 297302
This patch also renames the PR number the test points to. The previous
reference was PR29559, but that bug was somehow deleted and recreated under
PR30183.
llvm-svn: 297295
Summary: Use AA when scanning to find an available load value.
Reviewers: rengolin, mcrosier, hfinkel, trentxintong, dberlin
Reviewed By: rengolin, dberlin
Subscribers: aemerson, dberlin, llvm-commits
Differential Revision: https://reviews.llvm.org/D30352
llvm-svn: 297284
Recommitting patch which was previously reverted in r297159. These
changes should address the casting issues.
The original patch enables dbg.value intrinsics to be attached to
newly inserted PHI nodes.
Differential Review: https://reviews.llvm.org/D30701
llvm-svn: 297269
A block with an UnreachableInst does not transfer execution to a successor.
The problem was exposed by GVN-hoist. This patch fixes bug 32153.
Patch by Aditya Kumar.
Differential Revision: https://reviews.llvm.org/D30667
llvm-svn: 297254
Itanium ABI may have an address point one byte after the end of a
vtable. When such vtable global is split, the !type metadata needs to
follow the right vtable.
Differential Revision: https://reviews.llvm.org/D30716
llvm-svn: 297236
This was committed at r297155 and reverted at r297166 because of an
over-reaching clang test. That should be fixed with r297189.
This is one part of solving a recent bug report:
http://lists.llvm.org/pipermail/llvm-dev/2017-February/110293.html
This keeps with our general approach: changing arbitrary shuffles is off-limts,
but changing splat is ok. The transform is very similar to the existing
shrinkBitwiseLogic() canonicalization.
Differential Revision: https://reviews.llvm.org/D30123
llvm-svn: 297232
Summary:
The purpose of coro.end intrinsic is to allow frontends to mark the cleanup and
other code that is only relevant during the initial invocation of the coroutine
and should not be present in resume and destroy parts.
In landing pads coro.end is replaced with an appropriate instruction to unwind to
caller. The handling of coro.end differs depending on whether the target is
using landingpad or WinEH exception model.
For landingpad based exception model, it is expected that frontend uses the
`coro.end`_ intrinsic as follows:
```
ehcleanup:
%InResumePart = call i1 @llvm.coro.end(i8* null, i1 true)
br i1 %InResumePart, label %eh.resume, label %cleanup.cont
cleanup.cont:
; rest of the cleanup
eh.resume:
%exn = load i8*, i8** %exn.slot, align 8
%sel = load i32, i32* %ehselector.slot, align 4
%lpad.val = insertvalue { i8*, i32 } undef, i8* %exn, 0
%lpad.val29 = insertvalue { i8*, i32 } %lpad.val, i32 %sel, 1
resume { i8*, i32 } %lpad.val29
```
The `CoroSpit` pass replaces `coro.end` with ``True`` in the resume functions,
thus leading to immediate unwind to the caller, whereas in start function it
is replaced with ``False``, thus allowing to proceed to the rest of the cleanup
code that is only needed during initial invocation of the coroutine.
For Windows Exception handling model, a frontend should attach a funclet bundle
referring to an enclosing cleanuppad as follows:
```
ehcleanup:
%tok = cleanuppad within none []
%unused = call i1 @llvm.coro.end(i8* null, i1 true) [ "funclet"(token %tok) ]
cleanupret from %tok unwind label %RestOfTheCleanup
```
The `CoroSplit` pass, if the funclet bundle is present, will insert
``cleanupret from %tok unwind to caller`` before
the `coro.end`_ intrinsic and will remove the rest of the block.
Reviewers: majnemer
Reviewed By: majnemer
Subscribers: llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D25543
llvm-svn: 297223
When expanding the set of uniform instructions beyond the seed instructions
(e.g., consecutive pointers), we mark a new instruction uniform if all its
loop-varying users are uniform. We should also allow users that are consecutive
or interleaved memory accesses. This fixes cases where we have an instruction
that is used as the pointer operand of a consecutive access but also used by a
non-memory instruction that later becomes uniform as part of the expansion.
llvm-svn: 297179
This reverts commit r296488.
As noted by David Blaikie on llvm-commits, I overlooked the case of a
debug function being inlined into a nodebug function being inlined
into a debug function.
llvm-svn: 297163
Summary:
We should check if loop size allows us to peel at least one iteration
before we do so.
Patch by Max Kazantsev!
Reviewers: sanjoy, mkuper, efriedma
Reviewed By: mkuper
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30632
llvm-svn: 297122
Summary: We do not need that special handling because the debug info is more accurate now. Performance testing shows no regression on google internal benchmarks.
Reviewers: davidxl, aprantl
Reviewed By: aprantl
Subscribers: llvm-commits, aprantl
Differential Revision: https://reviews.llvm.org/D30658
llvm-svn: 297038
Any unsuccessful llvm.type.checked.load devirtualizations will be translated
into uses of llvm.type.test, so we need to add the resulting llvm.type.test
intrinsics to the function summaries so that the LowerTypeTests pass will
export them.
Differential Revision: https://reviews.llvm.org/D29808
llvm-svn: 296939
Summary:
If a loop contains a Phi node which has an invariant input from back
edge, it is profitable to peel such loops (rather than unroll them) to
use the advantage that this Phi is always invariant starting from 2nd
iteration. After the 1st iteration is peeled, other optimizations can
potentially simplify calculations with this invariant.
Patch by Max Kazantsev!
Reviewers: sanjoy, apilipenko, igor-laevsky, anna, mkuper, reames
Reviewed By: mkuper
Subscribers: mkuper, mzolotukhin, llvm-commits
Differential Revision: https://reviews.llvm.org/D30161
llvm-svn: 296898
for VectorizeTree() API.This API uses it for proper mask computation to be used in shufflevector IR.
The fix is to compute the mask for out of order memory accesses while building the vectorizable tree
instead of actual vectorization of vectorizable tree.It also needs to recompute the proper Lane for
external use of vectorizable scalars based on shuffle mask.
Reviewers: mkuper
Differential Revision: https://reviews.llvm.org/D30159
Change-Id: Ide8773ce0ad3562f3cf4d1a0ad0f487e2f60ce5d
llvm-svn: 296863
Such edges may otherwise result in infinite recursion if a pointer to a vtable
is reachable from the vtable itself. This can happen in practice if a TU
defines the ABI types used to implement RTTI, and is itself compiled with RTTI.
Fixes PR32121.
llvm-svn: 296839
ValueTracking is used for more thorough analysis of operands. Based on the
analysis, either run-time checks can be simplified (e.g. check only one operand
instead of two) or the transformation can be avoided. For example, it is quite
often the case that a divisor is promoted from a shorter type and run-time
checks for it are redundant.
With additional compile-time analysis of values, two special cases naturally
arise and are addressed by the patch:
1) Both operands are known to be short enough. Then, the long division can be
simply replaced with a short one without CFG modification.
2) If a division is unsigned and the dividend is known to be short then the
long division is not needed at all. Because if the divisor is too big for
short division then the quotient is obviously zero (and the remainder is
equal to the dividend). Actually, the division is not needed when
(divisor > dividend).
Differential Revision: https://reviews.llvm.org/D29897
llvm-svn: 296832
and also "clang-format GenericDomTreeConstruction.h, since the current
formatting makes it look like their is a bug in the loop indentation, and there
is not"
This reverts commit r296535.
There are still some open design questions which I would like to discuss. I
revert this for Daniel (who gave the OK), as he is on vacation.
llvm-svn: 296812
Summary:
Extend -unroll-partial-threshold to 200 for runtime-loop3.ll test
as epilogue unroll initially add 1 more IV to the loop.
From: Evgeny Stupachenko <evstupac@gmail.com>
llvm-svn: 296803
This re-applies r289696, which caused TSan perf regression, which has
since been addressed in separate changes (see PR for details).
See PR31382.
llvm-svn: 296759
Summary:
When InstCombine is optimizing certain select-cmp-br patterns
it replaces the result of the select in uses outside of the
basic block containing the select. This is only legal if the
path from the select to the outside use is disjoint from all
other paths out from the originating basic block.
The problem found was that InstCombiner::replacedSelectWithOperand
did not consider the case when both edges out from the br pointed
to the same label. In that case the paths aren't disjoint and the
transformation is illegal. This patch avoids the faulty rewrites
by verifying that there is a single flow to the successor where
we want to replace uses.
Reviewers: llvm-commits, spatel, majnemer
Differential Revision: https://reviews.llvm.org/D30455
llvm-svn: 296752
After r296750, we're able to match interleaved accesses having types wider than
128 bits. This patch updates the associated TTI costs.
Differential Revision: https://reviews.llvm.org/D29675
llvm-svn: 296751
This patch teaches (ARM|AArch64)ISelLowering.cpp to match illegal vector types
to interleaved access intrinsics as long as the types are multiples of the
vector register width. A "wide" access will now be mapped to multiple
interleave intrinsics similar to the way in which non-interleaved accesses with
illegal types are legalized into multiple accesses. I'll update the associated
TTI costs (in getInterleavedMemoryOpCost) as a follow-on.
Differential Revision: https://reviews.llvm.org/D29466
llvm-svn: 296750
When computing the smallest and largest types for selecting the maximum
vectorization factor, we currently ignore loads and stores of pointer types if
the memory access is non-consecutive. We do this because such accesses must be
scalarized regardless of vectorization factor, and thus shouldn't be considered
when determining the factor. This patch makes this check less aggressive by
also considering non-consecutive accesses that may be vectorized, such as
interleaved accesses. Because we don't know at the time of the check if an
accesses will certainly be vectorized (this is a cost model decision given a
particular VF), we consider all accesses that can potentially be vectorized.
Differential Revision: https://reviews.llvm.org/D30305
llvm-svn: 296747
Now that terminators can be EH pads, this code needs to iterate over the
immediate dominators of the EH pad to find a valid insertion point.
Fix for PR32107
Patch by Robert Olliff!
Differential Revision: https://reviews.llvm.org/D30511
llvm-svn: 296698
Summary:
The SLP vectorizer should propagate IR-level optimization hints/flags
(nsw, nuw, exact, fast-math) when converting scalar horizontal
reductions instructions into vectors, just like for other vectorized
instructions.
It doe not include IR propagation for extra arguments, we need to handle
original scalar operations for extra args to propagate correct flags.
Reviewers: mkuper, mzolotukhin, hfinkel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30418
llvm-svn: 296614
Summary:
We should preserve IR flags for extra args. These IR flags should be
taken from original scalar operations, not from the reduction
operations.
Reviewers: mkuper, mzolotukhin, hfinkel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30447
llvm-svn: 296613
Summary:
If horizontal reduction tree starts from the binary operation that is
used in PHI node, but this PHI is not used in horizontal reduction, we
may end up with extra addition of this PHI node after vectorization.
Here is an example:
```
%phi = phi i32 [ %tmp, %end], ...
...
%tmp = add i32 %tmp1, %tmp2
end:
```
after vectorization we always have something like:
```
%phi = phi i32 [ %tmp, %end], ...
...
%red = extractelement <8 x 32> %vec.red, 0
%tmp = add i32 %red, %phi
end:
```
even if `%phi` is not used in reduction tree. Patch considers these PHI
nodes as extra arguments and considers them in the final result iff they
really used in reduction.
Reviewers: mkuper, hfinkel, mzolotukhin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30409
llvm-svn: 296606
Summary:
Solves PR 31990.
The bad rewrite could replace a memcpy of one word with
store i4 -1
while it should actually be
store i8 -1
Hopefully opt and llc has improved enough so the original optimization
done by the code isn't needed anymore.
One already existing testcase is affected. It originally tested that
the memcpy was replaced with
load double
but since we now remove that rewrite it will be
load i64
instead.
Patch suggestion by Eli Friedman.
Reviewers: eli.friedman, majnemer, efriedma
Reviewed By: efriedma
Subscribers: efriedma, llvm-commits
Differential Revision: https://reviews.llvm.org/D30254
llvm-svn: 296585
The practice in LV is that we emit analysis remarks and then finally report
either a missed or applied remark on the final decision whether vectorization
is taking place. On this code path, we were closing with an analysis remark.
llvm-svn: 296578
for VectorizeTree() API.This API uses it for proper mask computation to be used in shufflevector IR.
The fix is to compute the mask for out of order memory accesses while building the vectorizable tree
instead of actual vectorization of vectorizable tree.
Reviewers: mkuper
Differential Revision: https://reviews.llvm.org/D30159
Change-Id: Id1e287f073fa4959713ba545fa4254db5da8b40d
llvm-svn: 296575
Summary:
Currently, our post-dom tree tries to ignore and remove the effects of
infinite loops. It fails miserably at this, because it tries to do it
ahead of time, and thus can only detect self-loops, and any other type
of infinite loop, it pretends doesn't exist at all.
This can, in a bunch of cases, lead to wrong answers and a completely
empty post-dom tree.
Wrong answer:
```
declare void foo()
define internal void @f() {
entry:
br i1 undef, label %bb35, label %bb3.i
bb3.i:
call void @foo()
br label %bb3.i
bb35.loopexit3:
br label %bb35
bb35:
ret void
}
```
We get:
```
Inorder PostDominator Tree:
[1] <<exit node>> {0,7}
[2] %bb35 {1,6}
[3] %bb35.loopexit3 {2,3}
[3] %entry {4,5}
```
This is a trivial modification of the testcase for PR 6047
Note that we pretend bb3.i doesn't exist.
We also pretend that bb35 post-dominates entry.
While it's true that it does not exit in a theoretical sense, it's not
really helpful to try to ignore the effect and pretend that bb35
post-dominates entry. Worse, we pretend the infinite loop does
nothing (it's usually considered a side-effect), and doesn't even
exist, even when it calls a function. Sadly, this makes it impossible
to use when you are trying to move code safely. All compilers also
create virtual or real single exit nodes (including us), and connect
infinite loops there (which this patch does). In fact, others have
worked around our behavior here, to the point of building their own
post-dom trees:
https://zneak.github.io/fcd/2016/02/17/structuring.html and pointing
out the region infrastructure is near-useless for them with postdom in
this state :(
Completely empty post-dom tree:
```
define void @spam() #0 {
bb:
br label %bb1
bb1: ; preds = %bb1, %bb
br label %bb1
bb2: ; No predecessors!
ret void
}
```
Printing analysis 'Post-Dominator Tree Construction' for function 'foo':
=============================--------------------------------
Inorder PostDominator Tree:
[1] <<exit node>> {0,1}
:(
(note that even if you ignore the effects of infinite loops, bb2
should be present as an exit node that post-dominates nothing).
This patch changes post-dom to properly handle infinite loops and does
root finding during calculation to prevent empty tress in such cases.
We match gcc's (and the canonical theoretical) behavior for infinite
loops (find the backedge, connect it to the exit block).
Testcases coming as soon as i finish running this on a ton of random graphs :)
Reviewers: chandlerc, davide
Subscribers: bryant, llvm-commits
Differential Revision: https://reviews.llvm.org/D29705
llvm-svn: 296535
Summary: For SamplePGO, the profile may contain cross-module inline stacks. As we need to make sure the profile annotation happens when all the hot inline stacks are expanded, we need to pass this info to the module importer so that it can import proper functions if necessary. This patch implemented this feature by emitting cross-module targets as part of function entry metadata. In the module-summary phase, the metadata is used to build call edges that points to functions need to be imported.
Reviewers: mehdi_amini, tejohnson
Reviewed By: tejohnson
Subscribers: davidxl, llvm-commits
Differential Revision: https://reviews.llvm.org/D30053
llvm-svn: 296498
The LLVM backend cannot produce any debug info for an llvm::Function
without a DISubprogram attachment. When inlining a debug-info-carrying
function into a nodebug function, there is therefore no reason to keep
any debug info intrinsic calls or debug locations on the instructions.
This fixes a problem discovered in PR32042.
rdar://problem/30679307
llvm-svn: 296488
Splitting critical edges when one of the source edges is an indirectbr
is hard in general (because it requires changing the memory the indirectbr
reads). But if a block only has a single indirectbr predecessor (which is
the common case), we can simulate splitting that edge by splitting
the destination block, and retargeting the *direct* branches.
This is motivated by the use of computed gotos in python 2.7: PyEval_EvalFrame()
ends up using an indirect branch with ~100 successors, and passing a constant to
each of those. Since MachineSink can't break indirect critical edges on demand
(and doing this in MIR doesn't look feasible), this causes us to emit about ~100
defs of registers containing constants, which we in the predecessor block, where
only one of those constants is used in each successor. So, at each computed goto,
we needlessly spill about a 100 constants to stack. The end result is that a
clang-compiled python interpreter can be about ~2.5x slower on a simple python
reduction loop than a gcc-compiled interpreter.
Differential Revision: https://reviews.llvm.org/D29916
llvm-svn: 296416
This was suggested in D27855: have the inliner add assumptions, so we don't
lose nonnull info provided by argument attributes.
This still doesn't solve PR28430 (dyn_cast), but this gets us closer.
https://reviews.llvm.org/D29999
llvm-svn: 296366
This is a fix for a loop predication bug which resulted in malformed IR generation.
Loop invariant side of the widened condition is not guaranteed to be available in the preheader as is, so we need to expand it as well. See added unsigned_loop_0_to_n_hoist_length test for example.
Reviewed By: sanjoy, mkazantsev
Differential Revision: https://reviews.llvm.org/D30099
llvm-svn: 296345
Summary:
Previously we used to return a bogus result, 0, for IR like `ashr %val,
-1`.
I've also added an assert checking that `ComputeNumSignBits` at least
returns 1. That assert found an already checked in test case where we
were returning a bad result for `ashr %val, -1`.
Fixes PR32045.
Reviewers: spatel, majnemer
Reviewed By: spatel, majnemer
Subscribers: efriedma, mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D30311
llvm-svn: 296273
Current internal option -static-func-full-module-prefix keeps all the
directory path the profile counter names for static functions. The default
of this option is false. This strips the directory names from the source
filename which is problematic:
(1) it creates linker errors for profile-generation compilation, exposed in
our internal benchmarks. We are seeing messages like
"warning: relocation refers to discarded section".
This is due to the name conflicts after the stripping.
(2) the stripping only applies to getPGOFuncName.
Current Thin-LTO module importing for the indirect-calls assumes
the source directory name not being stripped. Current default value
for this option can potentially prevent some inter-module
indirect-call-promotions.
This patch turns the default value for -static-func-full-module-prefix to true.
The second part of the patch is to have an alternative implementation under
the internal option -static-func-strip-dirname-prefix=<value>
This options specifies level of directories to be stripped from the source
filename. Using a large value as the parameter has the same effect as
-static-func-full-module-prefix.
Differential Revision: http://reviews.llvm.org/D29512
llvm-svn: 296206
When we construct addressing modes, we use isNoopAddrSpaceCast to ignore
addrspacecast instructions. Make sure we insert the correct addrspacecast
when we reconstruct the addressing mode.
Differential Revision: https://reviews.llvm.org/D30114
llvm-svn: 296167
This optimisation was crashing when there was a chain of more than one bitcast
instruction to replace, as a result of the changes in D27283.
Patch by James Price.
Differential Revision: https://reviews.llvm.org/D30347
llvm-svn: 296163
Splitting critical edges when one of the source edges is an indirectbr
is hard in general (because it requires changing the memory the indirectbr
reads). But if a block only has a single indirectbr predecessor (which is
the common case), we can simulate splitting that edge by splitting
the destination block, and retargeting the *direct* branches.
This is motivated by the use of computed gotos in python 2.7: PyEval_EvalFrame()
ends up using an indirect branch with ~100 successors, and passing a constant to
each of those. Since MachineSink can't break indirect critical edges on demand
(and doing this in MIR doesn't look feasible), this causes us to emit about ~100
defs of registers containing constants, which we in the predecessor block, where
only one of those constants is used in each successor. So, at each computed goto,
we needlessly spill about a 100 constants to stack. The end result is that a
clang-compiled python interpreter can be about ~2.5x slower on a simple python
reduction loop than a gcc-compiled interpreter.
Differential Revision: https://reviews.llvm.org/D29916
llvm-svn: 296149
This patch merges the existing floating-point induction variable widening code
into the integer induction variable widening code, creating a single set of
functions for both kinds of inductions. The primary motivation for doing this
is to enable vector phi node creation for floating-point induction variables.
Differential Revision: https://reviews.llvm.org/D30211
llvm-svn: 296145
The Fuchsia ABI defines slots from the thread pointer where the
stack-guard value for stack-protector, and the unsafe stack pointer
for safe-stack, are stored. This parallels the Android ABI support.
Patch by Roland McGrath
Differential Revision: https://reviews.llvm.org/D30237
llvm-svn: 296081
Splitting critical edges when one of the source edges is an indirectbr
is hard in general (because it requires changing the memory the indirectbr
reads). But if a block only has a single indirectbr predecessor (which is
the common case), we can simulate splitting that edge by splitting
the destination block, and retargeting the *direct* branches.
This is motivated by the use of computed gotos in python 2.7: PyEval_EvalFrame()
ends up using an indirect branch with ~100 successors, and passing a constant to
each of those. Since MachineSink can't break indirect critical edges on demand
(and doing this in MIR doesn't look feasible), this causes us to emit about ~100
defs of registers containing constants, which we in the predecessor block, where
only one of those constants is used in each successor. So, at each computed goto,
we needlessly spill about a 100 constants to stack. The end result is that a
clang-compiled python interpreter can be about ~2.5x slower on a simple python
reduction loop than a gcc-compiled interpreter.
Differential Revision: https://reviews.llvm.org/D29916
llvm-svn: 296060
Summary: In case we do not know what the condition is in an unswitched loop, but we know its definitely NOT a known constant. We can perform simplifcations based on this information.
Reviewers: sanjoy, hfinkel, chenli, efriedma
Reviewed By: efriedma
Subscribers: david2050, llvm-commits, mzolotukhin
Differential Revision: https://reviews.llvm.org/D28968
llvm-svn: 296041
While not CVP's fault, this caused miscompiles (PR31181). Reverting
until those are resolved.
(This also reverts the follow-ups r288154 and r288161 which removed the
flag.)
llvm-svn: 296030
Summary: SamplePGO uses branch_weight annotation to represent callsite hotness. When ICP promotes an indirect call to direct call, we need to make sure the direct call is annotated with branch_weight in SamplePGO mode, so that downstream function inliner can use hot callsite heuristic.
Reviewers: davidxl, eraman, xur
Reviewed By: davidxl, xur
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D30282
llvm-svn: 296028
In OptimizeAdd, we scan the operand list to see if there are any common factors
between operands that can be factored out to reduce the number of multiplies
(e.g., 'A*A+A*B*C+D' -> 'A*(A+B*C)+D'). For each operand of the operand list, we
only consider unique factors (which is tracked by the Duplicate set). Now if we
find a factor that is a negative constant, we add the negated value as a factor
as well, because we can percolate the negate out. However, we mistakenly don't
add this negated constant to the Duplicates set.
Consider the expression A*2*-2 + B. Obviously, nothing to factor.
For the added value A*2*-2 we over count 2 as a factor without this change,
which causes the assert reported in PR30256. The problem is that this code is
assuming that all the multiply operands of the add are already reassociated.
This change avoids the issue by making OptimizeAdd tolerate multiplies which
haven't been completely optimized; this sort of works, but we're doing wasted
work: we'll end up revisiting the add later anyway.
Another possible approach would be to enforce RPO iteration order more strongly.
If we have RedoInsts, we process them immediately in RPO order, rather than
waiting until we've finished processing the whole function. Intuitively, it
seems like the natural approach: reassociation works on expression trees, so
the optimization only works in one direction. That said, I'm not sure how
practical that is given the current Reassociate; the "optimal" form for an
expression depends on its use list (see all the uses of "user_back()"), so
Reassociate is really an iterative optimization of sorts, so any changes here
would probably get messy.
PR30256
Differential Revision: https://reviews.llvm.org/D30228
llvm-svn: 296003
Summary: The discriminator has been encoded, and only the base discriminator should be used during profile matching.
Reviewers: dblaikie, davidxl
Reviewed By: dblaikie, davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30218
llvm-svn: 295999
result
Summary:
If the same value is used several times as an extra value, SLP
vectorizer takes it into account only once instead of actual number of
using.
For example:
```
int val = 1;
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 8; x++) {
val = val + input[y * 8 + x] + 3;
}
}
```
We have 2 extra rguments: `1` - initial value of horizontal reduction
and `3`, which is added 8*8 times to the reduction. Before the patch we
added `1` to the reduction value and added once `3`, though it must be
added 64 times.
Reviewers: mkuper, mzolotukhin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30262
llvm-svn: 295972
result
Summary:
If the same value is used several times as an extra value, SLP
vectorizer takes it into account only once instead of actual number of
using.
For example:
```
int val = 1;
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 8; x++) {
val = val + input[y * 8 + x] + 3;
}
}
```
We have 2 extra rguments: `1` - initial value of horizontal reduction
and `3`, which is added 8*8 times to the reduction. Before the patch we
added `1` to the reduction value and added once `3`, though it must be
added 64 times.
Reviewers: mkuper, mzolotukhin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30262
llvm-svn: 295956
result
Summary:
If the same value is used several times as an extra value, SLP
vectorizer takes it into account only once instead of actual number of
using.
For example:
```
int val = 1;
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 8; x++) {
val = val + input[y * 8 + x] + 3;
}
}
```
We have 2 extra rguments: `1` - initial value of horizontal reduction
and `3`, which is added 8*8 times to the reduction. Before the patch we
added `1` to the reduction value and added once `3`, though it must be
added 64 times.
Reviewers: mkuper, mzolotukhin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30262
llvm-svn: 295949
Implement isLegalToVectorizeLoadChain for AMDGPU to avoid
producing private address spaces accesses that will need to be
split up later. This was doing the wrong thing in the case
where the queried chain was an even number of elements.
A possible <4 x i32> store was being split into
store <2 x i32>
store i32
store i32
rather than
store <2 x i32>
store <2 x i32>
when legal.
llvm-svn: 295933
Summary:
Depends on D29606 and D29682
Makes us pass GVN's edge.ll (we also will pass a few other testcases
they just need cleaning up).
Thoughts on the Predicate* hiearchy of classes especially welcome :)
(it's not clear to me how best to organize it, and currently, the getBlock* seems ... uglier than maybe wasting a field somewhere or something).
Reviewers: davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29747
llvm-svn: 295889
After rL294814, LSR formula can have multiple SCEVAddRecExprs inside of its BaseRegs.
Previous canonicalization will swap the first SCEVAddRecExpr in BaseRegs with ScaledReg.
But now we want to swap the SCEVAddRecExpr Reg related with current loop with ScaledReg.
Otherwise, we may generate code like this: RegA + lsr.iv + RegB, where loop invariant
parts RegA and RegB are not grouped together and cannot be promoted outside of loop.
With this patch, it will ensure lsr.iv to be generated later in the expr:
RegA + RegB + lsr.iv, so that RegA + RegB can be promoted outside of loop.
Differential Revision: https://reviews.llvm.org/D26781
llvm-svn: 295884
Summary:
If the same value is used several times as an extra value, SLP
vectorizer takes it into account only once instead of actual number of
using.
For example:
```
int val = 1;
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 8; x++) {
val = val + input[y * 8 + x] + 3;
}
}
```
We have 2 extra rguments: `1` - initial value of horizontal reduction
and `3`, which is added 8*8 times to the reduction. Before the patch we
added `1` to the reduction value and added once `3`, though it must be
added 64 times.
Reviewers: mkuper, mzolotukhin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30262
llvm-svn: 295868
Prevent memory objects of different address spaces to be part of
the same load/store groups when analysing interleaved accesses.
This is fixing pr31900.
Reviewers: HaoLiu, mssimpso, mkuper
Reviewed By: mssimpso, mkuper
Subscribers: llvm-commits, efriedma, mzolotukhin
Differential Revision: https://reviews.llvm.org/D29717
This reverts r295042 (re-applies r295038) with an additional fix for the
buildbot problem.
llvm-svn: 295858
Summary: The CallTargetProfile should be added to FProfile to be consistent with other profile readers.
Reviewers: dnovillo, davidxl
Reviewed By: davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30233
llvm-svn: 295852
This enables peeling of loops with low dynamic iteration count by default,
when profile information is available.
Differential Revision: https://reviews.llvm.org/D27734
llvm-svn: 295796
This is part of trying to clean up our handling of min/max patterns in IR.
By converting these to canonical form, we're more likely to recognize them
because there are various places in InstCombine that don't use
matchSelectPattern or m_SMax and friends.
The backend fixups referenced in the now deleted TODO comment were added with:
https://reviews.llvm.org/rL291392https://reviews.llvm.org/rL289738
If there's any codegen fallout from this change, we should be able to address
it in DAGCombiner or target-specific lowering.
llvm-svn: 295758
Summary:
This is a fix for assertion failure in
`getInverseMinMaxSelectPattern` when ABS is passed in as a select pattern.
We should not be invoking the simplification rule for
ABS(MIN(~ x,y))) or ABS(MAX(~x,y)) combinations.
Added a test case which would cause an assertion failure without the patch.
Reviewers: sanjoy, majnemer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30051
llvm-svn: 295719
The new method introduced under "-lsr-exp-narrow" option (currenlty set to true).
Summary:
The method is based on registers number mathematical expectation and should be
generally closer to optimal solution.
Please see details in comments to
"LSRInstance::NarrowSearchSpaceByDeletingCostlyFormulas()" function
(in lib/Transforms/Scalar/LoopStrengthReduce.cpp).
Reviewers: qcolombet
Differential Revision: http://reviews.llvm.org/D29862
From: Evgeny Stupachenko <evstupac@gmail.com>
llvm-svn: 295704
Summary: This begins using the predicateinfo pass in NewGVN.
Reviewers: davide
Subscribers: llvm-commits, Prazek
Differential Revision: https://reviews.llvm.org/D29682
llvm-svn: 295583
Changing to 'or' (rather than 'xor' when no wrapping flags are set)
allows icmp simplifies to happen as expected.
Differential Revision: https://reviews.llvm.org/D29729
llvm-svn: 295574
The change to InstCombine in:
https://reviews.llvm.org/D29729
...exposes this missing fold in InstSimplify, so adding this
first to avoid a regression.
llvm-svn: 295573
A line number doesn't make much sense if you don't say where it's
from. Add a verifier check for this and update some tests that had
bogus debug info.
llvm-svn: 295516
A future change will cause this byte offset to be inttoptr'd and then exported
via an absolute symbol. On the importing end we will expect the symbol to be
in range [0,2^32) so that it will fit into a 32-bit relocation. The problem
is that on 64-bit architectures if the offset is negative it will not be in
the correct range once we inttoptr it.
This change causes us to use a 32-bit integer so that it can be inttoptr'd
(which zero extends) into the correct range.
Differential Revision: https://reviews.llvm.org/D30016
llvm-svn: 295487
We previously only created a vector phi node for an induction variable if its
step had a constant integer type. However, the step actually only needs to be
loop-invariant. We only handle inductions having loop-invariant steps, so this
patch should enable vector phi node creation for all integer induction
variables that will be vectorized.
Differential Revision: https://reviews.llvm.org/D29956
llvm-svn: 295456
Summary:
JumpThreading for guards feature has been reverted at https://reviews.llvm.org/rL295200
due to the following problem: the feature used the following algorithm for detection of
diamond patters:
1. Find a block with 2 predecessors;
2. Check that these blocks have a common single parent;
3. Check that the parent's terminator is a branch instruction.
The problem is that these checks are insufficient. They may pass for a non-diamond
construction in case if those two predecessors are actually the same block. This may
happen if parent's terminator is a br (either conditional or unconditional) to a block
that ends with "switch" instruction with exactly two branches going to one block.
This patch re-enables the JumpThreading for guards and fixes this issue by adding the
check that those found predecessors are actually different blocks. This guarantees that
parent's terminator is a conditional branch with exactly 2 different successors, which
is now ensured by assertions. It also adds two more tests for this situation (with parent's
terminator being a conditional and an unconditional branch).
Patch by Max Kazantsev!
Reviewers: anna, sanjoy, reames
Reviewed By: sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30036
llvm-svn: 295410
In rL294814, we allow formula with SCEVAddRecExpr type of Reg from loops
other than current loop. This is good for the case when induction variable
of outerloop being used in expr in innerloop. But it is very bad to allow
such Reg from sibling loop because we may need to add lsr.iv in other sibling
loops when scev expanding those SCEVAddRecExpr type exprs. For the testcase
below, one loop can be inserted with a bunch of lsr.iv because of LSR for
other loops.
// The induction variable j from a loop in the middle will have initial
// value generated from previous sibling loop and exit value used by its
// next sibling loop.
void goo(long i, long j);
long cond;
void foo(long N) {
long i = 0;
long j = 0;
i = 0; do { goo(i, j); i++; j++; } while (cond);
i = 0; do { goo(i, j); i++; j++; } while (cond);
i = 0; do { goo(i, j); i++; j++; } while (cond);
i = 0; do { goo(i, j); i++; j++; } while (cond);
i = 0; do { goo(i, j); i++; j++; } while (cond);
i = 0; do { goo(i, j); i++; j++; } while (cond);
}
The fix is to only allow formula with SCEVAddRecExpr type of Reg from current
loop or its parents.
Differential Revision: https://reviews.llvm.org/D30021
llvm-svn: 295378
This is a short term solution to the problem that many passes currently fail
to update the assumption cache. In the long term the verifier should not
be controllable with a flag. We should either fix all passes to correctly
update the assumption cache and enable the verifier unconditionally or
somehow arrange for the assumption list to be updated automatically by passes.
Differential Revision: https://reviews.llvm.org/D30003
llvm-svn: 295236
Multiple blocks in the callee can be mapped to a single cloned block
since we prune the callee as we clone it. The existing code
iterates over the value map and clones the block frequency (and
eventually scales the frequencies of the cloned blocks). Value map's
iteration is not deterministic and so the cloned block might get the
frequency of any of the original blocks. The fix is to set the max of
the original frequencies to the cloned block. The first block in the
sequence must have this max frequency and, in the call context,
subsequent blocks must have its frequency.
Differential Revision: https://reviews.llvm.org/D29696
llvm-svn: 295115
Group calls into constant and non-constant arguments up front, and use uint64_t
instead of ConstantInt to represent constant arguments. The goal is to allow
the information from the summary to fit naturally into this data structure in
a future change (specifically, it will be added to CallSiteInfo).
This has two side effects:
- We disallow VCP for constant integer arguments of width >64 bits.
- We remove the restriction that the bitwidth of a vcall's argument and return
types must match those of the vfunc definitions.
I don't expect either of these to matter in practice. The first case is
uncommon, and the second one will lead to UB (so we can do anything we like).
Differential Revision: https://reviews.llvm.org/D29744
llvm-svn: 295110
Summary:
When setting debugloc for instructions created in SplitBlockPredecessors, current implementation copies debugloc from the first-non-phi instruction of the original basic block. However, if the first-non-phi instruction is a call for @llvm.dbg.value, the debugloc of the instruction may point the location outside of the block itself. For the example code of
```
1 typedef struct _node_t {
2 struct _node_t *next;
3 } node_t;
4
5 extern node_t *root;
6
7 int foo() {
8 node_t *node, *tmp;
9 int ret = 0;
10
11 node = tmp = root->next;
12 while (node != root) {
13 while (node) {
14 tmp = node;
15 node = node->next;
16 ret++;
17 }
18 }
19
20 return ret;
21 }
```
, below is the basicblock corresponding to line 12 after Reassociate expressions pass:
```
while.cond: ; preds = %while.cond2, %entry
%node.0 = phi %struct._node_t* [ %1, %entry ], [ null, %while.cond2 ]
%ret.0 = phi i32 [ 0, %entry ], [ %ret.1, %while.cond2 ]
tail call void @llvm.dbg.value(metadata i32 %ret.0, i64 0, metadata !19, metadata !20), !dbg !21
tail call void @llvm.dbg.value(metadata %struct._node_t* %node.0, i64 0, metadata !11, metadata !20), !dbg !31
%cmp = icmp eq %struct._node_t* %node.0, %0, !dbg !33
br i1 %cmp, label %while.end5, label %while.cond2, !dbg !35
```
As you can see, the first-non-phi instruction is a call for @llvm.dbg.value, and the debugloc is
```
!21 = !DILocation(line: 9, column: 7, scope: !6)
```
, which is a definition of 'ret' variable and outside of the scope of the basicblock itself. However, current implementation picks up this debugloc for the instructions created in SplitBlockPredecessors. This patch addresses this problem by picking up debugloc from the first-non-phi-non-dbg instruction.
Reviewers: dblaikie, samsonov, eugenis
Reviewed By: eugenis
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29867
llvm-svn: 295106
This reverts 295092 (re-applies 295084), with a fix for dangling
references from the array of coverage names passed down from frontends.
I missed this in my initial testing because I only checked test/Profile,
and not test/CoverageMapping as well.
Original commit message:
The profile name variables passed to counter increment intrinsics are dead
after we emit the finalized name data in __llvm_prf_nm. However, we neglect to
erase these name variables. This causes huge size increases in the
__TEXT,__const section as well as slowdowns when linker dead stripping is
disabled. Some affected projects are so massive that they fail to link on
Darwin, because only the small code model is supported.
Fix the issue by throwing away the name constants as soon as we're done with
them.
Differential Revision: https://reviews.llvm.org/D29921
llvm-svn: 295099
The profile name variables passed to counter increment intrinsics are
dead after we emit the finalized name data in __llvm_prf_nm. However, we
neglect to erase these name variables. This causes huge size increases
in the __TEXT,__const section as well as slowdowns when linker dead
stripping is disabled. Some affected projects are so massive that they
fail to link on Darwin, because only the small code model is supported.
Fix the issue by throwing away the name constants as soon as we're done
with them.
Differential Revision: https://reviews.llvm.org/D29921
llvm-svn: 295084
Summary:
As written in the comments above, LastCallToStaticBonus is already applied to
the cost if Caller has only one user, so it is redundant to reapply the bonus
here.
If the only user is not a caller, TotalSecondaryCost will not be adjusted
anyway because callerWillBeRemoved is false. If there's no caller at all, we
don't need to care about TotalSecondaryCost because
inliningPreventsSomeOuterInline is false.
Reviewers: chandlerc, eraman
Reviewed By: eraman
Subscribers: haicheng, davidxl, davide, llvm-commits, mehdi_amini
Differential Revision: https://reviews.llvm.org/D29169
llvm-svn: 295075
This reapplies commit r294967 with a fix for the execution time regressions
caught by the clang-cmake-aarch64-quick bot. We now extend the truncate
optimization to non-primary induction variables only if the truncate isn't
already free.
Differential Revision: https://reviews.llvm.org/D29847
llvm-svn: 295063
back into a vector
Previously the cost of the existing ExtractElement/ExtractValue
instructions was considered as a dead cost only if it was detected that
they have only one use. But these instructions may be considered
dead also if users of the instructions are also going to be vectorized,
like:
```
%x0 = extractelement <2 x float> %x, i32 0
%x1 = extractelement <2 x float> %x, i32 1
%x0x0 = fmul float %x0, %x0
%x1x1 = fmul float %x1, %x1
%add = fadd float %x0x0, %x1x1
```
This can be transformed to
```
%1 = fmul <2 x float> %x, %x
%2 = extractelement <2 x float> %1, i32 0
%3 = extractelement <2 x float> %1, i32 1
%add = fadd float %2, %3
```
because though `%x0` and `%x1` have 2 users each other, these users are
part of the vectorized tree and we can consider these `extractelement`
instructions as dead.
Differential Revision: https://reviews.llvm.org/D29900
llvm-svn: 295056
Prevent memory objects of different address spaces to be part of
the same load/store groups when analysing interleaved accesses.
This is fixing pr31900.
Reviewers: HaoLiu, mssimpso, mkuper
Reviewed By: mssimpso, mkuper
Subscribers: llvm-commits, efriedma, mzolotukhin
Differential Revision: https://reviews.llvm.org/D29717
llvm-svn: 295038
Summary:
Function isCompatibleIVType is already used as a guard before the call to
SE.getMinusSCEV(OperExpr, PrevExpr);
in LSRInstance::ChainInstruction. getMinusSCEV requires the expressions
to be of the same type, so we now consider two pointers with different
address spaces to be incompatible, since it is possible that the pointers
in fact have different sizes.
Reviewers: qcolombet, eli.friedman
Reviewed By: qcolombet
Subscribers: nhaehnle, Ka-Ka, llvm-commits, mzolotukhin
Differential Revision: https://reviews.llvm.org/D29885
llvm-svn: 295033
Extend our store promotion code to deal with unordered atomic accesses. Ordered atomics continue to be unhandled.
Most of the change is straight-forward, the only complicated bit is in the reasoning around mixing of atomic and non-atomic memory access. Rather than trying to reason about the complex semantics in these cases, I simply disallowed promotion when both atomic and non-atomic accesses are present. This is conservatively correct.
It seems really tempting to just promote all access to atomics, but the original accesses might have been conditional. Since we can't lower an arbitrary atomic type, it might not be safe to promote all access to atomic. Consider a loop like the following:
while(b) {
load i128 ...
if (can lower i128 atomic)
store atomic i128 ...
else
store i128
}
It could be there's no race on the location and thus the code is perfectly well defined even if we can't lower a i128 atomically.
It's not clear we need to be this conservative - arguably the program above is brocken since it can't be lowered unless the branch is folded - but I didn't want to have to fix any fallout which might result.
Differential Revision: https://reviews.llvm.org/D15592
llvm-svn: 295015
Make the whole thing testable by adding YAML I/O support for the WPD
summary information and adding some negative tests that exercise the
YAML support.
Differential Revision: https://reviews.llvm.org/D29782
llvm-svn: 294981
This reverts commit r294967. This patch caused execution time slowdowns in a
few LLVM test-suite tests, as reported by the clang-cmake-aarch64-quick bot.
I'm reverting to investigate.
llvm-svn: 294973
This patch extends the optimization of truncations whose operand is an
induction variable with a constant integer step. Previously we were only
applying this optimization to the primary induction variable. However, the cost
model assumes the optimization is applied to the truncation of all integer
induction variables (even regardless of step type). The transformation is now
applied to the other induction variables, and I've updated the cost model to
ensure it is better in sync with the transformation we actually perform.
Differential Revision: https://reviews.llvm.org/D29847
llvm-svn: 294967
reductions.
Currently, LLVM supports vectorization of horizontal reduction
instructions with initial value set to 0. Patch supports vectorization
of reduction with non-zero initial values. Also, it supports a
vectorization of instructions with some extra arguments, like:
```
float f(float x[], int a, int b) {
float p = a % b;
p += x[0] + 3;
for (int i = 1; i < 32; i++)
p += x[i];
return p;
}
```
Patch allows vectorization of this kind of horizontal reductions.
Differential Revision: https://reviews.llvm.org/D29727
llvm-svn: 294934
Summary:
This adds support for placing predicateinfo such that it affects critical edges.
This fixes the issues mentioned by Nuno on the mailing list.
Depends on D29519
Reviewers: davide, nlopes
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29606
llvm-svn: 294921
proven larger than the loop-count
This fixes PR31098: Try to resolve statically data-dependences whose
compile-time-unknown distance can be proven larger than the loop-count,
instead of resorting to runtime dependence checking (which are not always
possible).
For vectorization it is sufficient to prove that the dependence distance
is >= VF; But in some cases we can prune unknown dependence distances early,
and even before selecting the VF, and without a runtime test, by comparing
the distance against the loop iteration count. Since the vectorized code
will be executed only if LoopCount >= VF, proving distance >= LoopCount
also guarantees that distance >= VF. This check is also equivalent to the
Strong SIV Test.
Reviewers: mkuper, anemet, sanjoy
Differential Revision: https://reviews.llvm.org/D28044
llvm-svn: 294892
it is dead or unreachable, as it should be.
This also makes the leader of INITIAL undef, enabling us to handle
irreducibility properly.
Summary:
This lets us verify, more than we do now, that we didn't screw up
value numbering.
Reviewers: davide
Subscribers: Prazek, llvm-commits
Differential Revision: https://reviews.llvm.org/D29842
llvm-svn: 294844
Summary:
The patch adds instructions number generated by a solution
to LSR cost under "-lsr-insns-cost" option.
Reviewers: qcolombet, hfinkel
Differential Revision: http://reviews.llvm.org/D28307
From: Evgeny Stupachenko <evstupac@gmail.com>
llvm-svn: 294821
There are no vldN/vstN f16 variants, even with +fullfp16.
We could use the i16 variants, but, in practice, even with +fullfp16,
the f16 sequence leading to the i16 shuffle usually gets scalarized.
We'd need to improve our support for f16 codegen before getting there.
Teach the cost model to consider f16 interleaved operations as
expensive. Otherwise, we are all but guaranteed to end up with
a large block of scalarized vector code.
llvm-svn: 294819
There are no vldN/vstN f16 variants, even with +fullfp16.
We could use the i16 variants, but, in practice, even with +fullfp16,
the f16 sequence leading to the i16 shuffle usually gets scalarized.
We'd need to improve our support for f16 codegen before getting there.
Reject f16 interleaved accesses. If we try to emit the f16 intrinsics,
we'll just end up with a selection failure.
llvm-svn: 294818
The recommit includes some changes of testcases. No functional change to the patch.
In RateRegister of existing LSR, if a formula contains a Reg which is a SCEVAddRecExpr,
and this SCEVAddRecExpr's loop is an outerloop, the formula will be marked as Loser
and dropped.
Suppose we have an IR that %for.body is outerloop and %for.body2 is innerloop. LSR only
handle inner loop now so only %for.body2 will be handled.
Using the logic above, formula like
reg(%array) + reg({1,+, %size}<%for.body>) + 1*reg({0,+,1}<%for.body2>) will be dropped
no matter what because reg({1,+, %size}<%for.body>) is a SCEVAddRecExpr type reg related
with outerloop. Only formula like
reg(%array) + 1*reg({{1,+, %size}<%for.body>,+,1}<nuw><nsw><%for.body2>) will be kept
because the SCEVAddRecExpr related with outerloop is folded into the initial value of the
SCEVAddRecExpr related with current loop.
But in some cases, we do need to share the basic induction variable
reg{0 ,+, 1}<%for.body2> among LSR Uses to reduce the final total number of induction
variables used by LSR, so we don't want to drop the formula like
reg(%array) + reg({1,+, %size}<%for.body>) + 1*reg({0,+,1}<%for.body2>) unconditionally.
From the existing comment, it tries to avoid considering multiple level loops at the same time.
However, existing LSR only handles innermost loop, so for any SCEVAddRecExpr with a loop other
than current loop, it is an invariant and will be simple to handle, and the formula doesn't have
to be dropped.
Differential Revision: https://reviews.llvm.org/D26429
llvm-svn: 294814
For function-scope variables with large initialisation list, FE usually
generates a global variable to hold the initializer, then generates
memcpy intrinsic to initialize the alloca. InstCombiner::visitAllocaInst
identifies such allocas which are accessed only by reading and replaces
them with the global variable. This is done by casting the global variable
to the type of the alloca and replacing all references.
However, when the global variable is in a different address space which
is disjoint with addr space 0 (e.g. for IR generated from OpenCL,
global variable cannot be in private addr space i.e. addr space 0), casting
the global variable to addr space 0 results in invalid IR for certain
targets (e.g. amdgpu).
To fix this issue, when the global variable is not in addr space 0,
instead of casting it to addr space 0, this patch chases down the uses
of alloca until reaching the load instructions, then replaces load from
alloca with load from the global variable. If during the chasing
bitcast and GEP are encountered, new bitcast and GEP based on the global
variable are generated and used in the load instructions.
Differential Revision: https://reviews.llvm.org/D27283
llvm-svn: 294786
Summary:
This patch starts the implementation as discuss in the following RFC: http://lists.llvm.org/pipermail/llvm-dev/2016-October/106532.html
When optimization duplicates code that will scale down the execution count of a basic block, we will record the duplication factor as part of discriminator so that the offline process tool can find the duplication factor and collect the accurate execution frequency of the corresponding source code. Two important optimization that fall into this category is loop vectorization and loop unroll. This patch records the duplication factor for these 2 optimizations.
The recording will be guarded by a flag encode-duplication-in-discriminators, which is off by default.
Reviewers: probinson, aprantl, davidxl, hfinkel, echristo
Reviewed By: hfinkel
Subscribers: mehdi_amini, anemet, mzolotukhin, llvm-commits
Differential Revision: https://reviews.llvm.org/D26420
llvm-svn: 294782
We previously only created a vector phi node for an induction variable if its
type matched the type of the canonical induction variable.
Differential Revision: https://reviews.llvm.org/D29776
llvm-svn: 294755
Chandler mentioned at the last social that the need for BFI in the new pass manager was causing a slight hiccup for this pass. Given this code has been checked in, but off for over a year, it makes sense to just remove it for now.
Note that there's nothing wrong with the general idea - it's actually a quite good one - and once we have the infrastructure in place to implement this without the full recompuation on every loop, we absolutely should.
llvm-svn: 294715
Now that the call graph supports efficient replacement of a function and
spurious reference edges, we can port ArgumentPromotion to the new pass
manager very easily.
The old PM-specific bits are sunk into callbacks that the new PM simply
doesn't use. Unlike the old PM, the new PM simply does argument
promotion and afterward does the update to LCG reflecting the promoted
function.
Differential Revision: https://reviews.llvm.org/D29580
llvm-svn: 294667
This fold already existed for vectors but only when 'C1' was a splat
constant (but 'C2' could be any constant).
There were no tests for any vector constants, so I'm adding a test
that shows non-splat constants for both operands.
llvm-svn: 294650
Summary:
This patch allows JumpThreading also thread through guards.
Virtually, guard(cond) is equivalent to the following construction:
if (cond) { do something } else {deoptimize}
Yet it is not explicitly converted into IFs before lowering.
This patch enables early threading through guards in simple cases.
Currently it covers the following situation:
if (cond1) {
// code A
} else {
// code B
}
// code C
guard(cond2)
// code D
If there is implication cond1 => cond2 or !cond1 => cond2, we can transform
this construction into the following:
if (cond1) {
// code A
// code C
} else {
// code B
// code C
guard(cond2)
}
// code D
Thus, removing the guard from one of execution branches.
Patch by Max Kazantsev!
Reviewers: reames, apilipenko, igor-laevsky, anna, sanjoy
Reviewed By: sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29620
llvm-svn: 294617
It turns out that some of our negative tests were not in fact providing the
test coverage we expected: they were passing because the vtables were failing
an early check that they were constant. Fix this by changing the globals in
these tests to constants.
llvm-svn: 294550
This module will contain nothing but vtable definitions and (soon)
available_externally function definitions, so there is no point in keeping
debug info in the module.
Differential Revision: https://reviews.llvm.org/D28913
llvm-svn: 294511
Making the cost model selecting between Interleave, GatherScatter or Scalar vectorization form of memory instruction.
The right decision should be done for non-consecutive memory access instrcuctions that may have more than one vectorization solution.
This patch includes the following changes:
- Cost Model calculates the cost of Load/Store vector form and choose the better option between Widening, Interleave, GatherScactter and Scalarization. Cost Model keeps the widening decision.
- Arrays of Uniform and Scalar values are moved from Legality to Cost Model.
- Cost Model collects Uniforms and Scalars per VF. The collection is based on CM decision map of Loadis/Stores vectorization form.
- Vectorization of memory instruction is performed according to the CM decision.
Differential Revision: https://reviews.llvm.org/D27919
llvm-svn: 294503
This test is under 'ArgumentPromotion' but there are no arguments that
get promoted in the test case, so there seems to be no point. Also,
there are no assertions about the output at all, so this seems like
something we should just delete given the low value.
llvm-svn: 294428
renaming things to at least have somewhat spelled out names, and even
have meaningful names where I could guess at what they should be.
Also add FileCheck assertions that we're actually doing what we set out
to do for some of the tests, for example not promoting a type that would
result in infinite promotion.
llvm-svn: 294426
Currently IRCE relies on the loops it transforms to be (semantically) of
the form:
for (i = START; i < END; i++)
...
or
for (i = START; i > END; i--)
...
However, we were not verifying the presence of the START < END entry
check (i.e. check before the first iteration). We were only verifying
that the backedge was guarded by (i + 1) < END.
Usually this would work "fine" since (especially in Java) most loops do
actually have the START < END check, but of course that is not
guaranteed.
llvm-svn: 294375
Summary:
This patch adds a utility to build extended SSA (see "ABCD: eliminating
array bounds checks on demand"), and an intrinsic to support it. This
is then used to get functionality equivalent to propagateEquality in
GVN, in NewGVN (without having to replace instructions as we go). It
would work similarly in SCCP or other passes. This has been talked
about a few times, so i built a real implementation and tried to
productionize it.
Copies are inserted for operands used in assumes and conditional
branches that are based on comparisons (see below for more)
Every use affected by the predicate is renamed to the appropriate
intrinsic result.
E.g.
%cmp = icmp eq i32 %x, 50
br i1 %cmp, label %true, label %false
true:
ret i32 %x
false:
ret i32 1
will become
%cmp = icmp eq i32, %x, 50
br i1 %cmp, label %true, label %false
true:
; Has predicate info
; branch predicate info { TrueEdge: 1 Comparison: %cmp = icmp eq i32 %x, 50 }
%x.0 = call @llvm.ssa_copy.i32(i32 %x)
ret i32 %x.0
false:
ret i23 1
(you can use -print-predicateinfo to get an annotated-with-predicateinfo dump)
This enables us to easily determine what operations are affected by a
given predicate, and how operations affected by a chain of
predicates.
Reviewers: davide, sanjoy
Subscribers: mgorny, llvm-commits, Prazek
Differential Revision: https://reviews.llvm.org/D29519
Update for review comments
Fix a bug Nuno noticed where we are giving information about and/or on edges where the info is not useful and easy to use wrong
Update for review comments
llvm-svn: 294351
This patch removes unneeded instructions from the existing ARM/AArch64
interleaved access cost model tests. I'll be adding a similar set of tests in a
follow-on patch to increase coverage.
llvm-svn: 294336
This reverts commit r294250. It caused PR31891.
Add a test case that shows that inlinable calls retain location
information with an accurate scope.
llvm-svn: 294317
Summary: Checking CS.getCalledFunction() == nullptr does not necessary indicate indirect call. We also need to check if CS.getCalledValue() is not a constant.
Reviewers: davidxl
Reviewed By: davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29570
llvm-svn: 294260
This breaks when one of the extra values is also a scalar that
participates in the same vectorization tree which we'll end up
reducing.
llvm-svn: 294245
In ValueMapper we create new operands for MDNodes and
rely on MDNode::replaceWithUniqued to create a new MDNode
with the specified operands. However this doesn't always
actually happen correctly for DISubprograms because when we
uniquify the new node, we only odr-compare it with existing nodes
(MDNodeSubsetEqualImpl<DISubprogram>::isDeclarationOfODRMember). Although
the TemplateParameters field can refer to a distinct DICompileUnit via
DITemplateTypeParameter::type -> DICompositeType::scope -> DISubprogram::unit,
it is not currently included in the odr comparison. As a result, we can end
up getting our original DISubprogram back, which means we will have a cloned
module referring to the DICompileUnit in the original module, which causes
a verification error.
The fix I implemented was to consider TemplateParameters to be one of the
odr-equal properties. But I'm a little uncomfortable with this. In general it
seems unsound to rely on distinct MDNodes never being reachable from nodes
which we only check odr-equality of. My only long term suggestion would be
to separate odr-uniquing from full uniquing.
Differential Revision: https://reviews.llvm.org/D29240
llvm-svn: 294240
Summary: When type casting of the return value is needed, promoteIndirectCall will return the type casting instruction instead of the direct call. This patch changed to return the direct call instruction instead.
Reviewers: davidxl
Reviewed By: davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29569
llvm-svn: 294205
tests.
This also removes the use of instcombine as we can max the patterns
produced by argument promotion directly with the more powerful tools in
FileCheck.
llvm-svn: 294174
This patch is based on the llvm-dev discussion here:
http://lists.llvm.org/pipermail/llvm-dev/2017-January/109631.html
Folding to i1 should always be desirable because that's better for value tracking
and we have special folds for i1 types.
I checked for other users of shouldChangeType() where this might have an effect,
but we already handle the i1 case differently than other types in all of those cases.
Side note: the default datalayout includes i1, so it seems we only find this gap in
shouldChangeType + phi folding for the case when there is (1) an explicit datalayout
without i1, (2) casting to i1 from a legal type, and (3) a phi with exactly 2 incoming
casted operands (as Björn mentioned).
Differential Revision: https://reviews.llvm.org/D29336
llvm-svn: 294066
The code comments didn't match the code logic, and we didn't actually distinguish the fake unary (not/neg/fneg)
operators from arguments. Adding another level to the weighting scheme provides more structure and can help
simplify the pattern matching in InstCombine and other places.
I fixed regressions that would have shown up from this change in:
rL290067
rL290127
But that doesn't mean there are no pattern-matching logic holes left; some combines may just be missing regression tests.
Should fix:
https://llvm.org/bugs/show_bug.cgi?id=28296
Differential Revision: https://reviews.llvm.org/D27933
llvm-svn: 294049
This generalizes memory access sorting to use differences between SCEVs,
instead of relying on constant offsets. That allows us to properly do
SLP vectorization of non-sequentially ordered loads within loops bodies.
Differential Revision: https://reviews.llvm.org/D29425
llvm-svn: 294027
Currently LLVM supports vectorization of horizontal reduction
instructions with initial value set to 0. Patch supports vectorization
of reduction with non-zero initial values. Also it supports a
vectorization of instructions with some extra arguments, like:
float f(float x[], int a, int b) {
float p = a % b;
p += x[0] + 3;
for (int i = 1; i < 32; i++)
p += x[i];
return p;
}
Patch allows vectorization of this kind of horizontal reductions.
Differential Revision: https://reviews.llvm.org/D28961
llvm-svn: 293994
Exit loop analysis early if suitable private access found.
Do not account for GEPs which are invariant to loop induction variable.
Do not account for Allocas which are too big to fit into register file anyway.
Add option for tuning: -amdgpu-unroll-threshold-private.
Differential Revision: https://reviews.llvm.org/D29473
llvm-svn: 293991
Summary: While scanning predecessors to find an available loaded value, if the predecessor has a single predecessor, we can continue scanning through the single predecessor.
Reviewers: mcrosier, rengolin, reames, davidxl, haicheng
Reviewed By: rengolin
Subscribers: zzheng, llvm-commits
Differential Revision: https://reviews.llvm.org/D29200
llvm-svn: 293896
Summary:
We can hoist out loads that are dominated by invariant.start, to the preheader.
We conservatively assume the load is variant, if we see a corresponding
use of invariant.start (it could be an invariant.end or an escaping
call).
Reviewers: mkuper, sanjoy, reames
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29331
llvm-svn: 293887
Although this is 'no-functional-change-intended', I'm adding tests
for shl-shl and lshr-lshr pairs because there is no existing test
coverage for those folds.
It seems like we should be able to remove some code from foldShiftedShift()
at this point because we're handling those patterns on the general path.
llvm-svn: 293814
This tries to address what Hal defined (in the post-commit review of
r293727) a long-standing problem with noinline, where we end up
de facto inlining trivial functions e.g.
__attribute__((noinline)) int patatino(void) { return 5; }
because of return value propagation.
llvm-svn: 293799
Summary:
If there are two adjacent guards with different conditions, we can
remove one of them and include its condition into the condition of
another one. This patch allows InstCombine to merge them by the
following pattern:
guard(a); guard(b) -> guard(a & b).
Reviewers: reames, apilipenko, igor-laevsky, anna, sanjoy
Reviewed By: sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29378
llvm-svn: 293778
A program may contain llvm.assume info that disagrees with other analysis.
This may be caused by UB in the program, so we must not crash because of that.
As noted in the code comments:
https://llvm.org/bugs/show_bug.cgi?id=31809
...we can do better, but this at least avoids the assert/crash in the bug report.
Differential Revision: https://reviews.llvm.org/D29395
llvm-svn: 293773
Fix a bug where we would construct shufflevector instructions addressing
invalid elements.
Differential Revision: https://reviews.llvm.org/D29313
llvm-svn: 293673
Summary: In iterative sample pgo where profile is collected from PGOed binary, we may see indirect call targets promoted and inlined in the profile. Before profile annotation, we need to make this happen in order to annotate correctly on IR. This patch explicitly promotes these indirect calls and inlines them before profile annotation.
Reviewers: xur, davidxl
Reviewed By: davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29040
llvm-svn: 293657
transformToIndexedCompare
If they don't have the same type, the size of the constant
index would need to be adjusted (and this wouldn't be always
possible).
Alternatively we could try the analysis with the initial
RHS value, which would guarantee that the two sides have
the same type. However it is unlikely that in practice this
would pass our transformation requirements.
Fixes PR31808 (https://llvm.org/bugs/show_bug.cgi?id=31808).
llvm-svn: 293629
For now just port some of the existing NVPTX tests
and from an old HSAIL optimization pass which
approximately did the same thing.
Don't enable the pass yet until more testing is done.
llvm-svn: 293580
For targets with different addressing modes in each address space,
if this is dropped querying isLegalAddressingMode later with this
will give a nonsense result, breaking the isLegalUse assertions.
This is a candidate for the 4.0 release branch.
llvm-svn: 293542
This reverts commit r293196
Besides making things look nicer, ATM, we'd like to preserve analysis
more than we'd like to destroy the CFG. We'll probably revisit in the future
llvm-svn: 293501
The original shift is bigger, so this may qualify as 'obvious',
but here's an attempt at an Alive-based proof:
Name: exact
Pre: (C1 u< C2)
%a = shl i8 %x, C1
%b = lshr exact i8 %a, C2
=>
%c = lshr exact i8 %x, C2 - C1
%b = and i8 %c, ((1 << width(C1)) - 1) u>> C2
Optimization is correct!
llvm-svn: 293498
The jumbled scalar loads will be sorted while building the tree and these accesses will be marked to generate shufflevector after the vectorized load with proper mask.
Reviewers: hfinkel, mssimpso, mkuper
Differential Revision: https://reviews.llvm.org/D26905
Change-Id: I9c0c8e6f91a00076a7ee1465440a3f6ae092f7ad
llvm-svn: 293386
Summary: Along with https://reviews.llvm.org/D27804, debug locations need to be merged when hoisting store instructions as well. Not sure if just dropping debug locations would make more sense for this case, but as the branch instruction will have at least different discriminator with the hoisted store instruction, I think there will be no difference in practice.
Reviewers: aprantl, andreadb, danielcdh
Reviewed By: aprantl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29062
llvm-svn: 293372
This is a minimal patch to avoid the infinite loop in:
https://llvm.org/bugs/show_bug.cgi?id=31751
But the general problem is bigger: we're not canonicalizing all of the min/max forms reported
by value tracking's matchSelectPattern(), and we don't define min/max consistently. Some code
uses matchSelectPattern(), other code uses matchers like m_Umax, and others have their own
inline definitions which may be subtly different from any of the above.
The reason that the test cases in this patch need a cast op to trigger is because we don't
(yet) canonicalize all min/max forms based on matchSelectPattern() in
canonicalizeMinMaxWithConstant(), but we do make min/max+cast transforms based on
matchSelectPattern() in visitSelectInst().
The location of the icmp transforms that trigger the inf-loop seems arbitrary at best, so
I'm moving those behind the min/max fence in visitICmpInst() as the quick fix.
llvm-svn: 293345
The interleaved access pass is an IR-to-IR transformation that runs before code
generation. It matches interleaved memory operations to target-specific
intrinsics (that are later lowered to load and store multiple instructions on
ARM/AArch64). We place tests for similar passes (e.g., GlobalMergePass) under
test/Transforms. This patch moves the InterleavedAccessPass tests out of
test/CodeGen and into target-specific directories under
test/Transforms/InterleavedAccess.
Although the pass is an IR pass, many of the existing tests were llc tests
rather opt tests. For example, the tests would check for ldN/stN instructions
generated by llc rather than the intrinsic calls the pass actually inserts.
Thus, this patch updates all tests to be opt tests that check for the inserted
intrinsics. We already have separate CodeGen tests that ensure we lower the
interleaved access intrinsics to their corresponding ldN/stN instructions. In
addition to migrating the tests to opt, this patch also performs some minor
clean-up (to ensure consistent naming, etc.).
Differential Revision: https://reviews.llvm.org/D29184
llvm-svn: 293309
skip sub-subloops.
The logic to skip subloops dated from when this code was shared with the
cached case. Once it was factored out to only run in the case of
recomputed subloops it became a dangerous bug. If a subsubloop contained
an interfering instruction it would be silently skipped from the alias
sets for LICM.
With the old pass manager this was extremely hard to trigger as it would
require failing to visit these subloops with the LICM pass but then
visiting the outer loop somehow. I've not yet contrived any test case
that actually manages to trigger this.
But with the new pass manager we don't do the cross-loop caching hack
that the old PM does and so we recompute alias set information from
first principles. While this seems much cleaner and simpler it exposed
this bug and would subtly miscompile code due to failing to correctly
model the aliasing constraints of deeply nested loops.
llvm-svn: 293273
Summary:
This adds basic dead and redundant store elimination to
NewGVN. Unlike our current DSE, it will happily do cross-block DSE if
it meets our requirements.
We get a bunch of DSE's simple.ll cases, and some stuff it doesn't.
Unlike DSE, however, we only try to eliminate stores of the same value
to the same memory location, not just general stores to the same
memory location.
Reviewers: davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29149
llvm-svn: 293258
the main pipeline.
This is a very straight forward port. Nothing weird or surprising.
This brings the number of missing passes from the new PM's pipeline down
to three.
llvm-svn: 293249
Summary:
There are many NVVM intrinsics that we can't entirely get rid of, but
that nonetheless often correspond to target-generic LLVM intrinsics.
For example, if flush denormals to zero (ftz) is enabled, we can convert
@llvm.nvvm.ceil.ftz.f to @llvm.ceil.f32. On the other hand, if ftz is
disabled, we can't do this, because @llvm.ceil.f32 will be lowered to a
non-ftz PTX instruction. In this case, we can, however, simplify the
non-ftz nvvm ceil intrinsic, @llvm.nvvm.ceil.f, to @llvm.ceil.f32.
These transformations are particularly useful because they let us
constant fold instructions that appear in libdevice, the bitcode library
that ships with CUDA and essentially functions as its libm.
Reviewers: tra
Subscribers: hfinkel, majnemer, llvm-commits
Differential Revision: https://reviews.llvm.org/D28794
llvm-svn: 293244
This change reverts:
r293061: "[InstCombine] Canonicalize guards for NOT OR condition"
r293058: "[InstCombine] Canonicalize guards for AND condition"
They miscompile cases like:
```
declare void @llvm.experimental.guard(i1, ...)
define void @test_guard_not_or(i1 %A, i1 %B) {
%C = or i1 %A, %B
%D = xor i1 %C, true
call void(i1, ...) @llvm.experimental.guard(i1 %D, i32 20, i32 30)[ "deopt"() ]
ret void
}
```
because they do transfer the `i32 20, i32 30` parameters to newly
created guard instructions.
llvm-svn: 293227
Summary:
This does not actually fix the testcase in PR31761 (discussion is
ongoing on the testcase), but does fix a bug it exposes, where stores
were not properly clobbering loads.
We accomplish this by unifying the memory equivalence infratructure
back into the normal congruence infrastructure, and then properly
destroying congruence classes when memory state leaders disappear.
Reviewers: davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29195
llvm-svn: 293216
We already have this fold when the lshr has one use, but it doesn't need that
restriction. We may be able to remove some code from foldShiftedShift().
Also, move the similar:
(X << C) >>u C --> X & (-1 >>u C)
...directly into visitLShr to help clean up foldShiftByConstOfShiftByConst().
That whole function seems questionable since it is called by commonShiftTransforms(),
but there's really not much in common if we're checking the shift opcodes for every
fold.
llvm-svn: 293215
change the set of uniform instructions in the loop causing an assert
failure.
The problem is that the legalization checking also builds data
structures mapping various facts about the loop body. The immediate
cause was the set of uniform instructions. If these then change when
LCSSA is formed, the data structures would already have been built and
become stale. The included test case triggered an assert in loop
vectorize that was reduced out of the new PM's pipeline.
The solution is to form LCSSA early enough that no information is cached
across the changes made. The only really obvious position is outside of
the main logic to vectorize the loop. This also has the advantage of
removing one case where forming LCSSA could mutate the loop but we
wouldn't track that as a "Changed" state.
If it is significantly advantageous to do some legalization checking
prior to this, we can do a more careful positioning but it seemed best
to just back off to a safe position first.
llvm-svn: 293168
factory functions for the two modes the loop unroller is actually used
in in-tree: simplified full-unrolling and the entire thing including
partial unrolling.
I've also wired these up to nice names so you can express both of these
being in a pipeline easily. This is a precursor to actually enabling
these parts of the O2 pipeline.
Differential Revision: https://reviews.llvm.org/D28897
llvm-svn: 293136
Even when we don't create a remainder loop (that is, when we unroll by 2), we
may duplicate nested loops into the remainder. This is complicated by the fact
the remainder may itself be either inserted into an outer loop, or at the top
level. In the latter case, we may need to create new top-level loops.
Differential Revision: https://reviews.llvm.org/D29156
llvm-svn: 293124
Summary:
Previously we assumed that the result of sqrt(x) always had 0 as its
sign bit. But sqrt(-0) == -0.
Reviewers: hfinkel, efriedma, sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28928
llvm-svn: 293115
This patch introduces guard based loop predication optimization. The new LoopPredication pass tries to convert loop variant range checks to loop invariant by widening checks across loop iterations. For example, it will convert
for (i = 0; i < n; i++) {
guard(i < len);
...
}
to
for (i = 0; i < n; i++) {
guard(n - 1 < len);
...
}
After this transformation the condition of the guard is loop invariant, so loop-unswitch can later unswitch the loop by this condition which basically predicates the loop by the widened condition:
if (n - 1 < len)
for (i = 0; i < n; i++) {
...
}
else
deoptimize
This patch relies on an NFC change to make ScalarEvolution::isMonotonicPredicate public (revision 293062).
Reviewed By: sanjoy
Differential Revision: https://reviews.llvm.org/D29034
llvm-svn: 293064
This is a partial fix for Bug 31520 - [guards] canonicalize guards in instcombine
Reviewed By: apilipenko
Differential Revision: https://reviews.llvm.org/D29075
Patch by Maxim Kazantsev.
llvm-svn: 293061
This is a partial fix for Bug 31520 - [guards] canonicalize guards in instcombine
Reviewed By: apilipenko
Differential Revision: https://reviews.llvm.org/D29074
Patch by Maxim Kazantsev.
llvm-svn: 293058
This is a partial fix for Bug 31520 - [guards] canonicalize guards in instcombine
Reviewed By: majnemer, apilipenko
Differential Revision: https://reviews.llvm.org/D29071
Patch by Maxim Kazantsev.
llvm-svn: 293056
instructions.
If number of instructions in horizontal reduction list is not power of 2
then only PowerOf2Floor(NumberOfInstructions) last elements are actually
vectorized, other instructions remain scalar. Patch tries to vectorize
the remaining elements either.
Differential Revision: https://reviews.llvm.org/D28959
llvm-svn: 293042
Floating point intrinsics in LLVM are generally not speculatively
executed, since most of them are defined to behave the same as libm
functions, which set errno.
However, the @llvm.powi.* intrinsics do not correspond to any libm
function, and lacks any defined error handling semantics in LangRef.
It most certainly does not alter errno.
llvm-svn: 293041
Conservatively disable sinking and merging inline-asm instructions as doing so
can potentially create arguments that cannot satisfy the inline-asm constraints.
For example, SimplifyCFG used to do the following transformation:
(before)
if.then:
%0 = call i32 asm "rorl $2, $0", "=&r,0,n"(i32 %r6, i32 8)
br label %if.end
if.else:
%1 = call i32 asm "rorl $2, $0", "=&r,0,n"(i32 %r6, i32 6)
br label %if.end
(after)
%.sink = select i1 %tobool, i32 6, i32 8
%0 = call i32 asm "rorl $2, $0", "=&r,0,n"(i32 %r6, i32 %.sink)
This would result in a crash in the backend since only immediate integer operands
are permitted for constraint "n".
rdar://problem/30110806
Differential Revision: https://reviews.llvm.org/D29111
llvm-svn: 293025
loops.
We do this by reconstructing the newly added loops after the unroll
completes to avoid threading pass manager details through all the mess
of the unrolling infrastructure.
I've enabled some extra assertions in the LPM to try and catch issues
here and enabled a bunch of unroller tests to try and make sure this is
sane.
Currently, I'm manually running loop-simplify when needed. That should
go away once it is folded into the LPM infrastructure.
Differential Revision: https://reviews.llvm.org/D28848
llvm-svn: 293011
Summary:
When we decide that the result of the invoke instruction need to be spilled, we need to insert the spill into a block that is on the normal edge coming out of the invoke instruction. (Prior to this change the code would insert the spill immediately after the invoke instruction, which breaks the IR, since invoke is a terminator instruction).
In the following example, we will split the edge going into %cont and insert the spill there.
```
%r = invoke double @print(double 0.0) to label %cont unwind label %pad
cont:
%0 = call i8 @llvm.coro.suspend(token none, i1 false)
switch i8 %0, label %suspend [i8 0, label %resume
i8 1, label %cleanup]
resume:
call double @print(double %r)
```
Reviewers: majnemer
Reviewed By: majnemer
Subscribers: mehdi_amini, llvm-commits, EricWF
Differential Revision: https://reviews.llvm.org/D29102
llvm-svn: 293006
Summary: In iterative sample pgo where profile is collected from PGOed binary, we may see indirect call targets promoted and inlined in the profile. Before profile annotation, we need to make this happen in order to annotate correctly on IR. This patch explicitly promotes these indirect calls and inlines them before profile annotation.
Reviewers: xur, davidxl
Reviewed By: davidxl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29040
llvm-svn: 292979
Summary:
GVNHoist performs all the optimizations that MLSM does to loads, in a
more general way, and in a faster time bound (MLSM is N^3 in most
cases, N^4 in a few edge cases).
This disables the load portion.
Note that the way ld_hoist_st_sink.ll is written makes one think that
the loads should be moved to the while.preheader block, but
1. Neither MLSM nor GVNHoist do it (they both move them to identical places).
2. MLSM couldn't possibly do it anyway, as the while.preheader block
is not the head of the diamond, while.body is. (GVNHoist could do it
if it was legal).
3. At a glance, it's not legal anyway because the in-loop load
conflict with the in-loop store, so the loads must stay in-loop.
I am happy to update the test to use update_test_checks so that
checking is tighter, just was going to do it as a followup.
Note that i can find no particular benefit to the store portion on any
real testcase/benchmark i have (even size-wise). If we really still
want it, i am happy to commit to writing a targeted store sinker, just
taking the code from the MemorySSA port of MergedLoadStoreMotion
(which is N^2 worst case, and N most of the time).
We can do what it does in a much better time bound.
We also should be both hoisting and sinking stores, not just sinking
them, anyway, since whether we should hoist or sink to merge depends
basically on luck of the draw of where the blockers are placed.
Nonetheless, i have left it alone for now.
Reviewers: chandlerc, davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29079
llvm-svn: 292971
a lazy-asserting PoisoningVH.
AssertVH is fundamentally incompatible with cache-invalidation of
analysis results. The invaliadtion happens after the AssertingVH has
already fired. Instead, use a PoisoningVH that will assert if the
dangling handle is ever used rather than merely be assigned or
destroyed.
This patch also removes all of the (numerous) doomed attempts to work
around this fundamental incompatibility. It is a pretty significant
simplification IMO.
The most interesting change is in the Inliner where we still do some
clearing because we don't want to rely on the coarse grained
invalidation strategy of the containing pass manager. However, I prefer
the approach that contains this logic to the cleanup phase of the
Inliner, and I think we could enhance the CGSCC analysis management
layer to make this even better in the future if desired.
The rest is straight cleanup.
I've also added a test for one of the harder cases to work around: when
a *module analysis* contains many AssertingVHes pointing at functions.
Differential Revision: https://reviews.llvm.org/D29006
llvm-svn: 292928
With this change dominator tree remains in sync after each step of loop
peeling.
Differential Revision: https://reviews.llvm.org/D29029
llvm-svn: 292895
Running non-LCSSA-preserving LoopSimplify followed by LCSSA on (roughly) the
same loop is incorrect, since LoopSimplify may break LCSSA arbitrarily higher
in the loop nest. Instead, run LCSSA first, and then run LCSSA-preserving
LoopSimplify on the result.
This fixes PR31718.
Differential Revision: https://reviews.llvm.org/D29055
llvm-svn: 292854
Summary:
Next round of extra tests for MSSA.
I have a prototype invariant.group handling implementation
that fixes all the FIXMEs, and I think it will be
easier to see what is the difference if I firstly
post this, and then only fix fixits.
Reviewers: george.burgess.iv, dberlin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29022
llvm-svn: 292797
bots ever since d0k fixed the CHECK lines so that it did something at
all.
It isn't actually testing SCEV directly but LSR, so move it into LSR and
the x86-specific tree of tests that already exists there. Target
dependence is common and unavoidable with the current design of LSR.
llvm-svn: 292774
invalidation of deleted functions in GlobalDCE.
This was always testing a bug really triggered in GlobalDCE. Right now
we have analyses with asserting value handles into IR. As long as those
remain, when *deleting* an IR unit, we cannot wait for the normal
invalidation scheme to kick in even though it was designed to work
correctly in the face of these kinds of deletions. Instead, the pass
needs to directly handle invalidating the analysis results pointing at
that IR unit.
I've tought the Inliner about this and this patch teaches GlobalDCE.
This will handle the asserting VH case in the existing test as well as
other issues of the same fundamental variety. I've moved the test into
the GlobalDCE directory and added a comment explaining what is going on.
Note that we cannot simply require LVI here because LVI is too lazy.
llvm-svn: 292773
While this is covered by a clang test case, we should have something
locally to LLVM that immediately checks the inliner doesn't leave
analyses to dangling IR bodies.
llvm-svn: 292772
new PM's inliner.
The bug happens when we refine an SCC after having computed a proxy for
the FunctionAnalysisManager, and then proceed to compute fresh analyses
for functions in the *new* SCC using the manager provided by the old
SCC's proxy. *And* when we manage to mutate a function in this new SCC
in a way that invalidates those analyses. This can be... challenging to
reproduce.
I've managed to contrive a set of functions that trigger this and added
a test case, but it is a bit brittle. I've directly checked that the
passes run in the expected ways to help avoid the test just becoming
silently irrelevant.
This gets the new PM back to passing the LLVM test suite after the PGO
improvements landed.
llvm-svn: 292757
Summary:
This test had a bug: !llvm.invariant.group instead
of !invariant.group.
Also add some new test for future development.
All tests passes, when MSSA will support invariant.group
only the lines with FIXIT should be changed.
Reviewers: dberlin, george.burgess.iv
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28969
llvm-svn: 292730
We may be able to assert that no shl-shl or lshr-lshr pairs ever get here
because we should have already handled those in foldShiftedShift().
llvm-svn: 292726
This adds the last remaining core feature of the loop pass pipeline in
the new PM and removes the last of the really egregious hacks in the
LICM tests.
Sadly, this requires really substantial changes in the unittests in
order to provide and maintain simplified loops. This is particularly
hard because for example LoopSimplify will try to fold undef branches to
an ideal direction and simplify the loop accordingly.
Differential Revision: https://reviews.llvm.org/D28766
llvm-svn: 292709
Summary:
Under option -mergefunc-preserve-debug-info we:
- Do not create a new function for a thunk.
- Retain the debug info for a thunk's parameters (and associated
instructions for the debug info) from the entry block.
Note: -debug will display the algorithm at work.
- Create debug-info for the call (to the shared implementation) made by
a thunk and its return value.
- Erase the rest of the function, retaining the (minimally sized) entry
block to create a thunk.
- Preserve a thunk's call site to point to the thunk even when both occur
within the same translation unit, to aid debugability. Note that this
behaviour differs from the underlying -mergefunc implementation which
modifies the thunk's call site to point to the shared implementation
when both occur within the same translation unit.
Reviewers: echristo, eeckstein, dblaikie, aprantl, friss
Reviewed By: aprantl
Subscribers: davide, fhahn, jfb, mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D28075
llvm-svn: 292702
Summary:
Currently we return undef, but we're in the process of changing the
LangRef so that llvm.sqrt behaves like the other math intrinsics,
matching the return value of the standard libcall but not setting errno.
This change is legal even without the LangRef change because currently
calling llvm.sqrt(x) where x is negative is spec'ed to be UB. But in
practice it's also safe because we're simply constant-folding fewer
inputs: Inputs >= -0 get constant-folded as before, but inputs < -0 now
aren't constant-folded, because ConstantFoldFP aborts if the host math
function raises an fp exception.
Reviewers: hfinkel, efriedma, sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28929
llvm-svn: 292692
This adds the following to the new PM based inliner in PGO mode:
* Use block frequency analysis to derive callsite's profile count and use
that to adjust thresholds of hot and cold callsites.
* Incrementally update the BFI of the caller after a callee gets inlined
into it. This incremental update is only within an invocation of the run
method - BFI is not preserved across calls to run.
Update the function entry count of the callee after inlining it into a
caller.
* I've tuned the thresholds for the hot and cold callsites using a hacked
up version of the old inliner that explicitly computes BFI on a set of
internal benchmarks and spec. Once the new PM based pipeline stabilizes
(IIRC Chandler mentioned there are known issues) I'll benchmark this
again and adjust the thresholds if required.
Inliner PGO support.
Differential revision: https://reviews.llvm.org/D28331
llvm-svn: 292666
Unfortunately, recognizing these in value tracking may cause us to hit
a hack in InstCombiner::visitICmpInst() more often:
http://lists.llvm.org/pipermail/llvm-dev/2017-January/109340.html
...but besides being the obviously Right Thing To Do, there's a clear
codegen win from identifying these patterns for several targets.
llvm-svn: 292655
To import a type identifier we read the summary and create external
references to the symbols defined when exporting.
Differential Revision: https://reviews.llvm.org/D28546
llvm-svn: 292654
Summary:
This rewrites store expression/leader handling. We no longer use the
value operand as the leader, instead, we store it separately. We also
now store the stored value as part of the expression, and compare it
when comparing stores for equality. This enables us to get rid of a
bunch of our previous hacks and machinations, as the existing
machinery takes care of everything *except* updating the stored value
on classes. The only time we have to update it is if the storecount
goes to 0, and when we do, we destroy it.
Since we no longer use the value operand as the leader, during elimination, we have to use the value operand. Doing this also fixes a bunch of store forwarding cases we were missing.
Any value operand we use is guaranteed to either be updated by previous eliminations, or minimized by future ones.
(IE the fact that we don't use the most dominating value operand when it's not a constant does not affect anything).
Sadly, this change also exposes that we didn't pay attention to the
output of the pr31594.ll test, as it also very clearly exposes the
same store leader bug we are fixing here.
(I added pr31682.ll anyway, but maybe we think that's too large to be useful)
On the plus side, propagate-ir-flags.ll now passes due to the
corrected store forwarding.
This change was 3 stage'd on darwin and linux, with the full test-suite.
Reviewers:
davide
Subscribers:
llvm-commits
llvm-svn: 292648
This is the third attemp to recommit r292526.
The original summary:
Currently, a GEP is considered free only if its indices are all constant.
TTI::getGEPCost() can give target-specific more accurate analysis. TTI is
already used for the cost of many other instructions.
llvm-svn: 292633
This is the second attemp to recommit r292526.
The original summary:
Currently, a GEP is considered free only if its indices are all constant.
TTI::getGEPCost() can give target-specific more accurate analysis. TTI is
already used for the cost of many other instructions.
llvm-svn: 292616
Simplify a packss/packus truncation based on the elements of the mask that are actually demanded.
Differential Revision: https://reviews.llvm.org/D28777
llvm-svn: 292591
Like several other loop passes (the vectorizer, etc) this pass doesn't
really fit the model of a loop pass. The critical distinction is that it
isn't intended to be pipelined together with other loop passes. I plan
to add some documentation to the loop pass manager to make this more
clear on that side.
LoopSink is also different because it doesn't really need a lot of the
infrastructure of our loop passes. For example, if there aren't loop
invariant instructions causing a preheader to exist, there is no need to
form a preheader. It also doesn't need LCSSA because this pass is
only involved in sinking invariant instructions from a preheader into
the loop, not reasoning about live-outs.
This allows some nice simplifications to the pass in the new PM where we
can directly walk the loops once without restructuring them.
Differential Revision: https://reviews.llvm.org/D28921
llvm-svn: 292589
Part of the assert has been left active for further debugging.
The other part has been turned into a stat for tracking for the
moment.
llvm-svn: 292583
This recommits r292526 which is reverted in r292529 after fixing the test case.
The original summary:
Currently, a GEP is considered free only if its indices are all constant.
TTI::getGEPCost() can give target-specific more accurate analysis. TTI is
already used for the cost of many other instructions.
llvm-svn: 292570
Summary:
Fence instructions are currently marked as `ModRef` for all memory locations.
We can improve this for constant memory locations (such as constant globals),
since fence instructions cannot modify these locations.
This helps us to forward constant loads across fences (added test case in GVN).
There were no changes in behaviour for similar test cases in early-cse and licm.
Reviewers: dberlin, sanjoy, reames
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28914
llvm-svn: 292546
This can prove that:
extern int f;
int g() {
int x = 0;
for (int i = 0; i < 365; ++i) {
x /= f;
}
return x;
}
always returns zero. Thanks to Sanjoy for confirming this
transformation actually made sense (bugs are mine).
llvm-svn: 292531
Currently, a GEP is considered free only if its indices are all constant.
TTI::getGEPCost() can give target-specific more accurate analysis. TTI is
already used for the cost of many other instructions.
Differential Revision: https://reviews.llvm.org/D28693
llvm-svn: 292526
Summary:
In case of non-alloca pointers, we check for whether it is a pointer
from malloc-like calls and it is not captured. In such case, we can
promote the pointer, as the caller will have no way to access this pointer
even if there is unwinding in middle of the loop.
Reviewers: hfinkel, sanjoy, reames, eli.friedman
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28834
llvm-svn: 292510
Type identifiers are exported by:
- Adding coarse-grained information about how to test the type
identifier to the summary.
- Creating symbols in the object file (aliases and absolute symbols)
containing fine-grained information about the type identifier.
Differential Revision: https://reviews.llvm.org/D28424
llvm-svn: 292462
This changes the vectorizer to explicitly use the loopsimplify and lcssa utils,
instead of "requiring" the transformations as if they were analyses.
This is not NFC, since it changes the LCSSA behavior - we no longer run LCSSA
for all loops, but rather only for the loops we expect to modify.
Differential Revision: https://reviews.llvm.org/D28868
llvm-svn: 292456
We currently check whether a reduction has a single outside user. We don't
really need to require that - we just need to make sure a single value is
used externally. The number of external users of that value shouldn't actually
matter.
Differential Revision: https://reviews.llvm.org/D28830
llvm-svn: 292424
claims to test.
LoopSimplify was unifying the multiple exits in this test case, making
it never even test the multiple exit handling of LoopDeletion. Doh.
Now it works (thanks to a great idea from mkuper) and will fail if we
ever change something to make it stop working.
llvm-svn: 292331
Summary: Partial unrolling should have separate threshold with full unrolling.
Reviewers: efriedma, mzolotukhin
Reviewed By: efriedma, mzolotukhin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28831
llvm-svn: 292293
Summary: Add a test case for LICM when promoting locals that may be read after the throw within the loop.
Reviewers: eli.friedman, hfinkel, sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28822
llvm-svn: 292261
If a memory instruction will be vectorized, but it's pointer operand is
non-consecutive-like, the instruction is a gather or scatter operation. Its
pointer operand will be non-uniform. This should fix PR31671.
Reference: https://llvm.org/bugs/show_bug.cgi?id=31671
Differential Revision: https://reviews.llvm.org/D28819
llvm-svn: 292254
runnig LCSSA over them prior to running the loop pipeline.
This also teaches the loop PM to verify that LCSSA form is preserved
throughout the pipeline's run across the loop nest.
Most of the test updates just leverage this new functionality. One has to be
relaxed with the new PM as IVUsers is less powerful when it sees LCSSA input.
Differential Revision: https://reviews.llvm.org/D28743
llvm-svn: 292241
Also, add the corresponding match to the AssumptionCache's 'Affected Values' list.
Differential Revision: https://reviews.llvm.org/D28485
llvm-svn: 292239
Add missing fabs(fpext) optimzation that worked with the call,
and also fixes it creating a second fpext when there were multiple
uses.
llvm-svn: 292172
Simplify a pshufb shuffle mask based on the elements of the mask that are actually demanded.
Differential Revision: https://reviews.llvm.org/D28745
llvm-svn: 292101
First, I've moved a test of IVUsers from the LSR tree to a dedicated
IVUsers test directory. I've also simplified its RUN line now that the
new pass manager's loop PM is providing analyses on their own.
No functionality changed, but it makes subsequent changes cleaner.
llvm-svn: 292060
cover domtree and alias analysis. These are the pretty clear analyses
that we would always want to survive this pass.
To make these survive, we also need to preserve the assumption cache.
Added a test that verifies the important bits of this preservation.
llvm-svn: 292037
Allows LLVM to optimize sequences like the following:
%add = add nuw i32 %x, 1
%cmp = icmp ugt i32 %add, %y
Into:
%cmp = icmp uge i32 %x, %y
Previously, only signed comparisons were being handled.
Decrements could also be handled, but 'sub nuw %x, 1' is currently canonicalized to
'add %x, -1' in InstCombineAddSub, losing the nuw flag. Removing that canonicalization
seems like it might have far-reaching ramifications so I kept this simple for now.
Patch by Matti Niemenmaa!
Differential Revision: https://reviews.llvm.org/D24700
llvm-svn: 291975
Summary:
This is a testcase where phi node cycling happens, and because we do
not order the leaders by domination or anything similar, the leader
keeps changing.
Using std::set for the members is too expensive, and we actually don't
need them sorted all the time, only at leader changes.
We could keep both a set and a vector, and keep them mostly sorted and
resort as necessary, or use a set and a fibheap, but all of this seems
premature.
After running some statistics, we are able to avoid the vast majority
of sorting by keeping a "next leader" field. Most congruence classes only have
leader changes once or twice during GVN.
Reviewers: davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28594
llvm-svn: 291968
Summary:
Memory Dependence Analysis was limited to return only local dependencies
for invariant.group handling. Now it returns NonLocal when it finds it
and then by asking getNonLocalPointerDependency we get found dep.
Thanks to this we are able to devirtualize loops!
void indirect(A &a, int n) {
for (int i = 0 ; i < n; i++)
a.foo();
}
void test(int n) {
A a;
indirect(a);
}
After inlining a.foo() will be changed to direct call, even if foo and A::A()
is external (but only if vtable definition is be available).
Reviewers: nlewycky, dberlin, chandlerc, rsmith
Subscribers: mehdi_amini, davide, llvm-commits
Differential Revision: https://reviews.llvm.org/D28137
llvm-svn: 291762
This test seems to have largely been relying on asserts being tripped.
It had a very specific and somewhat uninteresting grep of the output,
but it never really did anything to cause SCEV to be preserved across
loop simplify, certainly not explicitly. And a later addition to it
actually added CHECK lines despite the test never running FileCheck.
Now we actually print SCEV before and after loop simplify to make sure
it is *changing* and being *updated*. Which seems to be much more likely
the point of the test.
llvm-svn: 291740
This means that we can use a shorter instruction sequence in the case where
the size is a power of two and on the boundary between two representations.
Differential Revision: https://reviews.llvm.org/D28421
llvm-svn: 291706
classes, and updating checking to allow for equivalence through
reachability.
(Sadly, the checking here is not perfect, and can't be made perfect,
so we'll have to disable it after we are satisfied with correctness.
Right now it is just "very unlikely" to happen.)
llvm-svn: 291698
The removed assert seems bogus - it's perfectly legal for the roots of the
vectorized subtrees to be equal even if the original scalar values aren't,
if the original scalars happen to be equivalent.
This fixes PR31599.
Differential Revision: https://reviews.llvm.org/D28539
llvm-svn: 291692
Summary:
Revert LowerTypeTests: Split the pass in two: a resolution phase and a lowering phase.
This change separates how type identifiers are resolved from how intrinsic
calls are lowered. All information required to lower an intrinsic call
is stored in a new TypeIdLowering data structure. The idea is that this
data structure can either be initialized using the module itself during
regular LTO, or using the module summary in ThinLTO backends.
Original URL: https://reviews.llvm.org/D28341
Reviewers: pcc
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D28532
llvm-svn: 291684
updated instructions:
pmulld, pmullw, pmulhw, mulsd, mulps, mulpd, divss, divps, divsd, divpd, addpd and subpd.
special optimization case which replaces pmulld with pmullw\pmulhw\pshuf seq.
In case if the real operands bitwidth <= 16.
Differential Revision: https://reviews.llvm.org/D28104
llvm-svn: 291657
These are interesting again because the user may not be aware that this
is a common reason preventing LICM.
A const is removed from an instruction pointer declaration in order to
pass it to ORE.
Differential Revision: https://reviews.llvm.org/D27940
llvm-svn: 291649
This patch reverts r291588: [PGO] Turn off comdat renaming in IR PGO by default,
as we are seeing some hash mismatches in our internal tests.
llvm-svn: 291621
Bail out instead of asserting when we encounter this situation,
which can actually happen.
The reason the test uses the new PM is that the "bad" phi, incidentally, gets
cleaned up by LoopSimplify. But LICM can create this kind of phi and preserve
loop simplify form, so the cleanup has no chance to run.
This fixes PR31190.
We may want to solve this in a less conservative manner, since this phi is
actually uniform within the inner loop (or we may want LICM to output a cleaner
promotion to begin with).
Differential Revision: https://reviews.llvm.org/D28490
llvm-svn: 291589
Summary:
In IR PGO we append the function hash to comdat functions to avoid the
potential hash mismatch. This turns out not legal in some cases: if the comdat
function is address-taken and used in comparison. Renaming changes the semantic.
This patch turns off comdat renaming by default.
To alleviate the hash mismatch issue, we now rename the profile variable
for comdat functions. Profile allows co-existing multiple versions of profiles
with different hash value. The inlined copy will always has the correct profile
counter. The out-of-line copy might not have the correct count. But we will
not have the bogus mismatch warning.
Reviewers: davidxl
Subscribers: llvm-commits, xur
Differential Revision: https://reviews.llvm.org/D28416
llvm-svn: 291588
In some cases StructurizeCfg updates root node, but dominator info
remains unchanges, it causes crash when expensive checks are enabled.
To cope with this problem a new method was added to DominatorTreeBase
that allows adding new root nodes, it is called in StructurizeCfg to
put dominator tree in sync.
This change fixes PR27488.
Differential Revision: https://reviews.llvm.org/D28114
llvm-svn: 291530
This patch delays the fix-up step for external induction variable users until
after the dominator tree has been properly updated. This should fix PR30742.
The SCEVExpander in InductionDescriptor::transform can generate code in the
wrong location if the dominator tree is not up-to-date. We should work towards
keeping the dominator tree up-to-date throughout the transformation.
Reference: https://llvm.org/bugs/show_bug.cgi?id=30742
Differential Revision: https://reviews.llvm.org/D28168
llvm-svn: 291462
Summary:
By using stripPointerCasts we can get to the root
value and then walk down the bitcast graph
Reviewers: reames
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28181
llvm-svn: 291405
fabs(x * x) is not generally safe to assume x is positive if x is a NaN.
This is also less general than it could be, so this will be replaced
with a transformation on the intrinsic.
llvm-svn: 291359
Summary: LLVM's non-standard notion of phi nodes means we can't both try to substitute for undef in phi nodes *and* use phi nodes as leaders all the time. This changes NewGVN to use the same semantics as SimplifyPHINode to decide which phi nodes are equivalent.
Reviewers: davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28312
llvm-svn: 291308
This is fixing a bug where Loop Vectorization is widening a load but
with a lower alignment. Hoisting the load without propagating the alignment
will allow inst-combine to later deduce a higher alignment that what the pointer
actually is.
Differential Revision: https://reviews.llvm.org/D28408
llvm-svn: 291281
This change separates how type identifiers are resolved from how intrinsic
calls are lowered. All information required to lower an intrinsic call
is stored in a new TypeIdLowering data structure. The idea is that this
data structure can either be initialized using the module itself during
regular LTO, or using the module summary in ThinLTO backends.
Differential Revision: https://reviews.llvm.org/D28341
llvm-svn: 291205
Promotion is always legal when a store within the loop is guaranteed to execute.
However, this is not a necessary condition - for promotion to be memory model
semantics-preserving, it is enough to have a store that dominates every exit
block. This is because if the store dominates every exit block, the fact the
exit block was executed implies the original store was executed as well.
Differential Revision: https://reviews.llvm.org/D28147
llvm-svn: 291171
This code seems to be target dependent which may not be the same for all targets.
Passed the decision whether the given stride is complex or not to the target by sending stride information via SCEV to getAddressComputationCost instead of 'IsComplex'.
Specifically at X86 targets we dont see any significant address computation cost in case of the strided access in general.
Differential Revision: https://reviews.llvm.org/D27518
llvm-svn: 291106
Set up basic YAML I/O support for module summaries, plumb the summary into
the pass and add a few command line flags to test YAML I/O support. Bitcode
support to come separately, as will the code in LowerTypeTests that actually
uses the summary. Also add a couple of tests that pass by virtue of the pass
doing nothing with the summary (which happens to be the correct thing to do
for those tests).
Differential Revision: https://reviews.llvm.org/D28041
llvm-svn: 291069
performing partial redundancy elimination (PRE). Not doing so can cause jumpy line
tables and confusing (though correct) source attributions.
Differential Revision: https://reviews.llvm.org/D27857
llvm-svn: 291037
We can perform the following:
(add (zext (add nuw X, C1)), C2) -> (zext (add nuw X, C1+C2))
This is only possible if C2 is negative and C2 is greater than or equal to negative C1.
llvm-svn: 290927
Summary:
Regardless how the loop body weight is distributed, we should preserve
total loop body weight. i.e. we should have same weight reaching the body of the loop
or its duplicates in peeled and unpeeled case.
Reviewers: mkuper, davidxl, anemet
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28179
llvm-svn: 290833
Summary:
gep 0, 0 is equivalent to bitcast. LLVM canonicalizes it
to getelementptr because it make SROA can then handle it.
Simple case like
void g(A &a) {
z(a);
if (glob)
a.foo();
}
void testG() {
A a;
g(a);
}
was not devirtualized with -fstrict-vtable-pointers because luck of
handling for gep 0 in Memory Dependence Analysis
Reviewers: dberlin, nlewycky, chandlerc
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D28126
llvm-svn: 290763
This is similar to the allocfn case - if an alloca is not captured, then it's
necessarily thread-local.
Differential Revision: https://reviews.llvm.org/D28170
llvm-svn: 290738
Summary:
The current loop complete unroll algorithm checks if unrolling complete will reduce the runtime by a certain percentage. If yes, it will apply a fixed boosting factor to the threshold (by discounting cost). The problem for this approach is that the threshold abruptly. This patch makes the boosting factor a function of runtime reduction percentage, capped by a fixed threshold. In this way, the threshold changes continuously.
The patch also simplified the code by reducing one parameter in UP.
The patch only affects code-gen of two speccpu2006 benchmark:
445.gobmk binary size decreases 0.08%, no performance change.
464.h264ref binary size increases 0.24%, no performance change.
Reviewers: mzolotukhin, chandlerc
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D26989
llvm-svn: 290737
The accidentally had trivially dead code. Also needed to adjust the rounding mode to not CUR_DIRECTION so the intrinsics don't get converted to native operations before going through SimplifyDemandedVectorElts.
llvm-svn: 290702
This is an orthogonal and separated layer instead of being embedded
inside the pass manager. While it adds a small amount of complexity, it
is fairly minimal and the composability and control seems worth the
cost.
The logic for this ends up being nicely isolated and targeted. It should
be easy to experiment with different iteration strategies wrapped around
the CGSCC bottom-up walk using this kind of facility.
The mechanism used to track devirtualization is the simplest one I came
up with. I think it handles most of the cases the existing iteration
machinery handles, but I haven't done a *very* in depth analysis. It
does however match the basic intended semantics, and we can tweak or
tune its exact behavior incrementally as necessary. One thing that we
may want to revisit is freshly building the value handle set on each
iteration. While I don't think this will be a significant cost (it is
strictly fewer value handles but more churn of value handes than the old
call graph), it is conceivable that we'll want a somewhat more clever
tracking mechanism. My hope is to layer that on as a follow up patch
with data supporting any implementation complexity it adds.
This code also provides for a basic count heuristic: if the number of
indirect calls decreases and the number of direct calls increases for
a given function in the SCC, we assume devirtualization is responsible.
This matches the heuristics currently used in the legacy pass manager.
Differential Revision: https://reviews.llvm.org/D23114
llvm-svn: 290665
analyses when we're about to break apart an SCC.
We can't wait until after breaking apart the SCC to invalidate things:
1) Which SCC do we then invalidate? All of them?
2) Even if we invalidate all of them, a newly created SCC may not have
a proxy that will convey the invalidation to functions!
Previously we only invalidated one of the SCCs and too late. This led to
stale analyses remaining in the cache. And because the caching strategy
actually works, they would get used and chaos would ensue.
Doing invalidation early is somewhat pessimizing though if we *know*
that the SCC structure won't change. So it turns out that the design to
make the mutation API force the caller to know the *kind* of mutation in
advance was indeed 100% correct and we didn't do enough of it. So this
change also splits two cases of switching a call edge to a ref edge into
two separate APIs so that callers can clearly test for this and take the
easy path without invalidating when appropriate. This is particularly
important in this case as we expect most inlines to be between functions
in separate SCCs and so the common case is that we don't have to so
aggressively invalidate analyses.
The LCG API change in turn needed some basic cleanups and better testing
in its unittest. No interesting functionality changed there other than
more coverage of the returned sequence of SCCs.
While this seems like an obvious improvement over the current state, I'd
like to revisit the core concept of invalidating within the CG-update
layer at all. I'm wondering if we would be better served forcing the
callers to handle the invalidation beforehand in the cases that they
can handle it. An interesting example is when we want to teach the
inliner to *update and preserve* analyses. But we can cross that bridge
when we get there.
With this patch, the new pass manager an build all of the LLVM test
suite at -O3 and everything passes. =D I haven't bootstrapped yet and
I'm sure there are still plenty of bugs, but this gives a nice baseline
so I'm going to increasingly focus on fleshing out the missing
functionality, especially the bits that are just turned off right now in
order to let us establish this baseline.
llvm-svn: 290664
when they are call edges at the leaf but may (transitively) be reached
via ref edges.
It turns out there is a simple rule: insert everything as a ref edge
which is a safe conservative default. Then we let the existing update
logic handle promoting some of those to call edges.
Note that it would be fairly cheap to make these call edges right away
if that is desirable by testing whether there is some existing call path
from the source to the target. It just seemed like slightly more
complexity in this code path that isn't strictly necessary. If anyone
feels strongly about handling this differently I'm happy to change it.
llvm-svn: 290649
This adds a combine that canonicalizes a chain of inserts which broadcasts
a value into a single insert + a splat shufflevector.
This fixes PR31286.
Differential Revision: https://reviews.llvm.org/D27992
llvm-svn: 290641
most of the inliner test cases.
The inliner involves a bunch of interesting code and tends to be where
most of the issues I've seen experimenting with the new PM lie. All of
these test cases pass, but I'd like to keep some more thorough coverage
here so doing a fairly blanket enabling.
There are a handful of interesting tests I've not enabled yet because
they're focused on the always inliner, or on functionality that doesn't
(yet) exist in the inliner.
llvm-svn: 290592
skipping indirectly recursive inline chains.
To do this, we implicitly build an inline stack for each callsite and
check prior to inlining that doing so would not form a cycle. This uses
the exact same technique and even shares some code with the legacy PM
inliner.
This solution remains deeply unsatisfying to me because it means we
cannot actually iterate the inliner externally. Doing so would not be
able to easily detect and avoid such cycles. Some day I would very much
like to have a solution that works without this internal state to detect
cycles, but this is not that day.
llvm-svn: 290590
Nothing really interesting here, but I had to improve the test to use
variables rather than hard coding value names as we happen to end up
with different value names in the new PM.
llvm-svn: 290589
We currently ignore the `allocsize` attribute on functions calls with
the `nobuiltin` attribute when trying to lower `@llvm.objectsize`. We
shouldn't care about `nobuiltin` here: `allocsize` is explicitly added
by the user, not inferred based on a function's symbol.
llvm-svn: 290588
PMULDQ/PMULUDQ vXi64 instructions only use the even numbered v2Xi32 input elements which SimplifyDemandedVectorElts should try and use.
This builds on r290554 which added supported for 128 and 256-bit.
llvm-svn: 290582
This mostly involved converting from grep to FileCheck and tidying up
the IR used.
In one case (invoke_test-3.ll) the test had become completely pointless
as we use 'resume' rather than 'unwind' now, and even then it did not
occur at the end of the line.
llvm-svn: 290570
An earlier commit added support for unmasked scalar operations. At that time isel wouldn't generate an optimal sequence for masked operations, but that has now been fixed.
llvm-svn: 290566
inside of `InlineFunction`. Prior to this, call instructions are
specifically being rewritten and replaced within the inlined region,
invalidating some of the call sites.
Several of these regions are using the same technique to walk the
inlined region so this seems clearly safe up to this point.
I've also added a short circuit to the scan for call sites based on what
other code is doing.
With this, the most common crash I've found in the new inliner code is
fixed. I've turned it on for another test case that covers this
scenario.
I'll make my way through most of the other inliner test cases
just to get some easy coverage next.
llvm-svn: 290562
removing fully-dead comdats without removing dead entries in comdats
with live members.
This factors the core logic out of the current inliner's internals to
a reusable utility and leverages that in both places. The factored out
code should also be (minorly) more efficient in cases where we have very
few dead functions or dead comdats to consider.
I've added a test case to cover this behavior of the always inliner.
This is the last significant bug in the new PM's always inliner I've
found (so far).
llvm-svn: 290557
PMULDQ/PMULUDQ vXi64 instructions only use the even numbered v2Xi32 input elements which SimplifyDemandedVectorElts should try and use.
Differential Revision: https://reviews.llvm.org/D28119
llvm-svn: 290554
The current GVN algorithm folds unconditional branches to, it claims,
expose more PRE oportunities. The folding, if really needed,
(which is not sure, as it's not really proved it improves analysis)
can be done by an earlier cleanup pass instead of GVN itself.
Ack'ed/SGTM'd by Daniel Berlin.
Differential Revision: https://reviews.llvm.org/D28117
llvm-svn: 290546
systematically and document in the test what all is going on.
This replaces the PR-named test that was the only coverage for GlobalDCE
and comdats previously. I wrote this because I wasn't certain how
comdat DCE was supposed to work and wanted to step through what
GlobalDCE did to fully understand it. After talking to folks and reading
the code and really staring at things it all makes sense but it seemed
good to help write down some of this in a more explicit and fully
covering test case.
For example, it seemed like a bug that GlobalDCE didn't consider comdat
participation of ifuncs. Specifically it seemed like an accident because
testing didn't really cover that case. But in fact, ifuncs specifically
cannot participate in a comdat despite having that API. The new test
case covers this and explicitly documents that DCE gets to fire here
even though there are comdats involved.
Also, we didn't have any positive tests for the challenging cases such
as usage cycles between comdat participants that might make them seem
alive except that there is no external edge into the cycle.
llvm-svn: 290537
Summary:
I only do this for unmasked cases for now because isel is failing to fold the mask. I'll try to fix that soon.
I'll do the same thing for packed add/sub/mul/div in a future patch.
Reviewers: delena, RKSimon, zvi, craig.topper
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D27879
llvm-svn: 290535
Summary:
This patch adds support for converting the masked vpermv intrinsics into shufflevector instructions if the indices are constants.
We also need to wrap a select instruction around the shuffle to take care of the masking part. InstCombine will take care of optimizing the select if the mask is constant so I didn't bother checking for that.
Reviewers: zvi, delena, spatel, RKSimon
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D27825
llvm-svn: 290530
whether functions are removed, and fix the new PM's always inliner to
actually pass this test.
Without this, the new PM's always inliner leaves all the functions
kicking around which won't work out very well given the semantics of
always inline.
Doing this really highlights how frustrating the current alwaysinline
semantic contract is though -- why can we put it on *external*
functions, etc?
Also I've added a number of tricky and interesting test cases for
removing functions with the always inliner. There is one remaining case
not handled -- fully removing comdats -- and I've left a FIXME about
this.
llvm-svn: 290457
The pass creates some state which expects to be cleaned up by
a later instance of the same pass. opt-bisect happens to expose
this not ideal design because calling skipLoop() will result in
this state not being cleaned up at times and an assertion firing
in `doFinalization()`. Chandler tells me the new pass manager will
give us options to avoid these design traps, but until it's not ready,
we need a workaround for the current pass infrastructure. Fix provided
by Andy Kaylor, see the review for a complete discussion.
Differential Revision: https://reviews.llvm.org/D25848
llvm-svn: 290427
Use a dummy private function with inline asm calls instead of module
level asm blocks for CFI jumptables.
The main advantage is that now jumptable codegen can be affected by
the function attributes (like target_cpu on ARM). Module level asm
gets the default subtarget based on the target triple, which is often
not good enough.
This change also uses asm constraints/arguments to reference
jumptable targets and aliases directly. We no longer do asm name
mangling in an IR pass.
Differential Revision: https://reviews.llvm.org/D28012
llvm-svn: 290384
The code have been developed by Daniel Berlin over the years, and
the new implementation goal is that of addressing shortcomings of
the current GVN infrastructure, i.e. long compile time for large
testcases, lack of phi predication, no load/store value numbering
etc...
The current code just implements the "core" GVN algorithm, although
other pieces (load coercion, phi handling, predicate system) are
already implemented in a branch out of tree. Once the core is stable,
we'll start adding pieces on top of the base framework.
The test currently living in test/Transform/NewGVN are a copy
of the ones in GVN, with proper `XFAIL` (missing features in NewGVN).
A flag will be added in a future commit to enable NewGVN, so that
interested parties can exercise this code easily.
Differential Revision: https://reviews.llvm.org/D26224
llvm-svn: 290346
This patch renumbers the metadata nodes in debug info testcases after
https://reviews.llvm.org/D26769. This is a separate patch because it
causes so much churn. This was implemented with a python script that
pipes the testcases through llvm-as - | llvm-dis - and then goes
through the original and new output side-by side to insert all
comments at a close-enough location.
Differential Revision: https://reviews.llvm.org/D27765
llvm-svn: 290292
In r267672, where the loop distribution pragma was introduced, I tried
it hard to keep the old behavior for opt: when opt is invoked
with -loop-distribute, it should distribute the loop (it's off by
default when ran via the optimization pipeline).
As MichaelZ has discovered this has the unintended consequence of
breaking a very common developer work-flow to reproduce compilations
using opt: First you print the pass pipeline of clang
with -debug-pass=Arguments and then invoking opt with the returned
arguments.
clang -debug-pass will include -loop-distribute but the pass is invoked
with default=off so nothing happens unless the loop carries the pragma.
While through opt (default=on) we will try to distribute all loops.
This changes opt's default to off as well to match clang. The tests are
modified to explicitly enable the transformation.
llvm-svn: 290235
We're currently doing nearly the same thing for @llvm.objectsize in
three different places: two of them are missing checks for overflow,
and one of them could subtly break if InstCombine gets much smarter
about removing alloc sites. Seems like a good idea to not do that.
llvm-svn: 290214
This doesn't implement *every* feature of the existing inliner, but
tries to implement the most important ones for building a functional
optimization pipeline and beginning to sort out bugs, regressions, and
other problems.
Notable, but intentional omissions:
- No alloca merging support. Why? Because it isn't clear we want to do
this at all. Active discussion and investigation is going on to remove
it, so for simplicity I omitted it.
- No support for trying to iterate on "internally" devirtualized calls.
Why? Because it adds what I suspect is inappropriate coupling for
little or no benefit. We will have an outer iteration system that
tracks devirtualization including that from function passes and
iterates already. We should improve that rather than approximate it
here.
- Optimization remarks. Why? Purely to make the patch smaller, no other
reason at all.
The last one I'll probably work on almost immediately. But I wanted to
skip it in the initial patch to try to focus the change as much as
possible as there is already a lot of code moving around and both of
these *could* be skipped without really disrupting the core logic.
A summary of the different things happening here:
1) Adding the usual new PM class and rigging.
2) Fixing minor underlying assumptions in the inline cost analysis or
inline logic that don't generally hold in the new PM world.
3) Adding the core pass logic which is in essence a loop over the calls
in the nodes in the call graph. This is a bit duplicated from the old
inliner, but only a handful of lines could realistically be shared.
(I tried at first, and it really didn't help anything.) All told,
this is only about 100 lines of code, and most of that is the
mechanics of wiring up analyses from the new PM world.
4) Updating the LazyCallGraph (in the new PM) based on the *newly
inlined* calls and references. This is very minimal because we cannot
form cycles.
5) When inlining removes the last use of a function, eagerly nuking the
body of the function so that any "one use remaining" inline cost
heuristics are immediately refined, and queuing these functions to be
completely deleted once inlining is complete and the call graph
updated to reflect that they have become dead.
6) After all the inlining for a particular function, updating the
LazyCallGraph and the CGSCC pass manager to reflect the
function-local simplifications that are done immediately and
internally by the inline utilties. These are the exact same
fundamental set of CG updates done by arbitrary function passes.
7) Adding a bunch of test cases to specifically target CGSCC and other
subtle aspects in the new PM world.
Many thanks to the careful review from Easwaran and Sanjoy and others!
Differential Revision: https://reviews.llvm.org/D24226
llvm-svn: 290161
This patch implements PR31013 by introducing a
DIGlobalVariableExpression that holds a pair of DIGlobalVariable and
DIExpression.
Currently, DIGlobalVariables holds a DIExpression. This is not the
best way to model this:
(1) The DIGlobalVariable should describe the source level variable,
not how to get to its location.
(2) It makes it unsafe/hard to update the expressions when we call
replaceExpression on the DIGLobalVariable.
(3) It makes it impossible to represent a global variable that is in
more than one location (e.g., a variable with multiple
DW_OP_LLVM_fragment-s). We also moved away from attaching the
DIExpression to DILocalVariable for the same reasons.
This reapplies r289902 with additional testcase upgrades and a change
to the Bitcode record for DIGlobalVariable, that makes upgrading the
old format unambiguous also for variables without DIExpressions.
<rdar://problem/29250149>
https://llvm.org/bugs/show_bug.cgi?id=31013
Differential Revision: https://reviews.llvm.org/D26769
llvm-svn: 290153
Background/motivation - I was circling back around to:
https://llvm.org/bugs/show_bug.cgi?id=28296
I made a simple patch for that and noticed some regressions, so added test cases for
those with rL281055, and this is hopefully the minimal fix for just those cases.
But as you can see from the surrounding untouched folds, we are missing commuted patterns
all over the place, and of course there are no regression tests to cover any of those cases.
We could sprinkle "m_c_" dust all over this file and catch most of the missing folds, but
then we still wouldn't have test coverage, and we'd still miss some fraction of commuted
patterns because they require adjustments to the match order.
I'm aware of the concern about the potential compile-time performance impact of adding
matches like this (currently being discussed on llvm-dev), but I don't think there's any
evidence yet to suggest that handling commutative pattern matching more thoroughly is not
a worthwhile goal of InstCombine.
Differential Revision: https://reviews.llvm.org/D24419
llvm-svn: 290067
This is recommit of r287553 after fixing the invalid loop info after eliminating an empty block and unit test failures in AVR and WebAssembly :
Summary: Merging an empty case block into the header block of switch could cause ISel to add COPY instructions in the header of switch, instead of the case block, if the case block is used as an incoming block of a PHI. This could potentially increase dynamic instructions, especially when the switch is in a loop. I added a test case which was reduced from the benchmark I was targetting.
Reviewers: t.p.northover, mcrosier, manmanren, wmi, joerg, davidxl
Subscribers: joerg, qcolombet, danielcdh, hfinkel, mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D22696
llvm-svn: 289988
This reverts commit 289920 (again).
I forgot to implement a Bitcode upgrade for the case where a DIGlobalVariable
has not DIExpression. Unfortunately it is not possible to safely upgrade
these variables without adding a flag to the bitcode record indicating which
version they are.
My plan of record is to roll the planned follow-up patch that adds a
unit: field to DIGlobalVariable into this patch before recomitting.
This way we only need one Bitcode upgrade for both changes (with a
version flag in the bitcode record to safely distinguish the record
formats).
Sorry for the churn!
llvm-svn: 289982
This patch reapplies r289863. The original patch was reverted because it
exposed a bug causing the loop vectorizer to crash in the Python runtime on
PPC. The underlying issue was fixed with r289958.
llvm-svn: 289975
`dropUnknownNonDebugMetadata` takes a list of "known" metadata IDs. The
only reason it worked at all is that `getMetadataID` returns something
unrelated -- it returns the subclass ID of the receiver (which is used
in `dyn_cast` etc.). That does not numerically match
`LLVMContext::MD_invariant_group` and ends up dropping `invariant_group`
along with every other metadata that does not numerically match
`LLVMContext::MD_invariant_group`.
llvm-svn: 289973
After r288909, instructions feeding predicated instructions may be scalarized
if profitable. Since these instructions will remain scalar, we shouldn't
attempt to type-shrink them. We should only truncate vector types to their
minimal bit widths. This bug was exposed by enabling the vectorization of loops
containing conditional stores by default.
llvm-svn: 289958
This is recommit of r287553 after fixing the invalid loop info after eliminating an empty block:
Summary: Merging an empty case block into the header block of switch could cause ISel to add COPY instructions in the header of switch, instead of the case block, if the case block is used as an incoming block of a PHI. This could potentially increase dynamic instructions, especially when the switch is in a loop. I added a test case which was reduced from the benchmark I was targetting.
Reviewers: t.p.northover, mcrosier, manmanren, wmi, joerg, davidxl
Subscribers: joerg, qcolombet, danielcdh, hfinkel, mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D22696
llvm-svn: 289951
stores by default
This uncovers a crasher in the loop vectorizer on PPC when building the
Python runtime. I'll send the testcase to the review thread for the
original commit.
llvm-svn: 289934
This patch implements PR31013 by introducing a
DIGlobalVariableExpression that holds a pair of DIGlobalVariable and
DIExpression.
Currently, DIGlobalVariables holds a DIExpression. This is not the
best way to model this:
(1) The DIGlobalVariable should describe the source level variable,
not how to get to its location.
(2) It makes it unsafe/hard to update the expressions when we call
replaceExpression on the DIGLobalVariable.
(3) It makes it impossible to represent a global variable that is in
more than one location (e.g., a variable with multiple
DW_OP_LLVM_fragment-s). We also moved away from attaching the
DIExpression to DILocalVariable for the same reasons.
This reapplies r289902 with additional testcase upgrades.
<rdar://problem/29250149>
https://llvm.org/bugs/show_bug.cgi?id=31013
Differential Revision: https://reviews.llvm.org/D26769
llvm-svn: 289920
This patch implements PR31013 by introducing a
DIGlobalVariableExpression that holds a pair of DIGlobalVariable and
DIExpression.
Currently, DIGlobalVariables holds a DIExpression. This is not the
best way to model this:
(1) The DIGlobalVariable should describe the source level variable,
not how to get to its location.
(2) It makes it unsafe/hard to update the expressions when we call
replaceExpression on the DIGLobalVariable.
(3) It makes it impossible to represent a global variable that is in
more than one location (e.g., a variable with multiple
DW_OP_LLVM_fragment-s). We also moved away from attaching the
DIExpression to DILocalVariable for the same reasons.
<rdar://problem/29250149>
https://llvm.org/bugs/show_bug.cgi?id=31013
Differential Revision: https://reviews.llvm.org/D26769
llvm-svn: 289902
This pass prepares a module containing type metadata for ThinLTO by splitting
it into regular and thin LTO parts if possible, and writing both parts to
a multi-module bitcode file. Modules that do not contain type metadata are
written unmodified as a single module.
All globals with type metadata are added to the regular LTO module, and
the rest are added to the thin LTO module.
Differential Revision: https://reviews.llvm.org/D27324
llvm-svn: 289899
This patch sets the default value of the "-enable-cond-stores-vec" command line
option to "true".
Differential Revision: https://reviews.llvm.org/D27814
llvm-svn: 289863
Min/max canonicalization (r287585) exposes the fact that we're missing combines for min/max patterns.
This patch won't solve the example that was attached to that thread, so something else still needs fixing.
The line between InstCombine and InstSimplify gets blurry here because sometimes the icmp instruction that
we want to fold to already exists, but sometimes it's the swapped form of what we want.
Corresponding changes for smax/umin/umax to follow.
Differential Revision: https://reviews.llvm.org/D27531
llvm-svn: 289855
This is split out from D27696, since it turned out to be a bug fix and
not part of the NFC efficiency change.
Keep the same adjusted (possibly decayed) threshold in both the worklist
and the ImportList. Otherwise if we encountered it first along a cold
path, the callee would be added to the worklist with a lower decayed
threshold than when it is later encountered along a hot path. But the
logic uses the threshold recorded in the ImportList entry to check if
we should re-add it, and without this patch the threshold recorded there
is the same along both paths so we don't re-add it. Using the
same possibly decayed threshold in the ImportList ensures we re-add it
later with the higher non-decayed hot path threshold.
llvm-svn: 289843
A number of new patterns for simplifying and/xor of icmp:
(icmp ne %x, 0) ^ (icmp ne %y, 0) => icmp ne %x, %y if the following is true:
1- (%x = and %a, %mask) and (%y = and %b, %mask)
2- %mask is a power of 2.
(icmp eq %x, 0) & (icmp ne %y, 0) => icmp ult %x, %y if the following is true:
1- (%x = and %a, %mask1) and (%y = and %b, %mask2)
2- Let %t be the smallest power of 2 where %mask1 & %t != 0. Then for any
%s that is a power of 2 and %s & %mask2 != 0, we must have %s <= %t.
For example if %mask1 = 24 and %mask2 = 16, setting %s = 16 and %t = 8
violates condition (2) above. So this optimization cannot be applied.
llvm-svn: 289813
After r289755, the AssumptionCache is no longer needed. Variables affected by
assumptions are now found by using the new operand-bundle-based scheme. This
new scheme is more computationally efficient, and also we need much less
code...
llvm-svn: 289756
There was an efficiency problem with how we processed @llvm.assume in
ValueTracking (and other places). The AssumptionCache tracked all of the
assumptions in a given function. In order to find assumptions relevant to
computing known bits, etc. we searched every assumption in the function. For
ValueTracking, that means that we did O(#assumes * #values) work in InstCombine
and other passes (with a constant factor that can be quite large because we'd
repeat this search at every level of recursion of the analysis).
Several of us discussed this situation at the last developers' meeting, and
this implements the discussed solution: Make the values that an assume might
affect operands of the assume itself. To avoid exposing this detail to
frontends and passes that need not worry about it, I've used the new
operand-bundle feature to add these extra call "operands" in a way that does
not affect the intrinsic's signature. I think this solution is relatively
clean. InstCombine adds these extra operands based on what ValueTracking, LVI,
etc. will need and then those passes need only search the users of the values
under consideration. This should fix the computational-complexity problem.
At this point, no passes depend on the AssumptionCache, and so I'll remove
that as a follow-up change.
Differential Revision: https://reviews.llvm.org/D27259
llvm-svn: 289755
Joerg Sonnenberger