We have bailout hacks based on min/max in various places in instcombine
that shouldn't be necessary. The affected test was added for:
D48930
...which is a consequence of the improvement in:
D48584 (https://reviews.llvm.org/rL336172)
I'm assuming the visitTrunc bailout in this patch was added specifically
to avoid a change from SimplifyDemandedBits, so I'm just moving that
below the EvaluateInDifferentType optimization. A narrow min/max is still
a min/max.
llvm-svn: 336293
When creating `phi` instructions to resume at the scalar part of the loop,
copy the DebugLoc from the original phi over to the new one.
Differential Revision: https://reviews.llvm.org/D48769
llvm-svn: 336256
When zext is EvaluatedInDifferentType, InstCombine
drops the dbg.value intrinsic. This patch tries to
preserve said DI, by inserting the zext's old DI in the
resulting instruction. (Only for integer type for now)
Differential Revision: https://reviews.llvm.org/D48331
llvm-svn: 336254
As the test diffs show, the current users of getBinOpIdentity()
are InstCombine and Reassociate. SLP vectorizer is a candidate
for using this functionality too (D28907).
The InstCombine shuffle improvements are part of the planned
enhancements noted in D48830.
InstCombine actually has several other uses of getBinOpIdentity()
via SimplifyUsingDistributiveLaws(), but we don't call that for
any FP ops. Fixing that might be another part of removing the
custom reassociation in InstCombine that is only done for fadd+fmul.
llvm-svn: 336215
This is the last significant change suggested in PR37806:
https://bugs.llvm.org/show_bug.cgi?id=37806#c5
...though there are several follow-ups noted in the code comments
in this patch to complete this transform.
It's possible that a binop feeding a select-shuffle has been eliminated
by earlier transforms (or the code was just written like this in the 1st
place), so we'll fail to match the patterns that have 2 binops from:
D48401,
D48678,
D48662,
D48485.
In that case, we can try to materialize identity constants for the remaining
binop to fill in the "ghost" lanes of the vector (where we just want to pass
through the original values of the source operand).
I added comments to ConstantExpr::getBinOpIdentity() to show planned follow-ups.
For now, we only handle the 5 commutative integer binops (add/mul/and/or/xor).
Differential Revision: https://reviews.llvm.org/D48830
llvm-svn: 336196
Summary:
When salvaging a dbg.declare/dbg.addr we should not add
DW_OP_stack_value to the DIExpression
(see test/Transforms/InstCombine/salvage-dbg-declare.ll).
Consider this example
%vla = alloca i32, i64 2
call void @llvm.dbg.declare(metadata i32* %vla, metadata !1, metadata !DIExpression())
Instcombine will turn it into
%vla1 = alloca [2 x i32]
%vla1.sub = getelementptr inbounds [2 x i32], [2 x i32]* %vla, i64 0, i64 0
call void @llvm.dbg.declare(metadata [2 x i32]* %vla1.sub, metadata !19, metadata !DIExpression())
If the GEP can be eliminated, then the dbg.declare will be salvaged
and we should get
%vla1 = alloca [2 x i32]
call void @llvm.dbg.declare(metadata [2 x i32]* %vla1, metadata !19, metadata !DIExpression())
The problem was that salvageDebugInfo did not recognize dbg.declare
as being indirect (%vla1 points to the value, it does not hold the
value), so we incorrectly got
call void @llvm.dbg.declare(metadata [2 x i32]* %vla1, metadata !19, metadata !DIExpression(DW_OP_stack_value))
I also made sure that llvm::salvageDebugInfo and
DIExpression::prependOpcodes do not add DW_OP_stack_value to
the DIExpression in case no new operands are added to the
DIExpression. That way we avoid to, unneccessarily, turn a
register location expression into an implicit location expression
in some situations (see test11 in test/Transforms/LICM/sinking.ll).
Reviewers: aprantl, vsk
Reviewed By: aprantl, vsk
Subscribers: JDevlieghere, llvm-commits
Differential Revision: https://reviews.llvm.org/D48837
llvm-svn: 336191
unswitching loops.
Original patch trying to address this was sent in D47624, but that
didn't quite handle things correctly. There are two key principles used
to select whether and how to invalidate SCEV-cached information about
loops:
1) We must invalidate any info SCEV has cached before unswitching as we
may change (or destroy) the loop structure by the act of unswitching,
and make it hard to recover everything we want to invalidate within
SCEV.
2) We need to invalidate all of the loops whose CFGs are mutated by the
unswitching. Notably, this isn't the *entire* loop nest, this is
every loop contained by the outermost loop reached by an exit block
relevant to the unswitch.
And we need to do this even when doing trivial unswitching.
I've added more focused tests that directly check that SCEV starts off
with imprecise information and after unswitching (and simplifying
instructions) re-querying SCEV will produce precise information. These
tests also specifically work to check that an *outer* loop's information
becomes precise.
However, the testing here is still a bit imperfect. Crafting test cases
that reliably fail to be analyzed by SCEV before unswitching and succeed
afterward proved ... very, very hard. It took me several hours and
careful work to build these, and I'm not optimistic about necessarily
coming up with more to cover more elaborate possibilities. Fortunately,
the code pattern we are testing here in the pass is really
straightforward and reliable.
Thanks to Max Kazantsev for the initial work on this as well as the
review, and to Hal Finkel for helping me talk through approaches to test
this stuff even if it didn't come to much.
Differential Revision: https://reviews.llvm.org/D47624
llvm-svn: 336183
This patch changes order of transform in InstCombineCompares to avoid
performing transforms based on ranges which produce complex bit arithmetics
before more simple things (like folding with constants) are done. See PR37636
for the motivating example.
Differential Revision: https://reviews.llvm.org/D48584
Reviewed By: spatel, lebedev.ri
llvm-svn: 336172
Summary:
Comment on Transforms/LoopVersioning/incorrect-phi.ll: With the change
SCEV is able to prove that the loop doesn't wrap-self (due to zext i16
to i64), disabling the entire loop versioning pass. Removed the zext and
just use i64.
Reviewers: sanjoy
Subscribers: jlebar, hiraditya, javed.absar, bixia, llvm-commits
Differential Revision: https://reviews.llvm.org/D48409
llvm-svn: 336140
Summary: It is common to have the following min/max pattern during the intermediate stages of SLP since we only optimize at the end. This patch tries to catch such patterns and allow more vectorization.
%1 = extractelement <2 x i32> %a, i32 0
%2 = extractelement <2 x i32> %a, i32 1
%cond = icmp sgt i32 %1, %2
%3 = extractelement <2 x i32> %a, i32 0
%4 = extractelement <2 x i32> %a, i32 1
%select = select i1 %cond, i32 %3, i32 %4
Author: FarhanaAleen
Reviewed By: ABataev, RKSimon, spatel
Differential Revision: https://reviews.llvm.org/D47608
llvm-svn: 336130
This extends D48485 to allow another pair of binops (add/or) to be combined either
with or without a leading shuffle:
or X, C --> add X, C (when X and C have no common bits set)
Here, we need value tracking to determine that the 'or' can be reversed into an 'add',
and we've added general infrastructure to allow extending to other opcodes or moving
to where other passes could use that functionality.
Differential Revision: https://reviews.llvm.org/D48662
llvm-svn: 336128
Due to current limitations in constant analysis, we need flags
on add or mul to show propagation for the potential transform
suggested in these tests (no other binops currently report
identity constants).
llvm-svn: 336101
This version contains a fix to add values for which the state in ParamState change
to the worklist if the state in ValueState did not change. To avoid adding the
same value multiple times, mergeInValue returns true, if it added the value to
the worklist. The value is added to the worklist depending on its state in
ValueState.
Original message:
For comparisons with parameters, we can use the ParamState lattice
elements which also provide constant range information. This improves
the code for PR33253 further and gets us closer to use
ValueLatticeElement for all values.
Also, as we are using the range information in the solver directly, we
do not need tryToReplaceWithConstantRange afterwards anymore.
Reviewers: dberlin, mssimpso, davide, efriedma
Reviewed By: mssimpso
Differential Revision: https://reviews.llvm.org/D43762
llvm-svn: 336098
We were always using the opcodes of the first 2 scalars for the costs of the alternate opcode + shuffle. This made sense when we used SK_Alternate and opcodes were guaranteed to be alternating, but this fails for the more general SK_Select case.
This fix exposes an issue demonstrated by the fmul_fdiv_v4f32_const test - the SLM model has v4f32 fdiv costs which are more than twice those of the f32 scalar cost, meaning that the cost model determines that the vectorization is not performant. Unfortunately it completely ignores the fact that the fdiv by a constant will be changed into a fmul by InstCombine for a much lower cost vectorization. But at least we're seeing this now...
llvm-svn: 336095
Summary:
This patch introduce new intrinsic -
strip.invariant.group that was described in the
RFC: Devirtualization v2
Reviewers: rsmith, hfinkel, nlopes, sanjoy, amharc, kuhar
Subscribers: arsenm, nhaehnle, JDevlieghere, hiraditya, xbolva00, llvm-commits
Differential Revision: https://reviews.llvm.org/D47103
Co-authored-by: Krzysztof Pszeniczny <krzysztof.pszeniczny@gmail.com>
llvm-svn: 336073
This is a simple implementation of the unroll-and-jam classical loop
optimisation.
The basic idea is that we take an outer loop of the form:
for i..
ForeBlocks(i)
for j..
SubLoopBlocks(i, j)
AftBlocks(i)
Instead of doing normal inner or outer unrolling, we unroll as follows:
for i... i+=2
ForeBlocks(i)
ForeBlocks(i+1)
for j..
SubLoopBlocks(i, j)
SubLoopBlocks(i+1, j)
AftBlocks(i)
AftBlocks(i+1)
Remainder Loop
So we have unrolled the outer loop, then jammed the two inner loops into
one. This can lead to a simpler inner loop if memory accesses can be shared
between the now jammed loops.
To do this we have to prove that this is all safe, both for the memory
accesses (using dependence analysis) and that ForeBlocks(i+1) can move before
AftBlocks(i) and SubLoopBlocks(i, j).
Differential Revision: https://reviews.llvm.org/D41953
llvm-svn: 336062
Alternate opcode handling only supports binary operators, these tests demonstrate missed opportunities to vectorize some sitofp/uitofp and fptosi/fptoui style casts as well as some (successful) float bits manipulations
llvm-svn: 336060
This was discussed in D48401 as another improvement for:
https://bugs.llvm.org/show_bug.cgi?id=37806
If we have 2 different variable values, then we shuffle (select) those lanes,
shuffle (select) the constants, and then perform the binop. This eliminates a binop.
The new shuffle uses the same shuffle mask as the existing shuffle, so there's no
danger of creating a difficult shuffle.
All of the earlier constraints still apply, but we also check for extra uses to
avoid creating more instructions than we'll remove.
Additionally, we're disallowing the fold for div/rem because that could expose a
UB hole.
Differential Revision: https://reviews.llvm.org/D48678
llvm-svn: 335974
Summary:
An alternative to D48597.
Fixes [[ https://bugs.llvm.org/show_bug.cgi?id=37936 | PR37936 ]].
The problem is as follows:
1. `indvars` marks `%dec` as `NUW`.
2. `loop-instsimplify` runs `instsimplify`, which constant-folds `%dec` to -1 (D47908)
3. `loop-reduce` tries to do some further modification, but crashes
with an type assertion in cast, because `%dec` is no longer an `Instruction`,
If the runline is split into two, i.e. you first run `-indvars -loop-instsimplify`,
store that into a file, and then run `-loop-reduce`, there is no crash.
So it looks like the problem is due to `-loop-instsimplify` not discarding SCEV.
But in this case we can just not crash if it's not an `Instruction`.
This is just a local fix, unlike D48597, so there may very well be other problems.
Reviewers: mkazantsev, uabelho, sanjoy, silviu.baranga, wmi
Reviewed By: mkazantsev
Subscribers: evstupac, javed.absar, spatel, llvm-commits
Differential Revision: https://reviews.llvm.org/D48599
llvm-svn: 335950
Summary:
The InlinerFunctionImportStats will collect and dump stats regarding how
many function inlined into the module were imported by ThinLTO.
Reviewers: wmi, dexonsmith
Subscribers: mehdi_amini, inglorion, llvm-commits, eraman
Differential Revision: https://reviews.llvm.org/D48729
llvm-svn: 335914
When rewriting an alloca partition copy the DL from the
old alloca over the the new one.
Differential Revision: https://reviews.llvm.org/D48640
llvm-svn: 335904
This is an enhancement to D48401 that was discussed in:
https://bugs.llvm.org/show_bug.cgi?id=37806
We can convert a shift-left-by-constant into a multiply (we canonicalize IR in the other
direction because that's generally better of course). This allows us to remove the shuffle
as we do in the regular opcodes-are-the-same cases.
This requires a small hack to make sure we don't introduce any extra poison:
https://rise4fun.com/Alive/ZGv
Other examples of opcodes where this would work are add+sub and fadd+fsub, but we already
canonicalize those subs into adds, so there's nothing to do for those cases AFAICT. There
are planned enhancements for opcode transforms such or -> add.
Note that there's a different fold needed if we've already managed to simplify away a binop
as seen in the test based on PR37806, but we manage to get that one case here because this
fold is positioned above the demanded elements fold currently.
Differential Revision: https://reviews.llvm.org/D48485
llvm-svn: 335888
SCCP does not change the CFG, so we can mark it as preserved.
Reviewers: dberlin, efriedma, davide
Reviewed By: davide
Differential Revision: https://reviews.llvm.org/D47149
llvm-svn: 335820
If a trunc has a user in a block which is not reachable from entry,
we can safely perform trunc elimination as if this user didn't exist.
llvm-svn: 335816
I think the intrinsics named 'avx512.mask.' should refer to the previous behavior of taking a mask argument in the intrinsic instead of using a 'select' or 'and' instruction in IR to accomplish the masking. This is more consistent with the goal that eventually we will have no intrinsics that have masking builtin. When we reach that goal, we should have no intrinsics named "avx512.mask".
llvm-svn: 335744
This patch adds a custom trunc store lowering for v4i8 vector types.
Since there is not v.4b register, the v4i8 is promoted to v4i16 (v.4h)
and default action for v4i8 is to extract each element and issue 4
byte stores.
A better strategy would be to extended the promoted v4i16 to v8i16
(with undef elements) and extract and store the word lane which
represents the v4i8 subvectores. The construction:
define void @foo(<4 x i16> %x, i8* nocapture %p) {
%0 = trunc <4 x i16> %x to <4 x i8>
%1 = bitcast i8* %p to <4 x i8>*
store <4 x i8> %0, <4 x i8>* %1, align 4, !tbaa !2
ret void
}
Can be optimized from:
umov w8, v0.h[3]
umov w9, v0.h[2]
umov w10, v0.h[1]
umov w11, v0.h[0]
strb w8, [x0, #3]
strb w9, [x0, #2]
strb w10, [x0, #1]
strb w11, [x0]
ret
To:
xtn v0.8b, v0.8h
str s0, [x0]
ret
The patch also adjust the memory cost for autovectorization, so the C
code:
void foo (const int *src, int width, unsigned char *dst)
{
for (int i = 0; i < width; i++)
*dst++ = *src++;
}
can be vectorized to:
.LBB0_4: // %vector.body
// =>This Inner Loop Header: Depth=1
ldr q0, [x0], #16
subs x12, x12, #4 // =4
xtn v0.4h, v0.4s
xtn v0.8b, v0.8h
st1 { v0.s }[0], [x2], #4
b.ne .LBB0_4
Instead of byte operations.
llvm-svn: 335735
This prevents InstCombine from creating mis-sized dbg.values when
replacing a sequence of casts with a simpler cast. For example, in:
(fptrunc (floor (fpext X))) -> (floorf X)
We no longer emit dbg.value(X) (with a 32-bit float operand) to describe
(fpext X) (which is a 64-bit float).
This was diagnosed by the debugify check added in r335682.
llvm-svn: 335696
It's not possible to get the fragment size of some dbg.values. Teach the
mis-sized dbg.value diagnostic to detect this scenario and bail out.
Tested with:
$ find test/Transforms -print -exec opt -debugify-each -instcombine {} \;
llvm-svn: 335695
Summary:
When recording uses we need to rewrite after cloning a loop we need to
check if the use is not dominated by the original def. The initial
assumption was that the cloned basic block will introduce a new path and
thus the original def will only dominate the use if they are in the same
BB, but as the reproducer from PR37745 shows it's not always the case.
This fixes PR37745.
Reviewers: haicheng, Ka-Ka
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D48111
llvm-svn: 335675
Not sure why this logic seems to be repeated in 2 different places,
one called by the other.
On AMDGPU addrspace(3) globals start allocating at 0, so these
checks will be incorrect (not that real code actually tries
to compare these addresses)
llvm-svn: 335649
I'm not sure why the code here is skipping calls since
TTI does try to do something for general calls, but it
at least should allow intrinsics.
Skip intrinsics that should not be omitted as calls, which
is by far the most common case on AMDGPU.
llvm-svn: 335645
Similar to other patches in this series:
https://reviews.llvm.org/rL335512https://reviews.llvm.org/rL335527https://reviews.llvm.org/rL335597https://reviews.llvm.org/rL335616
...this is filling a gap in analysis that is exposed by an unrelated select-of-constants transform.
I didn't see a way to unify the sext cases because each div/rem opcode results in a different fold.
Note that in this case, the backend might want to convert the select into math:
Name: sext urem
%e = sext i1 %x to i32
%r = urem i32 %y, %e
=>
%c = icmp eq i32 %y, -1
%z = zext i1 %c to i32
%r = add i32 %z, %y
llvm-svn: 335622
Since D46637 we are better at handling uniform/non-uniform constant Pow2 detection; this patch tweaks the SLP argument handling to support them.
As SLP works with arrays of values I don't think we can easily use the pattern match helpers here.
Differential Revision: https://reviews.llvm.org/D48214
llvm-svn: 335621
Note: I didn't add a hasOneUse() check because the existing,
related fold doesn't have that check. I suspect that the
improved analysis and codegen make these some of the rare
canonicalization cases where we allow an increase in
instructions.
llvm-svn: 335597
changeToUnreachable may remove PHI nodes from executable blocks we found values
for and we would fail to replace them. By changing dead blocks to unreachable after
we replaced constants in all executable blocks, we ensure such PHI nodes are replaced
by their known value before.
Fixes PR37780.
Reviewers: efriedma, davide
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D48421
llvm-svn: 335588
Summary:
This is a follow-up to r334830 and r335031.
In the valueCoversEntireFragment check we now also handle
the situation when there is a variable length array (VLA)
involved, and the length of the array has been reduced to
a constant.
The ConvertDebugDeclareToDebugValue functions that are related
to PHI nodes and load instructions now avoid inserting dbg.value
intrinsics when the value does not, for certain, cover the
variable/fragment that should be described.
In r334830 we assumed that the value always covered the entire
var/fragment and we had assertions in the code to show that
assumption. However, those asserts failed when compiling code
with VLAs, so we removed the asserts in r335031. Now when we
know that the valueCoversEntireFragment check can fail also for
PHI/Load instructions we avoid to insert the faulty dbg.value
intrinsic in such situations. Compared to the Store instruction
scenario we simply drop the dbg.value here (as the variable does
not change its value due to PHI/Load, so an earlier dbg.value
describing the variable should still be valid).
Reviewers: aprantl, vsk, efriedma
Reviewed By: aprantl
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D48547
llvm-svn: 335580
Turn canonicalized subtraction back into (-1 - B) and combine it with (A + 1) into (A - B).
This is similar to the folding already done for (B ^ -1) + Const into (-1 + Const) - B.
Differential Revision: https://reviews.llvm.org/D48535
llvm-svn: 335579
unswitching of switches.
This works much like trivial unswitching of switches in that it reliably
moves the switch out of the loop. Here we potentially clone the entire
loop into each successor of the switch and re-point the cases at these
clones.
Due to the complexity of actually doing nontrivial unswitching, this
patch doesn't create a dedicated routine for handling switches -- it
would duplicate far too much code. Instead, it generalizes the existing
routine to handle both branches and switches as it largely reduces to
looping in a few places instead of doing something once. This actually
improves the results in some cases with branches due to being much more
careful about how dead regions of code are managed. With branches,
because exactly one clone is created and there are exactly two edges
considered, somewhat sloppy handling of the dead regions of code was
sufficient in most cases. But with switches, there are much more
complicated patterns of dead code and so I've had to move to a more
robust model generally. We still do as much pruning of the dead code
early as possible because that allows us to avoid even cloning the code.
This also surfaced another problem with nontrivial unswitching before
which is that we weren't as precise in reconstructing loops as we could
have been. This seems to have been mostly harmless, but resulted in
pointless LCSSA PHI nodes and other unnecessary cruft. With switches, we
have to get this *right*, and everything benefits from it.
While the testing may seem a bit light here because we only have two
real cases with actual switches, they do a surprisingly good job of
exercising numerous edge cases. Also, because we share the logic with
branches, most of the changes in this patch are reasonably well covered
by existing tests.
The new unswitch now has all of the same fundamental power as the old
one with the exception of the single unsound case of *partial* switch
unswitching -- that really is just loop specialization and not
unswitching at all. It doesn't fit into the canonicalization model in
any way. We can add a loop specialization pass that runs late based on
profile data if important test cases ever come up here.
Differential Revision: https://reviews.llvm.org/D47683
llvm-svn: 335553
This avoids creating unnecessary casts if the IP used to be a dbg info
intrinsic. Fixes PR37727.
Reviewers: vsk, aprantl, sanjoy, efriedma
Reviewed By: vsk, efriedma
Differential Revision: https://reviews.llvm.org/D47874
llvm-svn: 335513
We canonicalize to select with a zext-add and either zext-sub or sext-sub,
so this shows a pattern that's not conforming to the general trend.
llvm-svn: 335506
FDiv is replaced with multiplication by reciprocal and invariant
reciprocal is hoisted out of the loop, while multiplication remains
even if invariant.
Switch checks for all invariant operands and only invariant
denominator to fix the issue.
Differential Revision: https://reviews.llvm.org/D48447
llvm-svn: 335411
This gets rid of a bunch of weird special cases; instead, just use SCEV
rewriting for everything. In addition to being simpler, this fixes a
bug where we would use the wrong stride in certain edge cases.
The one bit I'm not quite sure about is the trip count handling,
specifically the FIXME about overflow. In general, I think we need to
widen the exit condition, but that's probably not profitable if the new
type isn't legal, so we probably need a check somewhere. That said, I
don't think I'm making the existing problem any worse.
As a followup to this, a bunch of IV-related code in root-finding could
be cleaned up; with SCEV-based rewriting, there isn't any reason to
assume a loop will have exactly one or two PHI nodes.
Differential Revision: https://reviews.llvm.org/D45191
llvm-svn: 335400
Enable tryToVectorizeList to support InstructionsState alternate opcode patterns at a root (build vector etc.) as well as further down the vectorization tree.
NOTE: This patch reduces some of the debug reporting if there are opcode mismatches - I can try to add it back if it proves a problem. But it could get rather messy trying to provide equivalent verbose debug strings via getSameOpcode etc.
Differential Revision: https://reviews.llvm.org/D48488
llvm-svn: 335364
SLP currently only accepts (F)Add/(F)Sub alternate counterpart ops to be merged into an alternate shuffle.
This patch relaxes this to accept any pair of BinaryOperator opcodes instead, assuming the target's cost model accepts the vectorization+shuffle.
Differential Revision: https://reviews.llvm.org/D48477
llvm-svn: 335349
This one shows another pattern that we'll need to match
in some cases, but the current ordering of folds allows
us to match this as 2 binops before simplification takes
place.
llvm-svn: 335347
AArch64 was only setting costs for SK_Transpose, which meant that many of the simpler shuffles (e.g. SK_Select and SK_PermuteSingleSrc for larger vector elements) was being severely overestimated by the default shuffle expansion.
This patch adds costs to help improve SLP performance and avoid a regression in reductions introduced by D48174.
I'm not very knowledgeable about AArch64 shuffle lowering so I've kept the extra costs to a minimum - someone who knows this code can add extra costs which should improve vectorization a lot more.
Differential Revision: https://reviews.llvm.org/D48172
llvm-svn: 335329
clear out deleted loops from the current queue beyond just the current
loop.
This is important because SimpleLoopUnswitch will now enqueue the same
loop to be re-processed. When it does this with the legacy PM, we don't
have a way of canceling the rest of the pipeline and so we can end up
deleting the loop before we reprocess it. =/
This change also makes it easy to support deleting other loops in the
queue to process, although I don't have any use cases for that.
Differential Revision: https://reviews.llvm.org/D48470
llvm-svn: 335317
With non-commutative binops, we could be using the same
variable value as operand 0 in 1 binop and operand 1 in
the other, so we have to check for that possibility and
bail out.
llvm-svn: 335312
Summary:
A reprise of D25849.
This crash was found through fuzzing some time ago and was documented in PR28879.
No check for load size has been added due to the following tests:
- Transforms/GVN/invariant.group.ll
- Transforms/GVN/pr10820.ll
These tests expect load sizes that are not a multiple of eight.
Thanks to @davide for the original patch.
Reviewers: nlopes, davide, RKSimon, reames, efriedma
Reviewed By: efriedma
Subscribers: davide, llvm-commits, Prazek
Differential Revision: https://reviews.llvm.org/D48330
llvm-svn: 335294
This is the simplest case from PR37806:
https://bugs.llvm.org/show_bug.cgi?id=37806
If we have a common variable operand used in a pair of binops with vector constants
that are vector selected together, then we can constant shuffle the constant vectors
to eliminate the shuffle instruction.
This has some tricky parts that are hopefully addressed in the tests and their
respective comments:
1. If the shuffle mask contains an undef element, then that lane of the result is
undef:
http://llvm.org/docs/LangRef.html#shufflevector-instruction
Therefore, we can replace the constant in that lane with an undef value except
for div/rem. With div/rem, an undef in the divisor would cause the whole op to
be undef. So I'm using the same hack as in D47686 - replace the undefs with '1'.
2. Intersect the wrapping and FMF of the original binops for the new binop. There
should be no extra poison or fast-math potential in the new binop that wasn't
possible in the original code.
3. Disregard other uses. Given that we're eliminating uses (shortening the
dependency chain), I think that's always the right IR canonicalization. But
I purposely chose the udiv test to demonstrate the scenario where both
intermediate values have other uses because that seems likely worse for
codegen with an expensive math op. This seems like a very rare possibility to
me, so I don't think it requires a backend patch first.
Differential Revision: https://reviews.llvm.org/D48401
llvm-svn: 335283
This reverts commit r335206.
As discussed here: https://reviews.llvm.org/rL333740, a fix will come
tomorrow. In the meanwhile, revert this to fix some bots.
llvm-svn: 335272
The previous code worked with vectors, but it failed when the
vector constants contained undef elements.
The matchers handle those cases.
llvm-svn: 335262
Summary:
Use the expanded features of the TableGen generic tables to avoid manually
adding the combinatorially exploded set of intrinsics. The
getAMDGPUImageDimIntrinsic lookup function is early-out,
i.e. non-AMDGPU intrinsics will never look at the underlying table.
Use a generic approach for getting the new intrinsic overload to keep the
code simple, and make the image dmask handling more generic:
- handle non-sampler image loads
- handle the case where the set of demanded elements is not a prefix
There is some overlap between this code and an optimization that happens
in the backend during code generation. They currently complement each other:
- only the codegen optimization can generate vec3 loads
- only the InstCombine optimization can handle D16
The InstCombine optimization also likely covers more cases since the
codegen optimization is fairly ad-hoc. Ideally, we'll remove the optimization
in codegen once the infrastructure for vec3 is in place (which will probably
take a long time).
Modify the test cases to use dimension-aware intrinsics. This makes it
easier to see that the test coverage for the new intrinsics is equivalent,
and the old style intrinsics will be removed in a follow-up commit anyway.
Change-Id: I4b91ea661413d13004956fe4ef7d13d41b8ce3ad
Reviewers: arsenm, rampitec, majnemer
Subscribers: kzhuravl, wdng, mgorny, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D48165
llvm-svn: 335230
This enables da-delinearize in Dependence Analysis for delinearizing array
accesses into multiple dimensions. This can help to increase the power of
Dependence analysis on multi-dimensional arrays and prevent having to fall
back to the slower and less accurate MIV tests. It adds static checks on the
bounds of the arrays to ensure that one dimension doesn't overflow into
another, and brings our code in line with our tests.
Differential Revision: https://reviews.llvm.org/D45872
llvm-svn: 335217
These were being over cautious for costs for one/two op general shuffles - VSHUFPD doesn't have to replicate the same shuffle in both lanes like VSHUFPS does.
llvm-svn: 335216
r335150 should resolve the issues with the clang-with-thin-lto-ubuntu
and clang-with-lto-ubuntu builders.
Original message:
This patch updates IPSCCP to use PredicateInfo to propagate
facts to true branches predicated by EQ and to false branches
predicated by NE.
As a follow up, we should be able to extend it to also propagate additional
facts about nonnull.
Reviewers: davide, mssimpso, dberlin, efriedma
Reviewed By: davide, dberlin
llvm-svn: 335206
conditions feeding a chain of `and`s or `or`s for a branch.
Much like with full non-trivial unswitching, we rely on the pass manager
to handle iterating until all of the profitable unswitches have been
done. This is to allow other more profitable unswitches to fire on any
of the cloned, simpler versions of the loop if viable.
Threading the partial unswiching through the non-trivial unswitching
logic motivated some minor refactorings. If those are too disruptive to
make it reasonable to review this patch, I can separate them out, but
it'll be somewhat timeconsuming so I wanted to send it for initial
review as-is. Feel free to tell me whether it warrants pulling apart.
I've tried to re-use (and factor out) logic form the partial trivial
unswitching, but not as much could be shared as I had haped. Still, this
wasn't as bad as I naively expected.
Some basic testing is added, but I probably need more. Suggestions for
things you'd like to see tested more than welcome. One thing I'd like to
do is add some testing that when we schedule this with loop-instsimplify
it effectively cleans up the cruft created.
Last but not least, this uncovered a bug that has been in loop cloning
the entire time for non-trivial unswitching. Specifically, we didn't
correctly add the outer-most cloned loop to the list of cloned loops.
This meant that LCSSA wouldn't be updated for it hypothetically, and
more significantly that we would never visit it in the loop pass
manager. I noticed this while checking loop-instsimplify by hand. I'll
try to separate this bugfix out into its own patch with a more focused
test. But it is just one line, so shouldn't significantly confuse the
review here.
After this patch, the only missing "feature" in this unswitch I'm aware
of us non-trivial unswitching of switches. I'll try implementing *full*
non-trivial unswitching of switches (which is at least a sound thing to
implement), but *partial* non-trivial unswitching of switches is
something I don't see any sound and principled way to implement. I also
have no interesting test cases for the latter, so I'm not really
worried. The rest of the things that need to be ported are bug-fixes and
more narrow / targeted support for specific issues.
Differential Revision: https://reviews.llvm.org/D47522
llvm-svn: 335203
Summary:
Two utils methods have essentially the same functionality. This is an attempt to merge them into one.
1. lib/Transforms/Utils/Local.cpp : MergeBasicBlockIntoOnlyPred
2. lib/Transforms/Utils/BasicBlockUtils.cpp : MergeBlockIntoPredecessor
Prior to the patch:
1. MergeBasicBlockIntoOnlyPred
Updates either DomTree or DeferredDominance
Moves all instructions from Pred to BB, deletes Pred
Asserts BB has single predecessor
If address was taken, replace the block address with constant 1 (?)
2. MergeBlockIntoPredecessor
Updates DomTree, LoopInfo and MemoryDependenceResults
Moves all instruction from BB to Pred, deletes BB
Returns if doesn't have a single predecessor
Returns if BB's address was taken
After the patch:
Method 2. MergeBlockIntoPredecessor is attempting to become the new default:
Updates DomTree or DeferredDominance, and LoopInfo and MemoryDependenceResults
Moves all instruction from BB to Pred, deletes BB
Returns if doesn't have a single predecessor
Returns if BB's address was taken
Uses of MergeBasicBlockIntoOnlyPred that need to be replaced:
1. lib/Transforms/Scalar/LoopSimplifyCFG.cpp
Updated in this patch. No challenges.
2. lib/CodeGen/CodeGenPrepare.cpp
Updated in this patch.
i. eliminateFallThrough is straightforward, but I added using a temporary array to avoid the iterator invalidation.
ii. eliminateMostlyEmptyBlock(s) methods also now use a temporary array for blocks
Some interesting aspects:
- Since Pred is not deleted (BB is), the entry block does not need updating.
- The entry block was being updated with the deleted block in eliminateMostlyEmptyBlock. Added assert to make obvious that BB=SinglePred.
- isMergingEmptyBlockProfitable assumes BB is the one to be deleted.
- eliminateMostlyEmptyBlock(BB) does not delete BB on one path, it deletes its unique predecessor instead.
- adding some test owner as subscribers for the interesting tests modified:
test/CodeGen/X86/avx-cmp.ll
test/CodeGen/AMDGPU/nested-loop-conditions.ll
test/CodeGen/AMDGPU/si-annotate-cf.ll
test/CodeGen/X86/hoist-spill.ll
test/CodeGen/X86/2006-11-17-IllegalMove.ll
3. lib/Transforms/Scalar/JumpThreading.cpp
Not covered in this patch. It is the only use case using the DeferredDominance.
I would defer to Brian Rzycki to make this replacement.
Reviewers: chandlerc, spatel, davide, brzycki, bkramer, javed.absar
Subscribers: qcolombet, sanjoy, nemanjai, nhaehnle, jlebar, tpr, kbarton, RKSimon, wmi, arsenm, llvm-commits
Differential Revision: https://reviews.llvm.org/D48202
llvm-svn: 335183
The idea of partial unswitching is to take a *part* of a branch's
condition that is loop invariant and just unswitching that part. This
primarily makes sense with i1 conditions of branches as opposed to
switches. When dealing with i1 conditions, we can easily extract loop
invariant inputs to a a branch and unswitch them to test them entirely
outside the loop.
As part of this, we now create much more significant cruft in the loop
body, so this relies on adding cleanup passes to the loop pipeline and
revisiting unswitched loops to do that cleanup before continuing to
process them.
This already appears to be more powerful at unswitching than the old
loop unswitch pass, and so I'd appreciate pretty careful review in case
I'm just missing some correctness checks. The `LIV-loop-condition` test
case is not unswitched by the old unswitch pass, but is with this pass.
Thanks to Sanjoy and Fedor for the review!
Differential Revision: https://reviews.llvm.org/D46706
llvm-svn: 335156
Using OrderedInstructions::dominates as comparator for instructions in
BBs without dominance relation can cause a non-deterministic order
between such instructions. That in turn can cause us to materialize
copies in a non-deterministic order. While this does not effect
correctness, it causes some minor non-determinism in the final generated
code, because values have slightly different labels.
Without this patch, running -print-predicateinfo on a reasonably large
module produces slightly different output on each run.
This patch uses the dominator trees DFSInNum to order instruction from
different BBs, which should enforce a deterministic ordering and
guarantee that dominated instructions come after the instructions that
dominate them.
Reviewers: dberlin, efriedma, davide
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D48230
llvm-svn: 335150
D47985 saw the old SK_Alternate 'alternating' shuffle mask replaced with the SK_Select mask which accepts either input operand for each lane, equivalent to a vector select with a constant condition operand.
This patch updates SLPVectorizer to make full use of this SK_Select shuffle pattern by removing the 'isOdd()' limitation.
The AArch64 regression will be fixed by D48172.
Differential Revision: https://reviews.llvm.org/D48174
llvm-svn: 335130
For both operands are unsigned, the following optimizations are valid, and missing:
1. X > Y && X != 0 --> X > Y
2. X > Y || X != 0 --> X != 0
3. X <= Y || X != 0 --> true
4. X <= Y || X == 0 --> X <= Y
5. X > Y && X == 0 --> false
unsigned foo(unsigned x, unsigned y) { return x > y && x != 0; }
should fold to x > y, but I found we haven't done it right now.
besides, unsigned foo(unsigned x, unsigned y) { return x < y && y != 0; }
Has been folded to x < y, so there may be a bug.
Patch by: Li Jia He!
Differential Revision: https://reviews.llvm.org/D47922
llvm-svn: 335129
These are the baseline tests for the functional change in D47922.
Patch by Li Jia He!
Differential Revision: https://reviews.llvm.org/D48000
llvm-svn: 335128
This patch replaces calls to X86-specific intrinsics with floor-ceil semantics
with calls to target-independent @llvm.floor.* and @llvm.ceil.* intrinsics. This
doesn't affect the resulting machine code, as those intrinsics are lowered to
the same instructions, but exposes these specific rounding cases to generic
optimizations.
Differential Revision: https://reviews.llvm.org/D48067
llvm-svn: 335039
LoopSimplifyCFG, being a loop pass, needs to preserve scalar
evolution. This invalidates SE for the loops altered during
block merging.
Differential Revision: https://reviews.llvm.org/D48258
llvm-svn: 335036
This patch adds logic to deal with the following constructions:
%iv = phi i64 ...
%trunc = trunc i64 %iv to i32
%cmp = icmp <pred> i32 %trunc, %invariant
Replacing it with
%iv = phi i64 ...
%cmp = icmp <pred> i64 %iv, sext/zext(%invariant)
In case if it is legal. Specifically, if `%iv` has signed comparison users, it is
required that `sext(trunc(%iv)) == %iv`, and if it has unsigned comparison
uses then we require `zext(trunc(%iv)) == %iv`. The current implementation
bails if `%trunc` has other uses than `icmp`, but in theory we can handle more
cases here (e.g. if the user of trunc is bitcast).
Differential Revision: https://reviews.llvm.org/D47928
Reviewed By: reames
llvm-svn: 335020
This reverts r334428. It incorrectly marks some multiplications as nuw. Tim
Shen is working on a proper fix.
Original commit message:
[SCEV] Add nuw/nsw to mul ops in StrengthenNoWrapFlags where safe.
Summary:
Previously we would add them for adds, but not multiplies.
llvm-svn: 335016
This reverts commit f976cf4cca0794267f28b54e468007fd476d37d9.
I am reverting this because it causes break in a few bots and its going
to take me sometime to look at this.
llvm-svn: 334993
Summary:
Simplify blockaddress usage before giving up in MergeBlockIntoPredecessor
This is a missing small optimization in MergeBlockIntoPredecessor.
This helps with one simplifycfg test which expects this case to be handled.
Reviewers: davide, spatel, brzycki, asbirlea
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D48284
llvm-svn: 334992
redundant-vf2-cost.ll is X86 specific. Moved from
test/Transforms/LoopVectorize/redundant-vf2-cost.ll to
test/Transforms/LoopVectorize/X86/redundant-vf2-cost.ll
llvm-svn: 334854
Summary:
When iterating users of a multiply in processUMulZExtIdiom, the
call to setOperand in the truncation case may replace the use
being visited; make sure the iterator has been advanced before
doing that replacement.
Reviewers: majnemer, davide
Reviewed By: davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D48192
llvm-svn: 334844
This is a minor fix for LV cost model, where the cost for VF=2 was
computed twice when the vectorization of the loop was forced without
specifying a VF.
Reviewers: xusx595, hsaito, fhahn, mkuper
Reviewed By: hsaito, xusx595
Differential Revision: https://reviews.llvm.org/D48048
llvm-svn: 334840
This is r334704 (which was reverted in r334732) with a fix for
types like x86_fp80. We need to use getTypeAllocSizeInBits and
not getTypeStoreSizeInBits to avoid dropping debug info for
such types.
Original commit msg:
> Summary:
> Do not convert a DbgDeclare to DbgValue if the store
> instruction only refer to a fragment of the variable
> described by the DbgDeclare.
>
> Problem was seen when for example having an alloca for an
> array or struct, and there were stores to individual elements.
> In the past we inserted a DbgValue intrinsics for each store,
> just as if the store wrote the whole variable.
>
> When handling store instructions we insert a DbgValue that
> indicates that the variable is "undefined", as we do not know
> which part of the variable that is updated by the store.
>
> When ConvertDebugDeclareToDebugValue is used with a load/phi
> instruction we assert that the referenced value is large enough
> to cover the whole variable. Afaict this should be true for all
> scenarios where those methods are used on trunk. If the assert
> blows in the future I guess we could simply skip to insert a
> dbg.value instruction.
>
> In the future I think we should examine which part of the variable
> that is accessed, and add a DbgValue instrinsic with an appropriate
> DW_OP_LLVM_fragment expression.
>
> Reviewers: dblaikie, aprantl, rnk
>
> Reviewed By: aprantl
>
> Subscribers: JDevlieghere, llvm-commits
>
> Tags: #debug-info
>
> Differential Revision: https://reviews.llvm.org/D48024
llvm-svn: 334830
Summary:
We already do it for splat constants, but not just values.
Also, undef cases are mostly non-functional.
The original commit was reverted because
it broke tests for amdgpu backend, which i didn't check.
Now, the backed was updated to recognize these new
patterns, so we are good.
https://bugs.llvm.org/show_bug.cgi?id=37603https://rise4fun.com/Alive/cplX
Reviewers: spatel, craig.topper, mareko, bogner, rampitec, nhaehnle, arsenm
Reviewed By: spatel, rampitec, nhaehnle
Subscribers: wdng, nhaehnle, llvm-commits
Differential Revision: https://reviews.llvm.org/D47980
llvm-svn: 334818
IEEE 754 defines the expected result on overflow. As far as I know,
hardware implementations (of f16), and compiler-rt (__floatuntisf)
correctly return +-Inf on overflow. And I can't think of any useful
transform that would take advantage of overflow being undefined here.
Differential Revision: https://reviews.llvm.org/D47807
llvm-svn: 334777
This patches teaches EarlyCSE to figure out that if `and i1 %x, %y` is true then both
`%x` and `%y` are true in the taken branch, and if `or i1 %x, %y` is false then both
`%x` and `%y` are false in non-taken branch. Fix for PR37635.
Differential Revision: https://reviews.llvm.org/D47574
Reviewed By: reames
llvm-svn: 334707
Summary:
Do not convert a DbgDeclare to DbgValue if the store
instruction only refer to a fragment of the variable
described by the DbgDeclare.
Problem was seen when for example having an alloca for an
array or struct, and there were stores to individual elements.
In the past we inserted a DbgValue intrinsics for each store,
just as if the store wrote the whole variable.
When handling store instructions we insert a DbgValue that
indicates that the variable is "undefined", as we do not know
which part of the variable that is updated by the store.
When ConvertDebugDeclareToDebugValue is used with a load/phi
instruction we assert that the referenced value is large enough
to cover the whole variable. Afaict this should be true for all
scenarios where those methods are used on trunk. If the assert
blows in the future I guess we could simply skip to insert a
dbg.value instruction.
In the future I think we should examine which part of the variable
that is accessed, and add a DbgValue instrinsic with an appropriate
DW_OP_LLVM_fragment expression.
Reviewers: dblaikie, aprantl, rnk
Reviewed By: aprantl
Subscribers: JDevlieghere, llvm-commits
Tags: #debug-info
Differential Revision: https://reviews.llvm.org/D48024
llvm-svn: 334704
IndVarSimplify sometimes makes transforms basing on users that are trivially dead. In particular,
if DCE wasn't run before it, there may be a dead `sext/zext` in loop that will trigger widening
transforms, however it makes no sense to do it.
This patch teaches IndVarsSimplify ignore the mist trivial cases of that.
Differential Revision: https://reviews.llvm.org/D47974
Reviewed By: sanjoy
llvm-svn: 334567
Name table occupies a big chunk of size in current binary format sample profile.
In order to reduce its size, the patch changes the sample writer/reader to
save/restore MD5Hash of names in the name table. Sample annotation phase will
also use MD5Hash of name to query samples accordingly.
Experiment shows compact binary format can reduce the size of sample profile by
2/3 compared with binary format generally.
Differential Revision: https://reviews.llvm.org/D47955
llvm-svn: 334447
Summary:
Previously we would add them for adds, but not multiplies.
Reviewers: sanjoy
Subscribers: llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D48038
llvm-svn: 334428
Summary:
`%ret = add nuw i8 %x, C`
From [[ https://llvm.org/docs/LangRef.html#add-instruction | langref ]]:
nuw and nsw stand for “No Unsigned Wrap” and “No Signed Wrap”,
respectively. If the nuw and/or nsw keywords are present,
the result value of the add is a poison value if unsigned
and/or signed overflow, respectively, occurs.
So if `C` is `-1`, `%x` can only be `0`, and the result is always `-1`.
I'm not sure we want to use `KnownBits`/`LVI` here, because there is
exactly one possible value (all bits set, `-1`), so some other pass
should take care of replacing the known-all-ones with constant `-1`.
The `test/Transforms/InstCombine/set-lowbits-mask-canonicalize.ll` change *is* confusing.
What happening is, before this: (omitting `nuw` for simplicity)
1. First, InstCombine D47428/rL334127 folds `shl i32 1, %NBits`) to `shl nuw i32 -1, %NBits`
2. Then, InstSimplify D47883/rL334222 folds `shl nuw i32 -1, %NBits` to `-1`,
3. `-1` is inverted to `0`.
But now:
1. *This* InstSimplify fold `%ret = add nuw i32 %setbit, -1` -> `-1` happens first,
before InstCombine D47428/rL334127 fold could happen.
Thus we now end up with the opposite constant,
and it is all good: https://rise4fun.com/Alive/OA9https://rise4fun.com/Alive/sldC
Was mentioned in D47428 review.
Follow-up for D47883.
Reviewers: spatel, craig.topper
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47908
llvm-svn: 334298
There could be more than one PHIs in exit block using same loop recurrence.
Don't assume there is only one and fix each user.
Differential Revision: https://reviews.llvm.org/D47788
llvm-svn: 334271
%ret = add nuw i8 %x, C
From langref:
nuw and nsw stand for “No Unsigned Wrap” and “No Signed Wrap”,
respectively. If the nuw and/or nsw keywords are present,
the result value of the add is a poison value if unsigned
and/or signed overflow, respectively, occurs.
So if C is -1, %x can only be 0, and the result is always -1.
https://rise4fun.com/Alive/sldC
Was mentioned in D47428 review.
llvm-svn: 334236
Summary:
`%r = shl nuw i8 C, %x`
As per langref:
```
If the nuw keyword is present, then the shift produces
a poison value if it shifts out any non-zero bits.
```
Thus, if the sign bit is set on `C`, then `%x` can only be `0`,
which means that `%r` can only be `C`.
Or in other words, set sign bit means that the signed value
is negative, so the constant is `<= 0`.
https://rise4fun.com/Alive/WMkhttps://rise4fun.com/Alive/udv
Was mentioned in D47428 review.
We already handle the `0` constant, https://godbolt.org/g/UZq1sJ, so this only handles negative constants.
Could use computeKnownBits() / LazyValueInfo,
but the cost-benefit analysis (https://reviews.llvm.org/D47891)
suggests it isn't worth it.
Reviewers: spatel, craig.topper
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47883
llvm-svn: 334222
These weren't included in D19544 - probably just an oversight.
D40044 made it more likely that we'll have LLVM math intrinsics rather
than libcalls, so this bug was more easily exposed.
As the tests/code show, we already have the complete mappings for pow/exp/log.
I don't have any experience with SVML, so I don't know if anything else is
missing. It's also not clear to me that we should be doing this transform in
IR rather than DAG/isel, but that's a separate issue.
Differential Revision: https://reviews.llvm.org/D47610
llvm-svn: 334211
%r = shl nuw i8 C, %x
As per langref: If the nuw keyword is present, then the shift produces
a poison value if it shifts out any non-zero bits.
Thus, if the sign bit is set on C, then %x can only be 0,
which means that %r can only be C.
https://rise4fun.com/Alive/WMk
Was mentioned in D47428 review.
llvm-svn: 334200
We do the same thing in rewriteSingleStoreAlloca.
Fixes PR37632.
Reviewers: chandlerc, davide, efriedma
Reviewed By: davide
Differential Revision: https://reviews.llvm.org/D47825
llvm-svn: 334187
The bug report:
https://bugs.llvm.org/show_bug.cgi?id=36036
...requests a DAG change for this, but an IR canonicalization
probably handles most cases. If we still want to match this
pattern in the backend, there's a proposal for that too:
D47831
Alive proofs including nsw/nuw cases that were first noted in:
D46988
https://rise4fun.com/Alive/Kmp
This patch is largely copied from the existing code that was
initially added with:
D40984
...but I didn't see much gain from trying to share code.
llvm-svn: 334137
Summary:
This is [[ https://bugs.llvm.org/show_bug.cgi?id=37603 | PR37603 ]].
https://godbolt.org/g/VCMNpShttps://rise4fun.com/Alive/idM
When doing bit manipulations, it is quite common to calculate some bit mask,
and apply it to some value via `and`.
The typical C code looks like:
```
int mask_signed_add(int nbits) {
return (1 << nbits) - 1;
}
```
which is translated into (with `-O3`)
```
define dso_local i32 @mask_signed_add(int)(i32) local_unnamed_addr #0 {
%2 = shl i32 1, %0
%3 = add nsw i32 %2, -1
ret i32 %3
}
```
But there is a second, less readable variant:
```
int mask_signed_xor(int nbits) {
return ~(-(1 << nbits));
}
```
which is translated into (with `-O3`)
```
define dso_local i32 @mask_signed_xor(int)(i32) local_unnamed_addr #0 {
%2 = shl i32 -1, %0
%3 = xor i32 %2, -1
ret i32 %3
}
```
Since we created such a mask, it is quite likely that we will use it in `and` next.
And then we may get rid of `not` op by folding into `andn`.
But now that i have actually looked:
https://godbolt.org/g/VTUDmU
_some_ backend changes will be needed too.
We clearly loose `bzhi` recognition.
Reviewers: spatel, craig.topper, RKSimon
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47428
llvm-svn: 334127
Before this patch, debugify would insert debug value intrinsics before the
terminating instruction in a block. This had the advantage of being simple,
but was a bit too simple/unrealistic.
This patch teaches debugify to insert debug values immediately after their
operand defs. This enables better testing of the compiler.
For example, with this patch, `opt -debugify-each` is able to identify a
vectorizer DI-invariance bug fixed in llvm.org/PR32761. In this bug, the
vectorizer produced different output with/without debug info present.
Reverting Davide's bugfix locally, I see:
$ ~/scripts/opt-check-dbg-invar.sh ./bin/opt \
.../SLPVectorizer/AArch64/spillcost-di.ll -slp-vectorizer
Comparing: -slp-vectorizer .../SLPVectorizer/AArch64/spillcost-di.ll
Baseline: /var/folders/j8/t4w0bp8j6x1g6fpghkcb4sjm0000gp/T/tmp.iYYeL1kf
With DI : /var/folders/j8/t4w0bp8j6x1g6fpghkcb4sjm0000gp/T/tmp.sQtQSeet
9,11c9,11
< %5 = getelementptr inbounds %0, %0* %2, i64 %0, i32 1
< %6 = bitcast i64* %4 to <2 x i64>*
< %7 = load <2 x i64>, <2 x i64>* %6, align 8, !tbaa !0
---
> %5 = load i64, i64* %4, align 8, !tbaa !0
> %6 = getelementptr inbounds %0, %0* %2, i64 %0, i32 1
> %7 = load i64, i64* %6, align 8, !tbaa !5
12a13
> store i64 %5, i64* %8, align 8, !tbaa !0
14,15c15
< %10 = bitcast i64* %8 to <2 x i64>*
< store <2 x i64> %7, <2 x i64>* %10, align 8, !tbaa !0
---
> store i64 %7, i64* %9, align 8, !tbaa !5
:: Found a test case ^
Running this over the *.ll files in tree, I found four additional examples
which compile differently with/without DI present. I plan on filing bugs for
these.
llvm-svn: 334118
These were added for D5603 / rL219542, and there's a proposal to
change one side in D47807.
These are tests of constant propagation, so they shouldn't have
ever been tested/housed under InstCombine.
llvm-svn: 334107
If no alignment is set, the abi/preferred alignment of structs will be
used which may be higher than required. This can lead to extra padding
and in the end an increase in data size.
Differential Revision: https://reviews.llvm.org/D47633
llvm-svn: 334099
We should never get different CodeGen based on whether the code is being
compiled in debug mode so we must skip over @llvm.dbg.value (and similar)
calls.
Should fix at least the worst part of PR37690.
llvm-svn: 334090
CodeGenPrepare pass move extension instructions close to load instructions in different BB, so they can be combined later. But the extension instructions can't move through logical and shift instructions in current implementation. This patch enables this enhancement, so we can eliminate more extension instructions.
Differential Revision: https://reviews.llvm.org/D45537
This is re-commit of r331783, which was reverted by r333305. The performance regression was caused by some unlucky alignment, not a code generation problem.
llvm-svn: 334049
When adjusting a cmp in order to canonicalize an abs/nabs select pattern we need
to use the type of the existing operand when creating a new operand not the
type of a select operand, as the two may be different.
This fixes PR37686.
llvm-svn: 334019
As noted in rL333782, we can be both better for optimization and
safer with this transform:
BinOp (shuffle V1, Mask), C --> shuffle (BinOp V1, NewC), Mask
The only potentially unsafe-to-speculate binops are integer div/rem.
All other binops are always safe (although I don't see a way to
assert that in code here).
For opcodes like shifts that can produce poison, it can't matter
here because we know the lanes with undef are dropped by the
subsequent shuffle.
Differential Revision: https://reviews.llvm.org/D47686
llvm-svn: 333962
entries to reach the target. Since these calls don't require type checks,
we can short-circuit them to their real targets, except in cases when they
can be pre-empted.
Differential Revision: https://reviews.llvm.org/D46326
llvm-svn: 333937
When checking a select to see if it matches an abs, allow the true/false values
to be a sign-extension of the comparison value instead of requiring that they're
directly the comparison value, as all the comparison cares about is the sign of
the value.
This fixes a regression due to r333702, where we were no longer generating ctlz
due to isKnownNonNegative failing to match such a pattern.
Differential Revision: https://reviews.llvm.org/D47631
llvm-svn: 333927
After r333856, opt -debugify would just stop emitting debug value
intrinsics after encountering a musttail call. This wasn't sufficient to
avoid verifier failures.
Debug value intrinicss for all instructions preceding a musttail call
must also be emitted before the musttail call.
llvm-svn: 333866
When we optimize select basing on fact that div by 0 is undef
we should not traverse the instruction which are not guaranteed to
transfer execution to next instruction. Guard intrinsic is an example.
Reviewers: spatel, craig.topper
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47576
llvm-svn: 333864
There's a patchwork of existing transforms trying to handle
these cases, but as seen in the changed test, we weren't
catching them all.
llvm-svn: 333845
As noted in the review thread for rL333782, we could have
made a bug harder to hit if we were simplifying instructions
before trying other folds.
The shuffle transform in question isn't ever a simplification;
it's just a canonicalization. So I've renamed that to make that
clearer.
This is NFCI at this point, but I've regenerated the test file
to show the cosmetic value naming difference of using
instcombine's RAUW vs. the builder.
Possible follow-ups:
1. Move reassociation folds after simplifies too.
2. Refactor common code; we shouldn't have so much repetition.
llvm-svn: 333820
As noted in the review thread for rL333782, we're lacking coverage
for this transform, so add tests for each binop opcode with constant
operand.
llvm-svn: 333818
Summary:
I noticed this issue because we didn't put the primary cloned loop into
the `NonChildClonedLoops` vector and so never iterated on it. Once
I fixed that, it made it clear why I had to do a really complicated and
unnecesasry dance when updating the loops to remain in canonical form --
I was unwittingly working around the fact that the primary cloned loop
wasn't in the expected list of cloned loops. Doh!
Now that we include it in this vector, we don't need to return it and we
can consolidate the update logic as we correctly have a single place
where it can be handled.
I've just added a test for the iteration order aspect as every time
I changed the update logic partially or incorrectly here, an existing
test failed and caught it so that seems well covered (which is also
evidenced by the extensive working around of this missing update).
Reviewers: asbirlea, sanjoy
Subscribers: mcrosier, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D47647
llvm-svn: 333811
Summary:
Getelementptr returns a vector of pointers, instead of a single address,
when one or more of its arguments is a vector. In such case it is not
possible to simplify the expression by inserting a bitcast of operand(0)
into the destination type, as it will create a bitcast between different
sizes.
Reviewers: majnemer, mkuper, mssimpso, spatel
Reviewed By: spatel
Subscribers: lebedev.ri, llvm-commits
Differential Revision: https://reviews.llvm.org/D46379
llvm-svn: 333783
This bug:
https://bugs.llvm.org/show_bug.cgi?id=37648
...was created with the enhancement to this transform with rL332479.
The urem test shows the disaster potential: any undef divisor lane makes
the whole op undef.
The test diffs show that vector demanded elements turns some of the potential,
but not all, unused binop operands back into undef already.
llvm-svn: 333782
Summary:
Emit summaries for bitcode modules that are only destined for the
regular LTO portion of the build so they can participate in
summary-based dead stripping.
This change reduces the size of a nacl_helper build with cfi-icall
enabled by 7%, removing the majority of the overhead due to enabling
cfi-icall. The cfi-icall size increase was caused by compiling in lots
of unused code and cfi-icall generating jumptable references to unused
symbols that could no longer be removed by -Wl,-gc-sections. Increasing
the visibility of summary-based dead stripping prevented jumptable
entries being created for unused symbols from the regular LTO portion
of the build.
Reviewers: pcc
Reviewed By: pcc
Subscribers: dschuff, mehdi_amini, inglorion, eraman, llvm-commits, kcc
Differential Revision: https://reviews.llvm.org/D47594
llvm-svn: 333768
This patch updates IPSCCP to use PredicateInfo to propagate
facts to true branches predicated by EQ and to false branches
predicated by NE.
As a follow up, we should be able to extend it to also propagate additional
facts about nonnull.
Reviewers: davide, mssimpso, dberlin, efriedma
Reviewed By: davide, dberlin
Differential Revision: https://reviews.llvm.org/D45330
llvm-svn: 333740
Summary:
Loop idiom recognize tries to convert loops like
```
int foo(int x) {
int cnt = 0;
while (x) {
x >>= 1;
++cnt;
}
return cnt;
}
```
into calls to ctlz, but if x is initially negative this loop should be infinite.
It happens that the cases that motivated this change have an absolute value of x before the loop. So this patch restricts the transform to cases where we know x is positive. Note: We are relying on the absolute value of INT_MIN to be undefined so we can assume that the result is always positive.
Fixes PR37479
Reviewers: spatel, hfinkel, efriedma, javed.absar
Reviewed By: efriedma
Subscribers: dmgreen, llvm-commits
Differential Revision: https://reviews.llvm.org/D47348
llvm-svn: 333702
This is the planned enhancement to D47163 / rL333611.
We want to match cmp/select sizes because that will be recognized
as min/max more easily and lead to better codegen (especially for
vector types).
As mentioned in D47163, this improves some of the tests that would
also be folded by D46380, so we may want to adjust that patch to
match the new patterns where the extend op occurs after the select.
llvm-svn: 333689
Convert a vector load intrinsic into an llvm load instruction.
This is beneficial when the underlying object being addressed
comes from a constant, since we get constant-folding for free.
Differential Revision: https://reviews.llvm.org/D46273
llvm-svn: 333643
In post-commit review, Eric Christopher notes that many
new MSan warnings are being observed with this patch.
The probable reason is: if 'y' is undef here and we could
evaluate it twice and get different results.
We can't increase the number of uses of a value.
llvm-svn: 333631
Don't always:
cast (select (cmp x, y), z, C) --> select (cmp x, y), (cast z), C'
This is something that came up as far back as D26556, and I lost track of it.
I suspect that this transform is part of the underlying problem that is
inspiring some of the recent proposals that seek to match larger patterns
that include a cast op. Even if that's not true, this transform causes
problems for codegen (particularly with vector types).
A transform to actively match the size of cmp and select operand sizes should
follow. This patch just removes the harmful canonicalization in the other
direction.
Differential Revision: https://reviews.llvm.org/D47163
llvm-svn: 333611
Summary:
Fix PR37625. It's possible for an extern_weak declaration to be emitted
to the merged module when a definition exists in the ThinLTO portion of
the build; discard the linkage on the declaration in that case.
(otherwise we copy the linkage to the alias to the jumptable and fail)
Reviewers: pcc
Reviewed By: pcc
Subscribers: mehdi_amini, llvm-commits, kcc
Differential Revision: https://reviews.llvm.org/D47494
llvm-svn: 333604
Summary:
The isKnownNonZero() function have checks that abort the recursion when
it reaches the specified max depth. However one of the recursive calls
was placed before the max depth check was done, resulting in a endless
recursion that eventually triggered a segmentation fault.
Fixed the problem by moving the max depth check above the first
recursive call.
Reviewers: Prazek, nlopes, spatel, craig.topper, hfinkel
Reviewed By: hfinkel
Subscribers: hfinkel, bjope, llvm-commits
Differential Revision: https://reviews.llvm.org/D47531
llvm-svn: 333557
Turning a table lookup intrinsic into a shuffle vector instruction
can be beneficial. If the mask used for the lookup is the constant
vector {7,6,5,4,3,2,1,0}, then the back-end generates byte reverse
instructions instead.
Differential Revision: https://reviews.llvm.org/D46133
llvm-svn: 333550
loop-cleanup passes at the beginning of the loop pass pipeline, and
re-enqueue loops after even trivial unswitching.
This will allow us to much more consistently avoid simplifying code
while doing trivial unswitching. I've also added a test case that
specifically shows effective iteration using this technique.
I've unconditionally updated the new PM as that is always using the
SimpleLoopUnswitch pass, and I've made the pipeline changes for the old
PM conditional on using this new unswitch pass. I added a bunch of
comments to the loop pass pipeline in the old PM to make it more clear
what is going on when reviewing.
Hopefully this will unblock doing *partial* unswitching instead of just
full unswitching.
Differential Revision: https://reviews.llvm.org/D47408
llvm-svn: 333493
be both simpler and substantially more efficient.
Rather than use a hand-rolled iteration technique that isn't quite the
same as RPO, use the pre-built RPO loop body traversal utility.
Once visiting the loop body in RPO, we can assert that we visit defs
before uses reliably. When this is the case, the only need to iterate is
when simplifying a def that is used by a PHI node along a back-edge.
With this patch, the first pass over the loop body is just a complete
simplification of every instruction across the loop body. When we
encounter a use of a simplified instruction that stems from a PHI node
in the loop body that has already been visited (due to some cyclic CFG,
potentially the loop itself, or a nested loop, or unstructured control
flow), we recall that specific PHI node for the second iteration.
Nothing else needs to be preserved from iteration to iteration.
On the second and later iterations, only instructions known to have
simplified inputs are considered, each time starting from a set of PHIs
that had simplified inputs along the backedges.
Dead instructions are collected along the way, but deleted in a batch at
the end of each iteration making the iterations themselves substantially
simpler. This uses a new batch API for recursively deleting dead
instructions.
This alsa changes the routine to visit subloops. Because simplification
is fundamentally transitive, we may need to visit the entire loop body,
including subloops, to handle knock-on simplification.
I've added a basic test file that helps demonstrate that all of these
changes work. It includes both straight-forward loops with
simplifications as well as interesting PHI-structures, CFG-structures,
and a nested loop case.
Differential Revision: https://reviews.llvm.org/D47407
llvm-svn: 333461
This is a simple implementation of the unroll-and-jam classical loop
optimisation.
The basic idea is that we take an outer loop of the form:
for i..
ForeBlocks(i)
for j..
SubLoopBlocks(i, j)
AftBlocks(i)
Instead of doing normal inner or outer unrolling, we unroll as follows:
for i... i+=2
ForeBlocks(i)
ForeBlocks(i+1)
for j..
SubLoopBlocks(i, j)
SubLoopBlocks(i+1, j)
AftBlocks(i)
AftBlocks(i+1)
Remainder
So we have unrolled the outer loop, then jammed the two inner loops into
one. This can lead to a simpler inner loop if memory accesses can be shared
between the now-jammed loops.
To do this we have to prove that this is all safe, both for the memory
accesses (using dependence analysis) and that ForeBlocks(i+1) can move before
AftBlocks(i) and SubLoopBlocks(i, j).
Differential Revision: https://reviews.llvm.org/D41953
llvm-svn: 333358
Reverting this to see if this is causing the failures of the
clang-with-thin-lto-ubuntu bot.
[IPSCCP] Use PredicateInfo to propagate facts from cmp instructions.
This patch updates IPSCCP to use PredicateInfo to propagate
facts to true branches predicated by EQ and to false branches
predicated by NE.
As a follow up, we should be able to extend it to also propagate additional
facts about nonnull.
Reviewers: davide, mssimpso, dberlin, efriedma
Reviewed By: davide, dberlin
Differential Revision: https://reviews.llvm.org/D45330
llvm-svn: 333323
Libfuzzer tests have been fixed to prevent being optimized.
Original commit message:
If the nsw flag is used in the absolute value then it is undefined for INT_MIN. For all other value it will produce a positive number. So we can assume the result is positive.
This breaks some InstCombine abs/nabs combining tests because we simplify the second compare from known bits rather than as the whole pattern. Looks like we can probably fix it by adding a neg+abs/nabs combine to just swap the select operands. N
Differential Revision: https://reviews.llvm.org/D47041
llvm-svn: 333300
Summary:
Look past debug intrinsics when querying whether an instruction is the
first instruction in the header block. The commit includes a reproducer
for a case where LICM would not hoist an instruction, due to the presence
of the intrinsic.
A caveat with this commit is that the check will not work properly if
the instruction at hand is a debug intrinsic. I assume that no one
depends on isGuaranteedToExecute() to return true for debug intrinsics
for these cases (and that this might be an indication of another debug
invariant issue), so I thought that it was not worth adding that extra
bit of complexity.
Reviewers: reames, anna
Reviewed By: anna
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47197
llvm-svn: 333274
This patch updates IPSCCP to use PredicateInfo to propagate
facts to true branches predicated by EQ and to false branches
predicated by NE.
As a follow up, we should be able to extend it to also propagate additional
facts about nonnull.
Reviewers: davide, mssimpso, dberlin, efriedma
Reviewed By: davide, dberlin
Differential Revision: https://reviews.llvm.org/D45330
llvm-svn: 333268
Setting the "Debug Info Version" module flag makes it possible to pipe
synthetic debug info into llc, which is useful for testing backends.
llvm-svn: 333237
If the nsw flag is used in the absolute value then it is undefined for INT_MIN. For all other value it will produce a positive number. So we can assume the result is positive.
This breaks some InstCombine abs/nabs combining tests because we simplify the second compare from known bits rather than as the whole pattern. Looks like we can probably fix it by adding a neg+abs/nabs combine to just swap the select operands. Need to check alive to make sure there are no corner cases.
Differential Revision: https://reviews.llvm.org/D47041
llvm-svn: 333226
Reassociation of math ops in some contexts (especially vector contexts)
has generally only been happening when the 'fast' FMF was set. This
enables reassoication when only the finer grained controls 'reassoc' and
'nsz' are set.
Differential Revision: https://reviews.llvm.org/D47335
llvm-svn: 333221
Summary:
In LICM, CFG could be changed in splitPredecessorsOfLoopExit(), which update
only DT and LoopInfo. Therefore, we should preserve only DT and LoopInfo specifically,
instead of all analyses that depend on the CFG (setPreservesCFG()).
This change should fix PR37323.
Reviewers: uabelho, davide, dberlin, Ka-Ka
Reviewed By: dberlin
Subscribers: mzolotukhin, bjope, mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D46775
llvm-svn: 333198
The ARM/ARM64 AESE and AESD instructions have a builtin XOR as the first step in
the instruction. Therefore, if the AES key is zero and the AES data was
previously XORed, it can be combined into a single instruction.
Differential Revision: https://reviews.llvm.org/D47239
Patch by Michael Brase!
llvm-svn: 333193
Summary:
If NaryReassociate succeed it will, when replacing the old instruction
with the new instruction, also recursively delete trivially
dead instructions from the old instruction. However, if the input to the
NaryReassociate pass contain dead code it is not save to recursively
delete trivially deadinstructions as it might lead to deleting the newly
created instruction.
This patch will fix the problem by using WeakVH to detect this
rare case, when the newly created instruction is dead, and it will then
restart the basic block iteration from the beginning.
This fixes pr37539
Reviewers: tra, meheff, grosser, sanjoy
Reviewed By: sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47139
llvm-svn: 333155
Summary:
StructurizeCFG::orderNodes basically uses a reverse post-order (RPO) traversal of the region list to get the order.
The only problem with it is that sometimes backedges for outer loops will be visited before backedges for inner loops.
To solve this problem, a loop depth based approach has been used to make sure all blocks in this loop has been visited
before moving on to outer loop.
However, we found a problem for a SubRegion which is a loop itself:
--> BB1 --> BB2 --> BB3 -->
In this case, BB2 is a SubRegion (loop), and thus its loopdepth is different than that of BB1 and BB3. This fact will lead
BB2 to be placed in the wrong order.
In this work, we treat the SubRegion as a special case and use its exit block to determine the loop and its depth
to guard the sorting.
Reviewers:
arsenm, jlebar
Differential Revision:
https://reviews.llvm.org/D46912
llvm-svn: 333111
Summary:
Finally fixes [[ https://bugs.llvm.org/show_bug.cgi?id=6773 | PR6773 ]].
Now that the backend is all done, we can finally fold it!
The canonical unfolded masked merge pattern is
```(x & m) | (y & ~m)```
There is a second, equivalent variant:
```(x | ~m) & (y | m)```
Only one of them (the or-of-and's i think) is canonical.
And if the mask is not a constant, we should fold it to:
```((x ^ y) & M) ^ y```
https://rise4fun.com/Alive/ndQw
Reviewers: spatel, craig.topper
Reviewed By: spatel
Subscribers: nicholas, RKSimon, llvm-commits
Differential Revision: https://reviews.llvm.org/D46814
llvm-svn: 333106
Loop unswitching makes substantial changes to a loop that can also affect cached
SCEV info in its outer loops as well, but it only cares to invalidate SCEV cache for the
innermost loop in case of full unswitching and does not invalidate anything at all in
case of trivial unswitching. As result, we may end up with incorrect data in cache.
Differential Revision: https://reviews.llvm.org/D46045
Reviewed By: mzolotukhin
llvm-svn: 333072
Summary:
Patch for capture tracking broke
bootstrap of clang with -fstict-vtable-pointers
which resulted in debbugging nightmare. It was fixed
https://reviews.llvm.org/D46900 but as it turned
out, there were other parts like inliner (computing of
noalias metadata) that I found after bootstraping with enabled
assertions.
Reviewers: hfinkel, rsmith, chandlerc, amharc, kuhar
Subscribers: JDevlieghere, eraman, llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D47088
llvm-svn: 333070
Also, produce the canonical IR abs (s<0) to be more efficient.
This is the libcall equivalent of the clang builtin change from:
rL333038
Pasting from that commit message:
The stdlib functions are defined in section 7.20.6.1 of the C standard with:
"If the result cannot be represented, the behavior is undefined."
That lets us mark the negation with 'nsw' because "sub i32 0, INT_MIN" would
be UB/poison.
llvm-svn: 333042
Summary: Previous patch does not care if a value is changed between calloc and strlen. This needs to be removed from InstCombine and maybe moved to DSE later after some rework.
Reviewers: efriedma
Reviewed By: efriedma
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47218
llvm-svn: 333022
This patch fixes two bugs:
* test1: Previously assume(a >= 5) concluded that a == 5. That's only
valid for assume(a == 5)...
* test2: If operands were swapped, additional users were added to the
wrong cmp operand. This resulted in an "unsettled iteration"
assertion failure.
Patch by Nikita Popov
Differential Revision: https://reviews.llvm.org/D46974
llvm-svn: 333007
Summary:
When lowerswitch merge several cases into a new default block it's not
updating the PHI nodes accordingly. The code that update the PHI nodes
for the default edge only update the first entry and do not remove the
remaining ones, to make sure the number of entries match the number of
predecessors.
This is easily fixed by replacing the code that update the PHI node with
the already existing utility function for updating PHI nodes.
Reviewers: hans, reames, arsenm
Reviewed By: arsenm
Subscribers: wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D47055
llvm-svn: 332960
Summary:
In LoopVersioning::addPHINodes we need to iterate over all
users for a value "Inst", and if the user is outside of the
VersionedLoop we should replace the use of "Inst" by using
the value "PN" instead.
Replacing the use of "Inst" for a user of "Inst" also means
that Inst->users() is modified. So it is not safe to do the
replace while iterating over Inst->users() as we used to do.
This patch splits the task into two steps. First we iterate
over Inst->users() to find all users that should be updated.
Those users are saved into a local data structure on the stack.
And then, in the second step, we do the actual updates. This
time iterating over the local data structure.
Reviewers: mzolotukhin, anemet
Reviewed By: mzolotukhin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47134
llvm-svn: 332958
We can eliminate old value if bound_ctrl = 1 and row_mask = bank_mask = 0xf.
This is alternative implementation working with the intrinsic in InstCombine.
Original review for past-ISel optimization: D46570.
Differential Revision: https://reviews.llvm.org/D46596
llvm-svn: 332956
This usually results in better code. Fixes using
inline asm with short2, and also fixes having a different
ABI for function parameters between VI and gfx9.
Partially cleans up the mess used for lowering of the d16
operations. Making v4f16 legal will help clean this up more,
but this requires additional work.
llvm-svn: 332953
In all cases, we're pulling the cast above the select.
That's not a good canonicalization if we're creating
a select that then mismatches the operand size of its
condition.
llvm-svn: 332883
Change matchSelectPattern to return X and -X for ABS/NABS in a well defined order. Adjust EarlyCSE to account for this. Ensure the SPF result is some kind of min/max and not abs/nabs in one place in InstCombine that made me nervous.
Prevously we returned the two operands of the compare part of the abs pattern. The RHS is always going to be a 0i, 1 or -1 constant. This isn't a very meaningful thing to return for any one. There's also some freedom in the abs pattern as to what happens when the value is equal to 0. This freedom led to early cse failing to match when different constants were used in otherwise equivalent operations. By returning the input and its negation in a defined order we can ensure an exact match. This also makes sure both patterns use the exact same subtract instruction for the negation. I believe CSE should evebntually make this happen and properly merge the nsw/nuw flags. But I'm not familiar with CSE and what order it does things in so it seemed like it might be good to really enforce that they were the same.
Differential Revision: https://reviews.llvm.org/D47037
llvm-svn: 332865
r332654 was reverted due to an unused function warning in
release build. This commit includes the same code with the
warning silenced.
Differential Revision: https://reviews.llvm.org/D44338
llvm-svn: 332860
Summary:
This patch fixes PR37526 by simplifying the newly generated LoadInst
instructions. If the pointer address is a bitcast from the pointer to
the NewType, we can just remove this extra bitcast instead of creating
the new one. This fixes the PR37526 + may speed up the whole compilation
process.
Reviewers: spatel, RKSimon, hfinkel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D47144
llvm-svn: 332855
We were previously using a DT in CVP through SimplifyQuery, but not requiring it in
the new pass manager. Hence it would crash if DT was not already available. This now
gets DT directly and plumbs it through to where it is used (instead of using it
through SQ).
llvm-svn: 332836
We already do this for min/max (see the blob above the diff),
so we should do the same for abs/nabs.
A sign-bit check (<s 0) is used as a predicate for other IR
transforms and it's likely the best for codegen.
This might solve the motivating cases for D47037 and D47041,
but I think those patches still make sense. We can't guarantee
this canonicalization if the icmp has more than one use.
Differential Revision: https://reviews.llvm.org/D47076
llvm-svn: 332819
In the patch rL329547, we have lifted the over-restrictive limitation on collected range
checks, allowing to work with range checks with the end of their range not being
provably non-negative. However it appeared that the non-negativity of this value was
assumed in the utility function `ClampedSubtract`. In particular, its reasoning is based
on the fact that `0 <= SINT_MAX - X`, which is not true if `X` is negative.
The function `ClampedSubtract` is only called twice, once with `X = 0` (which is OK)
and the second time with `X = IRC.getEnd()`, where we may now see the problem if
the end is actually a negative value. In this case, we may sometimes miscompile.
This patch is the conservative fix of the miscompile problem. Rather than rejecting
non-provably non-negative `getEnd()` values, we will check it for non-negativity in
runtime. For this, we use function `smax(smin(X, 0), -1) + 1` that is equal to `1` if `X`
is non-negative and is equal to 0 if `X` is negative. If we multiply `Begin, End` of safe
iteration space by this function calculated for `X = IRC.getEnd()`, we will get the original
`[Begin, End)` if `IRC.getEnd()` was non-negative (and, thus, `ClampedSubtract` worked
correctly) and the empty range `[0, 0)` in case if ` IRC.getEnd()` was negative.
So we in fact prohibit execution of the main loop if at least one of range checks was
made against a negative value (and we figured it out in runtime). It is still better than
what we have before (non-negativity had to be proved in compile time) and prevents
us from miscompile, however it is sometiles too restrictive for unsigned range checks
against a negative value (which in fact can be eliminated).
Once we re-implement `ClampedSubtract` in a way that it handles negative `X` correctly,
this limitation can be lifted, too.
Differential Revision: https://reviews.llvm.org/D46860
Reviewed By: samparker
llvm-svn: 332809
The evaluator goes through BB and creates global vars as temporary values to evaluate
results of LLVM instructions. It creates undef for alloca, however it assumes alloca
in addr space 0. If the next instruction is addrspace cast to 0, then we get an invalid
cast instruction.
This patch let the temp global var have an address space matching alloca addr space,
so that the valuation can be done.
Differential Revision: https://reviews.llvm.org/D47081
llvm-svn: 332794
Summary:
This feature is not needed, but it might be usefull in the future
to use metadata to mark what which function should support it
(and strip it when not).
Reviewers: rsmith, sanjoy, amharc, kuhar
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D45419
llvm-svn: 332787
This patch aims to match the changes introduced in gcc by
https://gcc.gnu.org/ml/gcc-cvs/2018-04/msg00534.html. The
IBT feature definition is removed, with the IBT instructions
being freely available on all X86 targets. The shadow stack
instructions are also being made freely available, and the
use of all these CET instructions is controlled by the module
flags derived from the -fcf-protection clang option. The hasSHSTK
option remains since clang uses it to determine availability of
shadow stack instruction intrinsics, but it is no longer directly used.
Comes with a clang patch (D46881).
Patch by mike.dvoretsky
Differential Revision: https://reviews.llvm.org/D46882
llvm-svn: 332705
Summary:
Fix a case where FoldBranchToCommonDest() would bail out from doing CSE
when encountering a debug intrinsic. Handle that by skipping past the
debug intrinsics.
Also, as a minor refactoring, rename checkCSEInPredecessor() to
tryCSEWithPredecessor() to make it a bit more clear that the function
may remove instructions.
Reviewers: fhahn, craig.topper, dblaikie, xbolva00
Reviewed By: fhahn, xbolva00
Subscribers: vsk, davide, llvm-commits
Differential Revision: https://reviews.llvm.org/D46635
llvm-svn: 332698
CanProveNotTakenFirstIteration utility does not handle the case when
condition of the branch is a constant. Add its handling.
Reviewers: reames, anna, mkazantsev
Reviewed By: reames
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D46996
llvm-svn: 332695
Patch #3 from VPlan Outer Loop Vectorization Patch Series #1
(RFC: http://lists.llvm.org/pipermail/llvm-dev/2017-December/119523.html).
Expected to be NFC for the current inner loop vectorization path. It
introduces the basic algorithm to build the VPlan plain CFG (single-level
CFG, no hierarchical CFG (H-CFG), yet) in the VPlan-native vectorization
path using VPInstructions. It includes:
- VPlanHCFGBuilder: Main class to build the VPlan H-CFG (plain CFG without nested regions, for now).
- VPlanVerifier: Main class with utilities to check the consistency of a H-CFG.
- VPlanBlockUtils: Main class with utilities to manipulate VPBlockBases in VPlan.
Reviewers: rengolin, fhahn, mkuper, mssimpso, a.elovikov, hfinkel, aprantl.
Differential Revision: https://reviews.llvm.org/D44338
llvm-svn: 332654
Currently debugify prints it's output to stdout,
with this patch all the output generated goes to stderr.
This change lets us use debugify without taking away
the ability to pipe the output to other llvm tools.
llvm-svn: 332642
According to alive this is valid. I'm hoping to use this to make an assumption that the sign bit is zero after this sequence. The only way it wouldn't be is if the input was INT__MIN, but by preserving the flags we can make doing this to INT_MIN UB.
The nuw flags is weird because it creates such a contradiction that the original number would have to be positive meaning we could remove the select entirely, but we don't get that far.
Differential Revision: https://reviews.llvm.org/D46988
llvm-svn: 332623
entries to reach the target. Since these calls don't require type checks,
we can short-circuit them to their real targets.
Differential Revision: https://reviews.llvm.org/D46326
llvm-svn: 332610
Summary:
Require DominatorTree when requiring/preserving LoopInfo in the old pass manager
BreakCriticalEdges tries to keep LoopInfo and DominatorTree updated if they
exist. However, since commit r321653 and r321805, to update LoopInfo we
must have a DominatorTree, or we will hit an assert.
To fix this we now make a couple of passes that only required/preserved
LoopInfo also require DominatorTree.
This solves PR37334.
Reviewers: eli.friedman, efriedma
Reviewed By: efriedma
Subscribers: efriedma, llvm-commits
Differential Revision: https://reviews.llvm.org/D46829
llvm-svn: 332583
The existing comment said that the functions were available only
on GNU/Linux (and on certain Android versions), but only checked
T.isGNUEnvironment() which also is true on MinGW (for arch-windows-gnu
triplets), which doesn't have such functions.
Existing checks in the initialize function in TargetLibraryInfo.cpp
also use only T.isOSLinux() to check for glibc features.
This fixes use of stdio on MinGW.
Differential Revision: https://reviews.llvm.org/D47002
llvm-svn: 332581
Summary:
The verifier accepts PHI nodes with multiple entries for the
same basic block, as long as the value is the same.
As seen in PR37203, SROA did not handle such PHI nodes properly
when speculating loads over the PHI, since it inserted multiple
loads in the predecessor block and changed the PHI into having
multiple entries for the same basic block, but with different
values.
This patch teaches SROA to reuse the same speculated load for
each PHI duplicate entry in such situations.
Resolves: https://bugs.llvm.org/show_bug.cgi?id=37203
Reviewers: uabelho, chandlerc, hfinkel, bkramer, efriedma
Reviewed By: efriedma
Subscribers: dberlin, efriedma, llvm-commits
Differential Revision: https://reviews.llvm.org/D46426
llvm-svn: 332577
The current integer widening does not support rewriting partial split slices in rewriteIntegerStore (and rewriteIntegerLoad).
This patch adds explicit checks for this case in isIntegerWideningViableForSlice.
Before r322533, splitting is allowed only for the whole-alloca slice and hence the above case is implicitly rejected by another check `if (DL.getTypeStoreSize(ValueTy) > Size)` because whole-alloca slice is larger than the partition.
Differential Revision: https://reviews.llvm.org/D46750
llvm-svn: 332575
The canonicalization was restricted to shuffle masks with
a 1-to-1 mapping to the constant vector, but that disqualifies
the common splat pattern. This is part of solving PR37463:
https://bugs.llvm.org/show_bug.cgi?id=37463
llvm-svn: 332479
Summary: If file stream arg is not captured and source is fopen, we could replace IO calls by unlocked IO ("_unlocked" function variants) to gain better speed,
Reviewers: efriedma, RKSimon, spatel, sanjoy, hfinkel, majnemer, lebedev.ri, rja
Reviewed By: rja
Subscribers: rja, srhines, efriedma, lebedev.ri, llvm-commits
Differential Revision: https://reviews.llvm.org/D45736
llvm-svn: 332452
A catchswitch must be the only non-phi instruction in its basic block;
attempting to move a retain or release into a catchswitch basic block
will result in invalid IR. Explicitly mark a CFG hazard in this case to
prevent the code motion.
Differential Revision: https://reviews.llvm.org/D46482
llvm-svn: 332430
Author: Samuel Pitoiset
Without this patch, it appears to me that we are selecting
the wrong operand when inverting conditions. In the attached
test, it will select %tmp3 instead of %tmp4. To fix it, just
use 'A' as everywhere.
This fixes a regression introduced by
"[PatternMatch] define m_Not using m_Xor and cst_pred_ty"
https://reviews.llvm.org/D46351
llvm-svn: 332403
When two interposable functions are merged, we cannot replace
uses and have to emit calls to a common internal function. However,
writeThunk() will not actually emit a thunk if the function is too
small. This leaves us in a broken state where mergeTwoFunctions
already rewired the functions, but writeThunk doesn't do anything.
This patch changes the implementation so that:
* writeThunk() does just that.
* The direct replacement of calls is moved into mergeTwoFunctions()
into the non-interposable case only.
* isThunkProfitable() is extracted and will be called for
the non-iterposable case always, and in the interposable case
only if uses are still left after replacement.
This issue has been introduced in https://reviews.llvm.org/D34806,
where the code for checking thunk profitability has been moved.
Differential Revision: https://reviews.llvm.org/D46804
Reviewed By: whitequark
llvm-svn: 332342
This adds a -debugify-each mode to opt which, when enabled, wraps each
{Module,Function}Pass in a pipeline with logic to add, check, and strip
synthetic debug info for testing purposes.
This mode can be used to test complex pipelines for debug info bugs, or
to collect statistics about the number of debug values & locations lost
throughout various stages of a pipeline.
Patch by Son Tuan Vu!
Differential Revision: https://reviews.llvm.org/D46525
llvm-svn: 332312
Summary:
Part of the InstCombine code for simplifying GEPs looks through
addrspacecasts. However, this was done by updating a variable
also used by the next transformation, for marking GEPs as
inbounds. This led to replacing a GEP with a similar instruction
in a different addrspace, which caused an assertion failure in RAUW.
This caused julia issue https://github.com/JuliaLang/julia/issues/27055
Patch by Jeff Bezanson <jeff@juliacomputing.com>
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D46722
llvm-svn: 332302
This is a CodeExtractor improvement which adds support for extracting blocks
which have exception handling constructs if that is legal to do. CodeExtractor
performs validation checks to ensure that extraction is legal when it finds
invoke instructions or EH pads (landingpad, catchswitch, or cleanuppad) in
blocks to be extracted.
I have also added an option to allow extraction of blocks with alloca
instructions, but no validation is done for allocas. CodeExtractor caller has
to validate it himself before allowing alloca instructions to be extracted.
By default allocas are still not allowed in extraction blocks.
Differential Revision: https://reviews.llvm.org/D45904
llvm-svn: 332151
Let separate-const-offset-from-gep pass handle trunc() when it calculates
constant offset relative to base. The pass itself may insert trunc()
instructions when it canonicalises array indices to pointer-size integers
and needs to handle trunc() in order to evaluate the offset.
Differential Revision: https://reviews.llvm.org/D46732
llvm-svn: 332142
Summary:
This change adds handling of the atomic memset intrinsic to the
code path that simplifies the regular memset. In practice this means
that we will now also expand a small constant-length atomic memset
into a single unordered atomic store.
Reviewers: apilipenko, skatkov, mkazantsev, anna, reames
Reviewed By: reames
Subscribers: reames, llvm-commits
Differential Revision: https://reviews.llvm.org/D46660
llvm-svn: 332132
If the sprintf function is static (as on mingw-w64, where many stdio
functions are static inline wrappers), earlier optimization passes
could optimize out the return value altogether, and make it void,
which could break optimizations of this libcall that touch the
return value.
This fixes the issue discussed in PR37408 for the sprintf function.
Differential Revision: https://reviews.llvm.org/D46752
llvm-svn: 332106
Phi nodes can reside in live blocks but one of their incoming
arguments can come from a dead block. Dead blocks and reassociate
don't play nice together. In fact, reassociate performs an RPO
as a first step to avoid processing dead blocks.
The reason why Reassociate might not fixpoint when examining
dead blocks is that the following:
%xor0 = xor i16 %xor1, undef
%xor1 = xor i16 %xor0, undef
is perfectly valid LLVM IR (if it appears in a dead block),
so the worklist algorithm keeps pushing the two instructions for
reexamination. Note that this is not Reassociate fault, at least
not entirely. It's llvm that has a weird definition of dominance.
Fixes PR37390.
llvm-svn: 332100
Summary:
This change reworks the handling of atomic memcpy within the instcombine pass.
Previously, a constant length atomic memcpy would be lowered into loads & stores
as long as no more than 16 load/store pairs are created. This is quite different
from the lowering done for a non-atomic memcpy; which only ever lowers into a single
load/store pair of no more than 8 bytes. Larger constant-sized memcpy calls are
expanded to load/stores in later passes, such as SelectionDAG lowering.
In this change the behaviour for atomic memcpy is unified with non-atomic memcpy;
atomic memcpy is now treated in the same was as non-atomic memcpy has always been.
We leave it to later passes to lower longer-length atomic memcpy calls.
Due to the structure of the pass's handling of memtransfer intrinsics, this change
also gives us handling of atomic memmove that we did not previously have.
Reviewers: apilipenko, skatkov, mkazantsev, anna, reames
Reviewed By: reames
Subscribers: reames, llvm-commits
Differential Revision: https://reviews.llvm.org/D46658
llvm-svn: 332093
Summary:
https://bugs.llvm.org/show_bug.cgi?id=34897 demonstrates an incorrect
coroutine frame allocation elision in the coro-elide pass. The elision
is performed on the basis that the SSA variables from all llvm.coro.begin
are directly referenced in subsequent llvm.coro.destroy instructions.
However, this ignores the fact that the function may exit through paths
that do not run these destroy instructions. In the sample program from
PR34897, for example, the llvm.coro.destroy instruction is only
executed in exception handling code. When the coroutine function exits
normally, llvm.coro.destroy is not called. Eliding the allocation in
this case causes a subsequent reference to the coroutine handle from
outside of the function to access freed memory.
To fix the issue, when finding an llvm.coro.destroy for each llvm.coro.begin,
only consider llvm.coro.destroy that are executed along non-exceptional paths.
Test Plan:
1. Download the sample program from
https://bugs.llvm.org/show_bug.cgi?id=34897, compile it with
`clang++ -fcoroutines-ts -stdlib=libc++ -std=c++1z -O2`, and run it.
It should print `"run1\ncheck1\nrun2\ncheck2"` and then exit
successfully.
2. Compile https://godbolt.org/g/mCKfnr and confirm it is still
optimized to a single instruction, 'return 1190'.
3. `check-llvm`
Reviewers: rsmith, GorNishanov, eric_niebler
Reviewed By: GorNishanov
Subscribers: andrewrk, lewissbaker, EricWF, llvm-commits
Differential Revision: https://reviews.llvm.org/D43242
llvm-svn: 332077
These rotates take the form
(x << (n & mask)) | (x >> (-n & mask)) where mask is bitwidth - 1.
If x has been promoted to a wider type than its original bit width due to type promotion we fail to narrower it and therefore don't recognize it as a rotate.
llvm-svn: 332068
We found current sampleFDO had a performance issue when triaging a regression.
For a callsite with inline instance in the profile, even if hot callsite inliner
cannot inline it, it may still execute enough times and should not be treated as
cold in regular inliner later. However, currently if such callsite is not inlined
by hot callsite inliner, and the BB where the callsite locates doesn't get
samples from other instructions inside of it, the callsite will have no profile
metadata annotated. In regular inliner cost analysis, if the callsite has no
profile annotated and its caller has profile information, it will be treated as
cold.
The fix changes the isCallsiteHot check and chooses to compare
CallsiteTotalSamples with hot cutoff value computed by ProfileSummaryInfo.
Differential Revision: https://reviews.llvm.org/D45377
llvm-svn: 332058
This reverts commit SVN r331889, which could trigger failed
assertions for cases where the snprintf function is declared
with a vaguely differing signature (e.g. being defined as
static inline), see PR37408.
llvm-svn: 332043
Accessing the members of a large data structures needs a lot of GEPs which
usually have large offsets due to the size of the underlying data structure. If
the offsets are too large to fit into the r+i addressing mode, these GEPs cannot
be sunk to their users' blocks and many extra registers are needed then to carry
the values of these GEPs.
This patch tries to split a large data struct starting from %base like the
following.
Before:
BB0:
%base =
BB1:
%gep0 = gep %base, off0
%gep1 = gep %base, off1
%gep2 = gep %base, off2
BB2:
%load1 = load %gep0
%load2 = load %gep1
%load3 = load %gep2
After:
BB0:
%base =
%new_base = gep %base, off0
BB1:
%new_gep0 = %new_base
%new_gep1 = gep %new_base, off1 - off0
%new_gep2 = gep %new_base, off2 - off0
BB2:
%load1 = load i32, i32* %new_gep0
%load2 = load i32, i32* %new_gep1
%load3 = load i32, i32* %new_gep2
In the above example, the struct is split into two parts. The first part still
starts from %base and the second part starts from %new_base. After the
splitting, %new_gep1 and %new_gep2 have smaller offsets and then can be sunk to
BB2 and folded into their users.
The algorithm to split data structure is simple and very similar to the work of
merging SExts. First, it collects GEPs that have large offsets when iterating
the blocks. Second, it splits the underlying data structures and updates the
collected GEPs to use smaller offsets.
Differential Revision: https://reviews.llvm.org/D42759
llvm-svn: 332015
Summary:
This change teaches DSE that the atomic memory intrinsics can be overwriten
partially in the same way as the non-atomic forms. Specifically, that the
atomic memcpy & memset can be shortened at the end and that the atomic memset
can be shortened at the beginning, if they partially overwritten
by later stores.
Reviewers: mkazantsev, skatkov, apilipenko, efriedma, rsmith, spatel, filcab, sanjoy
Reviewed By: efriedma
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45584
llvm-svn: 331991
Fixes bug https://bugs.llvm.org/show_bug.cgi?id=37339.
InlineAsm is only uniqued if the FunctionTypes are exactly the
same, while cmpTypes() for example considers all pointer types
in the default address space to be the same. For this reason
the end of cmpInlineAsm() can be reached.
This patch replaces the unreachable assertion with a check that
the function types are not identical.
Differential Revision: https://reviews.llvm.org/D46495
Reviewers: jfb
llvm-svn: 331990
This is a follow-up to D45986. As suggested there, we should match the "all-bits-set"
pattern in addition to "any-bits-set".
This was a little more complicated than I thought it would be initially because the
"and 1" instruction can be anywhere in the chain. Hopefully, the code comments make
that logic understandable, but if you see a way to simplify or improve that, it's
most appreciated.
This transforms patterns that emerge from bitfield tests as seen in PR37098:
https://bugs.llvm.org/show_bug.cgi?id=37098
I think it would also help reduce the large test from:
D46336
D46595
but we need something to reassociate that case to the forms we're expecting here first.
Differential Revision: https://reviews.llvm.org/D46649
llvm-svn: 331937
The previous handling for guard widening in InstCombine was extremely restrictive. In particular, it didn't handle the common case where we had two guards separated by a single icmp. Handle this by scanning through a small fixed window of instructions to find the next guard if needed.
Differential Revision: https://reviews.llvm.org/D46203
llvm-svn: 331935
This is safe as long as the udiv is not exact. The pattern is not common in
C++ code, but comes up all the time in code generated by XLA's GPU backend.
Differential Revision: https://reviews.llvm.org/D46647
llvm-svn: 331933
Summary:
MergedLoadStoreMotion::mergeStores is using some heuristics
to limit the amount of stores that it tries to sink (see
MagicCompileTimeControl in MergedLoadStoreMotion.cpp). The
heuristic involves counting the number of instructions in
one of the basic blocks that is part of the transformation.
We now ignore dbg intrinsics when counting instruction for
the MagicCompileTimeControl heuristic. This to make sure that
the amount of stores that are sunk doesn't depend on the amount
of debug information (if -g is used or not).
Reviewers: Gerolf, davide, majnemer
Reviewed By: davide
Subscribers: dberlin, bjope, aprantl, JDevlieghere, llvm-commits
Differential Revision: https://reviews.llvm.org/D46600
llvm-svn: 331852
In order to set breakpoints on labels and list source code around
labels, we need collect debug information for labels, i.e., label
name, the function label belong, line number in the file, and the
address label located. In order to keep these information in LLVM
IR and to allow backend to generate debug information correctly.
We create a new kind of metadata for labels, DILabel. The format
of DILabel is
!DILabel(scope: !1, name: "foo", file: !2, line: 3)
We hope to keep debug information as much as possible even the
code is optimized. So, we create a new kind of intrinsic for label
metadata to avoid the metadata is eliminated with basic block.
The intrinsic will keep existing if we keep it from optimized out.
The format of the intrinsic is
llvm.dbg.label(metadata !1)
It has only one argument, that is the DILabel metadata. The
intrinsic will follow the label immediately. Backend could get the
label metadata through the intrinsic's parameter.
We also create DIBuilder API for labels to be used by Frontend.
Frontend could use createLabel() to allocate DILabel objects, and use
insertLabel() to insert llvm.dbg.label intrinsic in LLVM IR.
Differential Revision: https://reviews.llvm.org/D45024
Patch by Hsiangkai Wang.
llvm-svn: 331841
Summary:
The current LowerInvoke pass cannot handle invoke instructions with a
funclet bundle operand. The order of operands for an invoke instruction
is {call arguments, callee, funclet operand (if any), normal dest,
unwind dest}. The current code assumes there is no funclet operand and
incorrectly includes a funclet operand into call arguments.
Reviewers: rnk
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D46242
llvm-svn: 331832
We enter MergeBlockIntoPredecessor with a block looking like this:
for.inc.us-lcssa: ; preds = %cond.end
%k.1.lcssa.ph = phi i32 [ %conv15, %cond.end ]
%t.3.lcssa.ph = phi i32 [ %k.1.lcssa.ph, %cond.end ]
br label %for.inc, !dbg !66
[note the first arg of the PHI being a PHI].
FoldSingleEntryPHINodes gets rid of both PHIs (calling, eraseFromParent).
But right before we call the function, we push into IncomingValues the
only argument of the PHIs, and shortly after we try to iterate over
something which has been invalidated before :(
The fix its not trying to remove PHIs which have an incoming value
coming from the same BB we're looking at.
Fixes PR37300 and rdar://problem/39910460
Differential Revision: https://reviews.llvm.org/D46568
llvm-svn: 331824
Summary:
Broadcast code generation emitted instructions in pre-header, while the instruction they are dependent on in the vector loop body.
This resulted in an IL verification error ---- value used before defined.
Reviewers: rengolin, fhahn, hfinkel
Reviewed By: rengolin, fhahn
Subscribers: dcaballe, Ka-Ka, llvm-commits
Differential Revision: https://reviews.llvm.org/D46302
llvm-svn: 331799
CodeGenPrepare pass move extension instructions close to load instructions in different BB, so they can be combined later. But the extension instructions can't move through logical and shift instructions in current implementation. This patch enables this enhancement, so we can eliminate more extension instructions.
Differential Revision: https://reviews.llvm.org/D45537
llvm-svn: 331783
Summary:
In formLCSSAForInstructions we speculatively add new PHI
nodes, that sometimes ends up without having any uses. It
has been discovered that sometimes an added PHI node can
appear as being unused in one iteration of the Worklist,
although it can end up being used by a PHI node added in
a later iteration. We now check, a second time, that the
PHI node still is unused before we remove it. This avoids
an assert about "Trying to remove a phi with uses." for the
added test case.
Reviewers: davide, mzolotukhin, mattd, dberlin
Reviewed By: mzolotukhin, dberlin
Subscribers: dberlin, mzolotukhin, davide, bjope, uabelho, llvm-commits
Differential Revision: https://reviews.llvm.org/D46422
llvm-svn: 331741
Summary: Makes this consistent with the old PM.
Reviewers: eraman
Subscribers: mehdi_amini, llvm-commits
Differential Revision: https://reviews.llvm.org/D46526
llvm-svn: 331709
to be needed: jump table sections are created with .cfi.jumptable suffix. With
this change each jump table is placed in a separate section, which allows the
linker to re-order them.
Differential Revision: https://reviews.llvm.org/D46537
llvm-svn: 331680
This pattern came up in D46494.
I'm pretty sure we want to canonicalize it from
(x | ~m) & (y & m)
to
(x & m) | (y & ~m)
https://rise4fun.com/Alive/TEM
llvm-svn: 331625
Summary:
launder.invariant.group has the same rules of capturing as
bitcast, gep, etc - the original value is not captured
if the returned pointer is not captured.
With this patch, we mark 40% more functions as noalias when compiling with -fstrict-vtable-pointers;
1078 vs 1778 (39.37%)
Reviewers: sanjoy, davide, nlewycky, majnemer, mehdi_amini
Subscribers: JDevlieghere, llvm-commits
Differential Revision: https://reviews.llvm.org/D32673
llvm-svn: 331587
Computing this property within the existing walk ensures that the cost is linear with the size of the block. If we did this from within isGuaranteedToExecute, it would be quadratic without some very fancy caching.
This allows us to reliably catch a hoistable instruction within a header which may throw at some point *after* our hoistable instruction. It doesn't do anything for non-header cases, but given how common single block loops are, this seems very worthwhile.
llvm-svn: 331557
A catchswitch is both a pad and a terminator, meaning it must be the
only non-phi instruction in its basic block. When we're inserting a
bitcast in the incoming basic block for a phi, if that incoming block is
a catchswitch, we should go up the dominator tree to find a valid
insertion point rather than attempting to insert before the catchswitch
(which would result in invalid IR).
Differential Revision: https://reviews.llvm.org/D46412
llvm-svn: 331548
We currently recognize this idiom where x is signed and thus the shift in an ashr.
int cnt = 0;
while (x) {
x >>= 1; // arithmetic shift right
++cnt;
}
and turn it into (bitwidth - ctlz(x)). And if there is anything else in the loop we will create a new loop that runs that many times.
If x is initially negative, the shift result will never be 0 and thus the loop is infinite. If you put something with side effects in the loop, that side effect will now only happen bitwidth times instead of an infinite number of times.
So this transform is only safe for logical shift right (which we don't currently recognize) or if we can prove that x cannot be negative before the loop.
llvm-svn: 331493
Add logic for the special case when a cmp+select can clearly be
reduced to just a bitwise logic instruction, and remove an
over-reaching chunk of general purpose bit magic. The primary goal
is to remove cases where we are not improving the IR instruction
count when doing these select transforms, and in all cases here that
is true.
In the motivating 3-way compare tests, there are further improvements
because we can combine/propagate select values (not sure if that
belongs in instcombine, but it's there for now).
DAGCombiner has folds to turn some of these selects into bit magic,
so there should be no difference in the end result in those cases.
Not all constant combinations are handled there yet, however, so it
is possible that some targets will see more cmov/csel codegen with
this change in IR canonicalization.
Ideally, we'll go further to *not* turn selects into multiple
logic/math ops in instcombine, and we'll canonicalize to selects.
But we should make sure that this step does not result in regressions
first (and if it does, we should fix those in the backend).
The general direction for this change was discussed here:
http://lists.llvm.org/pipermail/llvm-dev/2016-September/105373.htmlhttp://lists.llvm.org/pipermail/llvm-dev/2017-July/114885.html
Alive proofs for the new bit magic:
https://rise4fun.com/Alive/XG7
Differential Revision: https://reviews.llvm.org/D46086
llvm-svn: 331486
The code fails to check that the same value is used twice. We only make sure the left hand side of the and is part of the loop recurrence. The 'x' in the subtract can be any value.
llvm-svn: 331436
This patch was temporarily reverted because it has exposed bug 37229 on
PowerPC platform. The bug is unrelated to the patch and was just a general
bug in the optimization done for PowerPC platform only. The bug was fixed
by the patch rL331410.
This patch returns the disabled commit since the bug was fixed.
llvm-svn: 331427
Summary:
Prior to this change, LLVM would in some cases emit *massive* writeout
functions with many 10s of 1000s of function calls in straight-line
code. This is a very wasteful way to represent what are fundamentally
loops and creates a number of scalability issues. Among other things,
register allocating these calls is extremely expensive. While D46127 makes this
less severe, we'll still run into scaling issues with this eventually. If not
in the compile time, just from the code size.
Now the pass builds up global data structures modeling the inputs to
these functions, and simply loops over the data structures calling the
relevant functions with those values. This ensures that the code size is
a fixed and only data size grows with larger amounts of coverage data.
A trivial change to IRBuilder is included to make it easier to build
the constants that make up the global data.
Reviewers: wmi, echristo
Subscribers: sanjoy, mcrosier, llvm-commits, hiraditya
Differential Revision: https://reviews.llvm.org/D46357
llvm-svn: 331407
Summary:
Some of our internal testing detected a major compile time regression which I've
tracked down to:
r278938 - Revert "Reassociate: Reprocess RedoInsts after each inst".
It appears that processing long chains of reassociatable instructions causes
non-linear (potentially exponential) growth in the number of times an
instruction is revisited. For example, the included test revisits instructions
220 times in a 20-instruction test.
It appears that r278938 reversed the order instructions were visited and that
this is preventing scheduled revisits from being cancelled as a result of
visiting the instructions naturally during normal processing. However, simply
reversing the order also harmed the generated code. Upon closer inspection, it
was discovered that revisits occurred in the opposite order to the first pass
(Thanks to escha for spotting that).
This patch makes the revisit order consistent with the first pass which allows
more revisits to be cancelled. This does appear to have a small impact on the
generated code in few cases but it significantly reduces compile-time.
After this patch, our internal test that was most affected by the regression
dropped from ~2 million revisits to ~4k resulting in Reassociate having 0.46%
of the runtime it had before (99.54% improvement).
Here's the summaries reported by lnt for the LLVM test-suite with --benchmarking-only:
| metric | geomean before patch | geomean after patch | delta |
| ----- | ----- | ----- | ----- |
| compile time | 0.1956 | 0.1261 | -35.54% |
| execution time | 0.3240 | 0.3237 | - |
| code size | 7365.4459 | 7365.6079 | - |
The results have a few wins and losses on compile-time, mostly in the +/- 2.5% range. There was one outlier though:
| Performance Regressions - compile_time | Δ | Previous | Current |
| MultiSource/Benchmarks/ASC_Sequoia/CrystalMk/CrystalMk | 9.82% | 2.0473 | 2.2483 |
Reviewers: javed.absar, dberlin
Reviewed By: dberlin
Subscribers: kristof.beyls, llvm-commits
Differential Revision: https://reviews.llvm.org/D45734
llvm-svn: 331381
and (or (lshr X, C), ...), 1 --> (X & C') != 0
I initially thought about implementing the minimal pattern in instcombine as mentioned here:
https://bugs.llvm.org/show_bug.cgi?id=37098#c6
...but we need to do better to catch the more general sequence from the motivating test
(more than 2 bits in the compare). And a test-suite run with statistics showed that this
pattern only happened 2 times currently. It would potentially happen more often if
reassociation worked better (D45842), but it's probably still not too frequent?
This is small enough that I didn't see a need to create a whole new class/file within
AggressiveInstCombine. There are likely other relatively small matchers like what was
discussed in D44266 that would slide under foldUnusualPatterns() (name suggestions welcome).
We could potentially also consolidate matchers for ctpop, bswap, etc under here.
Differential Revision: https://reviews.llvm.org/D45986
llvm-svn: 331311
As mentioned in D45986, there's a potential ordering dependency
between instcombine and aggressive-instcombine for detecting these,
so I'm adding a few tests to confirm that the expected folds occur
using -O3 (because aggressive-instcombine only runs at -O3 currently).
llvm-svn: 331308
Summary:
This fixes a build break with r331269.
test/Transforms/LoopVectorize/pr23997.ll
should be in:
test/Transforms/LoopVectorize/X86/pr23997.ll
llvm-svn: 331281
Summary:
This is a fix for PR23997.
The loop vectorizer is not preserving the inbounds property of GEPs that it creates.
This is inhibiting some optimizations. This patch preserves the inbounds property in
the case where a load/store is being fed by an inbounds GEP.
Reviewers: mkuper, javed.absar, hsaito
Reviewed By: hsaito
Subscribers: dcaballe, hsaito, llvm-commits
Differential Revision: https://reviews.llvm.org/D46191
llvm-svn: 331269
phi is on lhs of a comparison op.
For the following testcase,
L1:
%t0 = add i32 %m, 7
%t3 = icmp eq i32* %t2, null
br i1 %t3, label %L3, label %L2
L2:
%t4 = load i32, i32* %t2, align 4
br label %L3
L3:
%t5 = phi i32 [ %t0, %L1 ], [ %t4, %L2 ]
%t6 = icmp eq i32 %t0, %t5
br i1 %t6, label %L4, label %L5
We know if we go through the path L1 --> L3, %t6 should always be true. However
currently, if the rhs of the eq comparison is phi, JumpThreading fails to
evaluate %t6 to true. And we know that Instcombine cannot guarantee always
canonicalizing phi to the left hand side of the comparison operation according
to the operand priority comparison mechanism in instcombine. The patch handles
the case when rhs of the comparison op is a phi.
Differential Revision: https://reviews.llvm.org/D46275
llvm-svn: 331266
instcombine should transform the relevant cases if the OverflowingBinaryOperator/PossiblyExactOperator can be proven to be safe.
Change-Id: I7aec62a31a894e465e00eb06aed80c3ea0c9dd45
llvm-svn: 331265
This test had values that differed in only in capitalization,
and that causes problems for the auto-generating check line
script. So I changed that in rL331226, but I accidentally
forgot to change a subsequent use of a param.
llvm-svn: 331228
Summary:
As discussed in D45733, we want to do this in InstCombine.
https://rise4fun.com/Alive/LGk
Reviewers: spatel, craig.topper
Reviewed By: spatel
Subscribers: chandlerc, xbolva00, llvm-commits
Differential Revision: https://reviews.llvm.org/D45867
llvm-svn: 331205
Summary:
Masked merge has a pattern of: `((x ^ y) & M) ^ y`.
But, there is no difference between `((x ^ y) & M) ^ y` and `((x ^ y) & ~M) ^ x`,
We should canonicalize the pattern to non-inverted mask.
https://rise4fun.com/Alive/Yol
Reviewers: spatel, craig.topper
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45664
llvm-svn: 331112
Summary:
Masked merge has a pattern of: `((x ^ y) & M) ^ y`.
But, there is no difference between `((x ^ y) & M) ^ y` and `((x ^ y) & ~M) ^ x`,
We should canonicalize the pattern to non-inverted mask.
Differential Revision: https://reviews.llvm.org/D45663
llvm-svn: 331111
The effect of doing so is not disrupting the LoopPassManager when mixing this pass with other loop passes. This should help locality of access substaintially and avoids the cost of computing PostDom.
The assumption here is that the full GuardWidening (which does use PostDom) is run as a canonicalization before loop opts and that this version is just catching cases exposed by other loop passes. (i.e. LoopPredication, IndVarSimplify, LoopUnswitch, etc..)
llvm-svn: 331094
This patch adds support for fragment expressions
TryToShrinkGlobalToBoolean() which were previously just dropped.
Thanks to Reid Kleckner for providing me a reproducer!
llvm-svn: 331086
Summary:
Currently, we
1. match `LHS` matcher to the `first` operand of binary operator,
2. and then match `RHS` matcher to the `second` operand of binary operator.
If that does not match, we swap the `LHS` and `RHS` matchers:
1. match `RHS` matcher to the `first` operand of binary operator,
2. and then match `LHS` matcher to the `second` operand of binary operator.
This works ok.
But it complicates writing of commutative matchers, where one would like to match
(`m_Value()`) the value on one side, and use (`m_Specific()`) it on the other side.
This is additionally complicated by the fact that `m_Specific()` stores the `Value *`,
not `Value **`, so it won't work at all out of the box.
The last problem is trivially solved by adding a new `m_c_Specific()` that stores the
`Value **`, not `Value *`. I'm choosing to add a new matcher, not change the existing
one because i guess all the current users are ok with existing behavior,
and this additional pointer indirection may have performance drawbacks.
Also, i'm storing pointer, not reference, because for some mysterious-to-me reason
it did not work with the reference.
The first one appears trivial, too.
Currently, we
1. match `LHS` matcher to the `first` operand of binary operator,
2. and then match `RHS` matcher to the `second` operand of binary operator.
If that does not match, we swap the ~~`LHS` and `RHS` matchers~~ **operands**:
1. match ~~`RHS`~~ **`LHS`** matcher to the ~~`first`~~ **`second`** operand of binary operator,
2. and then match ~~`LHS`~~ **`RHS`** matcher to the ~~`second`~ **`first`** operand of binary operator.
Surprisingly, `$ ninja check-llvm` still passes with this.
But i expect the bots will disagree..
The motivational unittest is included.
I'd like to use this in D45664.
Reviewers: spatel, craig.topper, arsenm, RKSimon
Reviewed By: craig.topper
Subscribers: xbolva00, wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D45828
llvm-svn: 331085
As suggested in D45842
(although still not sure if we're going to advance that),
we must invalidate references to instructions that have
been recycled (operands were changed, so result is different).
llvm-svn: 331083
We currently have a hard to solve analysis problem around the order of instructions within a potentially throwing block. We can't cheaply determine whether a given instruction is before the first potential throw in the block. While we're working on that in the background, special case the first instruction within the header.
why this particular special case? Well, headers are guaranteed to execute if the loop does, and it turns out we tend to produce this form in practice.
In a follow on patch, I tend to extend LICM with an alternate approach which works for any instruction in the header before the first throw, but this is the best I can come up with other users of the analysis (such as store promotion.)
Note: I can't show the difference in the analysis result since we're ORing in the expensive instruction walk used by SCEV. Using the full walk is not suitable for a general solution.
llvm-svn: 331079
The idea is to have a pass which performs the same transformation as GuardWidening, but can be run within a loop pass manager without disrupting the pass manager structure. As demonstrated by the test case, this doesn't quite get there because of issues with post dom, but it gives a good step in the right direction. the motivation is purely to reduce compile time since we can now preserve locality during the loop walk.
This patch only includes a legacy pass. A follow up will add a new style pass as well.
llvm-svn: 331060
We currently support LCSSA PHI nodes in the outer loop exit, if their
incoming values do not come from the outer loop latch or if the
outer loop latch has a single predecessor. In that case, the outer loop latch
will be executed only if the inner loop gets executed. If we have multiple
predecessors for the outer loop latch, it may be executed even if the inner
loop does not get executed.
This is a first step to support the case described in
https://bugs.llvm.org/show_bug.cgi?id=30472
Reviewers: efriedma, karthikthecool, mcrosier
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D43237
llvm-svn: 331037
Since PTX has grown a <2 x half> datatype vectorization has become more
important. The late LoadStoreVectorizer intentionally only does loads
and stores, but now arithmetic has to be vectorized for optimal
throughput too.
This is still very limited, SLP vectorization happily creates <2 x half>
if it's a legal type but there's still a lot of register moving
happening to get that fed into a vectorized store. Overall it's a small
performance win by reducing the amount of arithmetic instructions.
I haven't really checked what the loop vectorizer does to PTX code, the
cost model there might need some more tweaks. I didn't see it causing
harm though.
Differential Revision: https://reviews.llvm.org/D46130
llvm-svn: 331035
It doesn't unwind, and the wrong marking leads to the creation of an
.eh_frame section when it isn't necessary.
Differential Revision: https://reviews.llvm.org/D46082
llvm-svn: 331008
Summary: If file stream arg is not captured and source is fopen, we could replace IO calls by unlocked IO ("_unlocked" function variants) to gain better speed,
Reviewers: efriedma, RKSimon, spatel, sanjoy, hfinkel, majnemer
Subscribers: lebedev.ri, llvm-commits
Differential Revision: https://reviews.llvm.org/D45736
llvm-svn: 331002
Summary:
Simplify integer add expression X % C0 + (( X / C0 ) % C1) * C0 to
X % (C0 * C1). This is a common pattern seen in code generated by the XLA
GPU backend.
Add test cases for this new optimization.
Patch by Bixia Zheng!
Reviewers: sanjoy
Reviewed By: sanjoy
Subscribers: efriedma, craig.topper, lebedev.ri, llvm-commits, jlebar
Differential Revision: https://reviews.llvm.org/D45976
llvm-svn: 330992
remainder expressions as operands.
Summary:
Add test cases to prepare for the new optimization that Simplifies integer add
expression X % C0 + (( X / C0 ) % C1) * C0 to X % (C0 * C1).
Patch by Bixia Zheng!
Reviewers: sanjoy
Reviewed By: sanjoy
Subscribers: jlebar, llvm-commits
Differential Revision: https://reviews.llvm.org/D46017
llvm-svn: 330991
This reverts commit 023c8be90980e0180766196cba86f81608b35d38.
This patch triggers miscompile of zlib on PowerPC platform. Most likely it is
caused by some pre-backend PPC-specific pass, but we don't clearly know the
reason yet. So we temporally revert this patch with intention to return it
once the problem is resolved. See bug 37229 for details.
llvm-svn: 330893
Summary:
When performing indirect call promotion, current implementation inspects "all" parameters of the callsite and attemps to match with the formal argument type of the callee function. However, it is not possible to find the type for variable length arguments, and the compiler crashes when it attemps to match the type for variable lenght argument.
It seems that the bug is introduced with D40658. Prior to that, the type matching is performed only for the parameters whose ID is less than callee->getFunctionNumParams(). The attached test case will crash without the patch.
Reviewers: mssimpso, davidxl, davide
Reviewed By: mssimpso
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D46026
llvm-svn: 330844
As discussed in D45862, we want these folds sometimes
because they're good improvements.
But as we can see here, the current logic doesn't
check uses and doesn't produce optimal code in all
cases.
llvm-svn: 330837
This also means we have to check if the latch is the exiting block now,
as `transform` expects the latches to be the exiting blocks too.
https://bugs.llvm.org/show_bug.cgi?id=36586
Reviewers: efriedma, davide, karthikthecool
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D45279
llvm-svn: 330806
Summary:
When Reassociate is rewriting an expression tree it may
reuse old binary expression nodes, for new expressions.
Whenever an expression node is reused, but with a non-trivial
change in the result, we need to invalidate any debug info
that is associated with the node.
If for example rewriting
x = mul a, b
y = mul c, x
into
x = mul c, b
y = mul a, x
we still get the same result for 'y', but 'x' is a new expression.
All debug info referring to 'x' must be invalidated (marked as
optimized out) since we no longer calculate the expected value.
As a side-effect this patch avoid (at least some) problems where
reassociate could end up creating IR with debug-use before def.
Earlier the dbg.value nodes where left untouched in the IR, while
the reused binary nodes where sinked to just before the root node
of the rewritten expression tree. See PR27273 for more info about
such problems.
Reviewers: dblaikie, aprantl, dexonsmith
Reviewed By: aprantl
Subscribers: JDevlieghere, llvm-commits
Tags: #debug-info
Differential Revision: https://reviews.llvm.org/D45975
llvm-svn: 330804
These are all but 1 of the select-of-constant tests that appear
to be transformed within foldSelectICmpAnd() and the block above
it predicated by decomposeBitTestICmp().
As discussed in D45862 (and can be seen in several tests here),
we probably want to stop doing those transforms because they
can increase the instruction count without benefitting other
passes or codegen.
The 1 test not included here is a urem test where the bit hackery
allows us to remove a urem. To preserve killing that urem, we
should do some stronger known-bits analysis or pattern matching of
'urem x, (select-of-pow2-constants)'.
llvm-svn: 330768
Patch #2 from VPlan Outer Loop Vectorization Patch Series #1
(RFC: http://lists.llvm.org/pipermail/llvm-dev/2017-December/119523.html).
This patch introduces the basic infrastructure to detect, legality check
and process outer loops annotated with hints for explicit vectorization.
All these changes are protected under the feature flag
-enable-vplan-native-path. This should make this patch NFC for the existing
inner loop vectorizer.
Reviewers: hfinkel, mkuper, rengolin, fhahn, aemerson, mssimpso.
Differential Revision: https://reviews.llvm.org/D42447
llvm-svn: 330739
After D43236, we started interchanging loops with empty dependence
matrices. In isProfitableForVectorization, we try to determine if
interchanging makes the loop dependences more friendly to the
vectorizer. If there are no dependences, we should not interchange,
based on that heuristic.
Reviewers: efriedma, mcrosier, karthikthecool, blitz.opensource
Reviewed By: mcrosier
Differential Revision: https://reviews.llvm.org/D45208
llvm-svn: 330738
The first step in fixing problems raised in D45862
is to make the problems visible. Now we can more easily
see/update cases where selects have been turned into
multiple instructions with no apparent improvement in
analysis or benefits for other passes (vectorization).
llvm-svn: 330731
The current version of the script uses regex for params.
This could mask a bug (param values got wrongly swapped),
but it seems unlikely in practice, so let's just update
the whole file to reduce diffs when there is a meaningful
change here.
llvm-svn: 330729
The memory location an invariant load is using can never be clobbered by
any store, so it's safe to move the load ahead of the store.
Differential Revision: https://reviews.llvm.org/D46011
llvm-svn: 330725
loop unswitch.
This code incorrectly added the header to the loop block set early. As
a consequence we would incorrectly conclude that a nested loop body had
already been visited when the header of the outer loop was the preheader
of the nested loop. In retrospect, adding the header eagerly doesn't
really make sense. It seems nicer to let the cycle be formed naturally.
This will catch crazy bugs in the CFG reconstruction where we can't
correctly form the cycle earlier rather than later, and makes the rest
of the logic just fall out.
I've also added various asserts that make these issues *much* easier to
debug.
llvm-svn: 330707
This code path can very clearly be called in a context where we have
baselined all the cloned blocks to a particular loop and are trying to
handle nested subloops. There is no harm in this, so just relax the
assert. I've added a test case that will make sure we actually exercise
this code path.
llvm-svn: 330680
I was reminded today that this patch got reverted in r301885. I can no
longer reproduce the failure that caused the revert locally (...almost
one year later), and the patch applied pretty cleanly, so I guess we'll
see if the bots still get angry about it.
The original breakage was InstSimplify complaining (in "assertion
failed" form) about getting passed some crazy IR when running `ninja
check-sanitizer`. I'm unable to find traces of what, exactly, said crazy
IR was. I suppose we'll find out pretty soon if that's still the case.
:)
Original commit:
Author: gbiv
Date: Mon May 1 18:12:08 2017
New Revision: 301880
URL: http://llvm.org/viewvc/llvm-project?rev=301880&view=rev
Log:
[InstSimplify] Handle selects of GEPs with 0 offset
In particular (since it wouldn't fit nicely in the summary):
(select (icmp eq V 0) P (getelementptr P V)) -> (getelementptr P V)
Differential Revision: https://reviews.llvm.org/D31435
llvm-svn: 330667
I'm not sure if this is where we should try to fold these
patterns inspired by:
https://bugs.llvm.org/show_bug.cgi?id=37098
...if this isn't the right place, we can move the tests.
llvm-svn: 330642
Summary: We do not need nonull attribute if we know an argument is going to be constant.
Reviewers: junbuml, davide, fhahn
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45608
llvm-svn: 330641
Summary:
Skip basic blocks not reachable from the entry node
in MemCpyOptPass::iterateOnFunction.
Code that is unreachable may have properties that do not exist
for reachable code (an instruction in a basic block can for
example be dominated by a later instruction in the same basic
block, for example if there is a single block loop).
MemCpyOptPass::processStore is only safe to use for reachable
basic blocks, since it may iterate past the basic block
beginning when used for unreachable blocks. By simply skipping
to optimize unreachable basic blocks we can avoid asserts such
as "Assertion `!NodePtr->isKnownSentinel()' failed."
in MemCpyOptPass::processStore.
The problem was detected by fuzz tests.
Reviewers: eli.friedman, dneilson, efriedma
Reviewed By: efriedma
Subscribers: efriedma, llvm-commits
Differential Revision: https://reviews.llvm.org/D45889
llvm-svn: 330635
Summary:
This change teaches DSE that the atomic memory intrinsics are stores
that can be eliminated, and can allow other stores to be eliminated.
This change specifically does not teach DSE that these intrinsics
can be partially eliminated (i.e. length reduced, and dest/src changed);
that will be handled in another change.
Reviewers: mkazantsev, skatkov, apilipenko, efriedma, rsmith
Reviewed By: efriedma
Subscribers: dmgreen, llvm-commits
Differential Revision: https://reviews.llvm.org/D45535
llvm-svn: 330629
LoopRotate only invalidates innermost loops while the changes that it makes may
also affert any of this parents. With patch rL329047, SCEV becomes much smarter
about calculation of exit counts for outer loops, so we cannot assume that they are
not affected.
Differential Revision: https://reviews.llvm.org/D45945
llvm-svn: 330582
In the function `simplifyOneLoop` we optimistically assume that changes in the
inner loop only affect this very loop and have no impact on its parents. In fact,
after rL329047 has been merged, we can now calculate exit counts for outer
loops which may depend on inner loops. Thus, we need to invalidate all parents
when we do something to a loop.
There is an evidence of incorrect behavior of `simplifyOneLoop`: when we insert
`SE->verify()` check in the end of this funciton, it fails on a bunch of existing
test, in particular:
LLVM :: Transforms/LoopUnroll/peel-loop-not-forced.ll
LLVM :: Transforms/LoopUnroll/peel-loop-pgo.ll
LLVM :: Transforms/LoopUnroll/peel-loop.ll
LLVM :: Transforms/LoopUnroll/peel-loop2.ll
Note that previously we have fixed issues of this variety, see rL328483.
This patch makes this function invalidate the outermost loop properly.
Differential Revision: https://reviews.llvm.org/D45937
Reviewed By: chandlerc
llvm-svn: 330576
The condition this was asserting doesn't actually hold. I've added
comments to explain why, removed the assert, and added a fun test case
reduced out of 403.gcc.
llvm-svn: 330564
This is the last step in getting constant pattern matchers to allow
undef elements in constant vectors.
I'm adding a dedicated m_ZeroInt() function and building m_Zero() from
that. In most cases, calling code can be updated to use m_ZeroInt()
directly when there's no need to match pointers, but I'm leaving that
efficiency optimization as a follow-up step because it's not always
clear when that's ok.
There are just enough icmp folds in InstSimplify that can be used for
integer or pointer types, that we probably still want a generic m_Zero()
for those cases. Otherwise, we could eliminate it (and possibly add a
m_NullPtr() as an alias for isa<ConstantPointerNull>()).
We're conservatively returning a full zero vector (zeroinitializer) in
InstSimplify/InstCombine on some of these folds (see diffs in InstSimplify),
but I'm not sure if that's actually necessary in all cases. We may be
able to propagate an undef lane instead. One test where this happens is
marked with 'TODO'.
llvm-svn: 330550
When creating a call to storeStrong in ObjCARCContract, ensure the call
gets the correct funclet token, otherwise WinEHPrepare will turn the
call (and all subsequent instructions) into unreachable.
We already have logic to do this for the ARC autorelease elision marker;
factor that out into a common function that's used for both. These are
the only two places in this transform that create call instructions.
Differential Revision: https://reviews.llvm.org/D45857
llvm-svn: 330487
The callback used to create an ORE for the legacy PI pass caches the allocated
object in a unique_ptr in the runOnModule function, and returns a reference to
that object. Under certian circumstances we can end up holding onto that
reference after the OREs destruction. Rather then allowing the new and legacy
passes to create ORE object in diffrent ways, create the ORE at the point of
use.
Differential Revision: https://reviews.llvm.org/D43219
llvm-svn: 330473
It also adds a check making sure PHIs for operands are all in the same
block.
Patch by Daniel Berlin <dberlin@dberlin.org>
Reviewers: dberlin, davide
Differential Revision: https://reviews.llvm.org/D43865
llvm-svn: 330444
Reapply the patches with a fix. Thanks Ilya and Hans for the reproducer!
This reverts commit r330416.
The issue was that removing predecessors invalidated uses that we stored
for rewrite. The fix is to finish manipulating with CFG before we select
uses for rewrite.
llvm-svn: 330431
Revert r330413: "[SSAUpdaterBulk] Use SmallVector instead of DenseMap for storing rewrites."
Revert r330403 "Reapply "[PR16756] Use SSAUpdaterBulk in JumpThreading." one more time."
r330403 commit seems to crash clang during our integrate while doing PGO build with the following stacktrace:
#2 llvm::SSAUpdaterBulk::RewriteAllUses(llvm::DominatorTree*, llvm::SmallVectorImpl<llvm::PHINode*>*)
#3 llvm::JumpThreadingPass::ThreadEdge(llvm::BasicBlock*, llvm::SmallVectorImpl<llvm::BasicBlock*> const&, llvm::BasicBlock*)
#4 llvm::JumpThreadingPass::ProcessThreadableEdges(llvm::Value*, llvm::BasicBlock*, llvm::jumpthreading::ConstantPreference, llvm::Instruction*)
#5 llvm::JumpThreadingPass::ProcessBlock(llvm::BasicBlock*)
The crash happens while compiling 'lib/Analysis/CallGraph.cpp'.
r3340413 is reverted due to conflicting changes.
llvm-svn: 330416
Hopefully, changing set to vector removes nondeterminism detected by
some bots, or the new assert will catch something.
This reverts commit r330180.
llvm-svn: 330403
Summary:
This change fixes https://crbug.com/834474, a build failure caused by
LowerTypeTests not preserving .symver symbol versioning directives for
exported functions. Emit symver information to ThinLTO summary data and
then propagate symver directives for exported functions to the merged
module.
Emitting symver information to the summaries increases the size of
intermediate build artifacts for a Chromium build by less than 0.2%.
Reviewers: pcc
Reviewed By: pcc
Subscribers: tejohnson, mehdi_amini, eraman, llvm-commits, eugenis, kcc
Differential Revision: https://reviews.llvm.org/D45798
llvm-svn: 330387
Similar to rL330086, I don't know if we want to do these
transforms here, but we might as well have the tests
here either way to show that this pass is missing
potential functionality (intentionally or not).
llvm-svn: 330368
Summary:
This fixes the bug pointed out in review with non-trivial unswitching.
This also provides a basis that should make it pretty easy to finish
fleshing out a routine to scan an entire function body for irreducible
control flow, but this patch remains minimal for disabling loop
unswitch.
Reviewers: sanjoy, fedor.sergeev
Subscribers: mcrosier, hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D45754
llvm-svn: 330357
If those operands change, we might find a leader for ValueOp, which
could enable new phi-of-op creation.
This fixes a case where we missed creating a phi-of-ops node. With D43865
and this patch, bootstrapping clang/llvm works with -enable-newgvn, whereas
without it, the "value changed after iteration" assertion is triggered.
Reviewers: dberlin, davide
Reviewed By: dberlin
Differential Revision: https://reviews.llvm.org/D42180
llvm-svn: 330334
After investigation discussed in D45439, it would seem that the nsw
flag restriction is unnecessary in most cases. So the IsInductionVar
lambda has been removed, the functionality extracted, and now only
require nsw when using eq/ne predicates.
Differential Revision: https://reviews.llvm.org/D45617
llvm-svn: 330256
If a predicate does not become known after peeling, peeling is unlikely
to be beneficial.
Reviewers: mcrosier, efriedma, mkazantsev, junbuml
Reviewed By: mkazantsev
Differential Revision: https://reviews.llvm.org/D44983
llvm-svn: 330250
Summary:
When sinking an instruction in InstCombine we now also sink
the DbgInfoIntrinsics that are using the sunken value.
Example)
When sinking the load in this input
bb.X:
%0 = load i64, i64* %start, align 4, !dbg !31
tail call void @llvm.dbg.value(metadata i64 %0, ...)
br i1 %cond, label %for.end, label %for.body.lr.ph
for.body.lr.ph:
br label %for.body
we now also move the dbg.value, like this
bb.X:
br i1 %cond, label %for.end, label %for.body.lr.ph
for.body.lr.ph:
%0 = load i64, i64* %start, align 4, !dbg !31
tail call void @llvm.dbg.value(metadata i64 %0, ...)
br label %for.body
In the past we haven't moved the dbg.value so we got
bb.X:
tail call void @llvm.dbg.value(metadata i64 %0, ...)
br i1 %cond, label %for.end, label %for.body.lr.ph
for.body.lr.ph:
%0 = load i64, i64* %start, align 4, !dbg !31
br label %for.body
So in the past we got a debug-use before the def of %0.
And that dbg.value was also on the path jumping to %for.end, for
which %0 never was defined.
CodeGenPrepare normally comes to rescue later (when not moving
the dbg.value), since it moves dbg.value instrinsics quite
brutally, without really analysing if it is correct to move
the intrinsic (see PR31878).
So at the moment this patch isn't expected to have much impact,
besides that it is moving the dbg.value already in opt, making
the IR look more sane directly.
This can be seen as a preparation to (hopefully) make it possible
to turn off CodeGenPrepare::placeDbgValues later as a solution
to PR31878.
I also adjusted test/DebugInfo/X86/sdagsplit-1.ll to make the
IR in the test case up-to-date with this behavior in InstCombine.
Reviewers: rnk, vsk, aprantl
Reviewed By: vsk, aprantl
Subscribers: mattd, JDevlieghere, llvm-commits
Tags: #debug-info
Differential Revision: https://reviews.llvm.org/D45425
llvm-svn: 330243
The bitcast may be interfering with other combines or vectorization
as shown in PR16739:
https://bugs.llvm.org/show_bug.cgi?id=16739
Most pointer-related optimizations are probably able to look through
this bitcast, but removing the bitcast shrinks the IR, so it's at
least a size savings.
Differential Revision: https://reviews.llvm.org/D44833
llvm-svn: 330237
Track the debug locations of the incoming values to newly-created phis,
and apply merged debug locations to the phis.
A merged location will be on line 0, but will have the correct scope
set. This improves crash reporting when an inlined instruction with a
merged location triggers a machine exception. A debugger will be able to
narrow down the crash to the correct inlined scope, instead of simply
pointing to the outer scope of the caller.
Taken together with a change allows generating merged line-0 locations
for instructions which aren't calls, this results in a 0.5% increase in
the uncompressed size of the .debug_line section of a stage2+Release
build of clang (-O3 -g).
rdar://33858697
Differential Revision: https://reviews.llvm.org/D45397
llvm-svn: 330227
One more, hopefully the last, bug is fixed: when forming UsesToRewrite
we should ignore phi operands coming from edges that we want to delete.
This reverts r329910.
llvm-svn: 330175
We use getExtractWithExtendCost to calculate the cost of extractelement and
s|zext together when computing the extract cost after vectorization, but we
calculate the cost of extractelement and s|zext separately when computing the
scalar cost which is larger than it should be.
Differential Revision: https://reviews.llvm.org/D45469
llvm-svn: 330143
Two cleanups:
1. As noted in D45453, we had tests that don't need FMF that were misplaced in the 'fast-math.ll' test file.
2. This removes the final uses of dyn_castFNegVal, so that can be deleted. We use 'match' now.
llvm-svn: 330126
Summary:
In order to get the whole fold as specified in [[ https://bugs.llvm.org/show_bug.cgi?id=6773 | PR6773 ]],
let's first handle the simple straight-forward things.
Let's start with the `and` -> `or` simplification.
The one obvious thing missing here: the constant mask is not handled.
I have an idea how to handle it, but it will require some thinking,
and is not strictly required here, so i've left that for later.
https://rise4fun.com/Alive/Pkmg
Reviewers: spatel, craig.topper, eli.friedman, jingyue
Reviewed By: spatel
Subscribers: llvm-commits
Was reviewed as part of https://reviews.llvm.org/D45631
llvm-svn: 330103
Summary:
In order to get the whole fold as specified in [[ https://bugs.llvm.org/show_bug.cgi?id=6773 | PR6773 ]],
let's first handle the simple straight-forward things.
Let's start with the `and` -> `or` simplification.
The one obvious thing missing here: the constant mask is not handled.
I have an idea how to handle it, but it will require some thinking,
and is not strictly required here, so i've left that for later.
https://rise4fun.com/Alive/Pkmg
Reviewers: spatel, craig.topper, eli.friedman, jingyue
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45631
llvm-svn: 330101
These simplifications were previously enabled only with isFast(), but that
is more restrictive than required. Since r317488, FMF has 'reassoc' to
control these cases at a finer level.
llvm-svn: 330089
It debateable whether instcombine should be in the business of
reassociation, but it is currently.
These tests and PR37098 demonstrate a missing ability to do a
simple reassociation that allows eliminating shifts.
If we decide that functionality belongs somewhere else, then we
should still have some tests to show that we've intentionally
limited instcombine to *not* include this ability.
llvm-svn: 330086
(and plain 'or', for completeness sake.)
After submitting D45631, i have realized that it will *already*
affect 'and' pattern, and it was obvious that there were no
good test patterns to show that.
Since the masked-merge.ll is getting kinda big,
unify naming schemes a bit, and split into 'xor'/'and'/'or'
testfiles, with the only difference being the last operation.
llvm-svn: 330072
The function getMinimumVF(ElemWidth) will return the minimum VF for
a vector with elements of size ElemWidth bits. This value will only
apply to targets for which TTI::shouldMaximizeVectorBandwidth returns
true. The value of 0 indicates that there is no minimum VF.
Differential Revision: https://reviews.llvm.org/D45271
llvm-svn: 330062
This was intended since initially, but i did not really think
about it, and did not know how to force that. Now that the
xor->or fold is working (patch upcoming), this came up
to improve the test coverage.
A followup for rL330003, rL330007
https://bugs.llvm.org/show_bug.cgi?id=6773
llvm-svn: 330039
Summary:
The fold added in D45108 did not account for the fact that
the and instruction is commutative, and if the mask is a variable,
the mask variable and the fold variable may be swapped.
I have noticed this by accident when looking into [[ https://bugs.llvm.org/show_bug.cgi?id=6773 | PR6773 ]]
This extends/generalizes that fold, so it is handled too.
Reviewers: spatel, craig.topper
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45539
llvm-svn: 330001
This completes the work started in r329604 and r329605 when we changed clang to no longer use the intrinsics.
We lost some InstCombine SimplifyDemandedBit optimizations through this change as we aren't able to fold 'and', bitcast, shuffle very well.
llvm-svn: 329990
This lifts a restriction on DILocation::getMergedLocation(), allowing it
to create merged locations for instructions other than calls.
Instruction::applyMergedLocation() now defaults to creating merged
locations for all instructions.
The default behavior of getMergedLocation() is unchanged: callers which
invoke it directly are unaffected.
This change will enable a follow-up Mem2Reg fix which improves crash
reporting.
Differential Revision: https://reviews.llvm.org/D45396
llvm-svn: 329955
Created a helper function to query for non negative SCEVs. Uses the
SGE predicate to catch constants that could be interpreted as
negative.
Differential Revision: https://reviews.llvm.org/D45481
llvm-svn: 329907
Summary:
The fold added in D45108 did not account for the fact that
the and instruction is commutative, and if the mask is a variable,
the mask variable and the fold variable may be swapped.
I have noticed this by accident when looking into [[ https://bugs.llvm.org/show_bug.cgi?id=6773 | PR6773 ]]
Reviewers: spatel, craig.topper
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45538
llvm-svn: 329901
We're already removing allocsize attributes from Functions that we
remove args from, since removing arguments from a function may make the
allocsize attribute incorrect. It appears we forgot to also remove them
from callsites.
Without this, I get verifier errors on `@Test2`.
It probably wouldn't be too hard to make DAE properly update allocsize
attributes instead of dropping them, but I can't think of a scenario
where that'd be useful in practice.
llvm-svn: 329868
Summary:
These tests show that DSE currently does nothing with the atomic memory
intrinsics. Future work will teach DSE how to simplify these.
llvm-svn: 329845
Summary:
In preparation for a future commit, this regenerates the test checks for
test/Transforms/DeadStoreElimination/OverwriteStoreBegin.ll
test/Transforms/DeadStoreElimination/OverwriteStoreEnd.ll
llvm-svn: 329839
Summary:
In preparation for a future commit, this regenerates the test checks for
test/Transforms/DeadStoreElimination/simple.ll
test/Transforms/DeadStoreElimination/memintrinsics.ll
llvm-svn: 329824
Author: Samuel Pitoiset
ds_read_b128 and ds_write_b128 have been recently enabled
under the amdgpu-ds128 option because the performance benefit
is unclear.
Though, using 128-bit loads/stores for the local address space
appears to introduce regressions in tessellation shaders. Not
sure what is broken, but as ds_read_b128/ds_write_b128 are not
enabled by default, just introduce a global option and enable
128-bit only if requested (until it's fixed/used correctly).
v2: - fix regressions in merge-stores.ll and multiple_tails.ll
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105464
llvm-svn: 329764
This is based on an example that was recently posted on llvm-dev:
void *propagate_null(void* b, int* g) {
if (!b) {
return 0;
}
(*g)++;
return b;
}
https://godbolt.org/g/xYk3qG
The original code or constant propagation in other passes has obscured the fact
that the phi can be removed completely.
Differential Revision: https://reviews.llvm.org/D45448
llvm-svn: 329755
Fix PR36484, as suggested:
<quote>
during moves, mark the direct users of the erased things that were phis as "not to be optimized"
<quote>
llvm-svn: 329621
Summary:
We do not try to move the instructions and split the block till we
know the blocks can be split, i.e. BCE-cmp-insts can be separated from
non-BCE-cmp-insts.
Reviewers: davide, courbet
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D44443
llvm-svn: 329564
In IRCE, we have a very old legacy check that works when we collect comparisons that we
treat as range checks. It ensures that the value against which the indvar is compared is
loop invariant and is also positive.
This latter condition remained there since the times when IRCE was only able to handle
signed latch comparison. As the optimization evolved, it now learned how to intersect
signed or unsigned ranges, and this logic has no reliance on the fact that the right border
of each range should be positive.
The old implementation of this non-negativity check was also naive enough and just looked
into ranges (while most of other IRCE logic tries to use power of SCEV implications), so this
check did not allow to deal with the most simple case that looks like follows:
int size; // not known non-negative
int length; //known non-negative;
i = 0;
if (size != 0) {
do {
range_check(i < size);
range_check(i < length);
++i;
} while (i < size)
}
In this case, even if from some dominating conditions IRCE could parse loop
structure, it could only remove the range check against `length` and simply
ignored the check against `size`.
In this patch we remove this obsolete check. It will allow IRCE to pick comparison
against `size` as a potential range check and then let Range Intersection logic
decide whether it is OK to eliminate it or not.
Differential Revision: https://reviews.llvm.org/D45362
Reviewed By: samparker
llvm-svn: 329547
There are a pair of folds that try to merge fneg into fsub
with an intervening cast, but as shown in the FIXME tests,
they can create extra instructions.
llvm-svn: 329501
D45344 is proposing to remove the use restriction that made the calloc
transform safe, but it doesn't currently address the problematic example
given inD16337. Add a test to make sure that doesn't break.
llvm-svn: 329412
Summary:
Fixing an issue where initializations of globals where constructors use
casts were silently translated to 0-initialization.
Reviewers: davidxl, evgeny777
Reviewed By: evgeny777
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45198
llvm-svn: 329409
Inserting instrumentation between a musttail call and ret instruction
would create invalid IR. Instead, treat musttail calls as function
exits.
llvm-svn: 329385
This restores what was lost with rL73243 but without
re-introducing the bug that was present in the old code.
Note that we already have these transforms if the ops are
marked 'fast' (and I assume that's happening somewhere in
the code added with rL170471), but we clearly don't need
all of 'fast' for these transforms.
llvm-svn: 329362
A fold for this pattern was removed at rL73243 to fix PR4374:
https://bugs.llvm.org/show_bug.cgi?id=4374
...and apparently there were no tests that went with that fold.
llvm-svn: 329360
Summary:
More tests for D45108:
* One use tests
* allow shift to be a variable, too
Reviewers: spatel, craig.topper
Reviewed By: spatel
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45329
llvm-svn: 329348
Summary:
This is a fix to PR37005.
Essentially, rL328539 ([InstCombine] reassociate loop invariant GEP chains to enable LICM) contains a bug
whereby it will convert:
%src = getelementptr inbounds i8, i8* %base, <2 x i64> %val
%res = getelementptr inbounds i8, <2 x i8*> %src, i64 %val2
into:
%src = getelementptr inbounds i8, i8* %base, i64 %val2
%res = getelementptr inbounds i8, <2 x i8*> %src, <2 x i64> %val
By swapping the index operands if the GEPs are in a loop, and %val is loop variant while %val2
is loop invariant.
This fix recreates new GEP instructions if the index operand swap would result in the type
of %src changing from vector to scalar, or vice versa.
Reviewers: sebpop, spatel
Reviewed By: sebpop
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45287
llvm-svn: 329331
There used to be a fold that would handle this case more generally,
but it was removed at rL73243 to fix PR4374:
https://bugs.llvm.org/show_bug.cgi?id=4374
llvm-svn: 329322
Using cstfp_pred_ty in the definition allows us to match vectors with undef elements.
This replicates the change for m_Not from D44076 / rL326823 and continues
towards making all pattern matchers allow undef elements in vectors.
llvm-svn: 329303
LoopInterchange relies on LoopInfo being up-to-date, so we should
preserve it after interchanging. This patch updates restructureLoops to
move the BBs of the interchanged loops to the right place.
Reviewers: davide, efriedma, karthikthecool, mcrosier
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D45278
llvm-svn: 329264
Summary:
If the callsite is inside landing pad, do not perform callsite splitting.
Callsite splitting uses utility function llvm::DuplicateInstructionsInSplitBetween, which eventually calls llvm::SplitEdge. llvm::SplitEdge calls llvm::SplitCriticalEdge with an assumption that the function returns nullptr only when the target edge is not a critical edge (and further assumes that if the return value was not nullptr, the predecessor of the original target edge always has a single successor because critical edge splitting was successful). However, this assumtion is not true because SplitCriticalEdge returns nullptr if the destination block is a landing pad. This invalid assumption results assertion failure.
Fundamental solution might be fixing llvm::SplitEdge to not to rely on the invalid assumption. However, it'll involve a lot of work because current API assumes that llvm::SplitEdge never fails. Instead, this patch makes callsite splitting to not to attempt splitting if the callsite is in a landing pad.
Attached test case will crash with assertion failure without the fix.
Reviewers: fhahn, junbuml, dberlin
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45130
llvm-svn: 329250
Summary: @llvm.icall.branch.funnel is musttail with variable number of
arguments. After inlining current backend can't separate call targets from call
arguments.
Reviewers: pcc
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D45116
llvm-svn: 329235
Summary:
Clang's __builtin_operator_new/delete was recently taught about the aligned allocation overloads (r328134). This patch makes LLVM aware of them as well.
This allows the compiler to perform certain optimizations including eliding new/delete calls.
Reviewers: rsmith, majnemer, dblaikie, vsk, bkramer
Reviewed By: bkramer
Subscribers: ckennelly, llvm-commits
Differential Revision: https://reviews.llvm.org/D44769
llvm-svn: 329218
Summary:
Clang's __builtin_operator_new/delete was recently taught about the aligned allocation overloads (r328134). This patch makes LLVM aware of them as well.
This allows the compiler to perform certain optimizations including eliding new/delete calls.
Reviewers: rsmith, majnemer, dblaikie, vsk, bkramer
Reviewed By: bkramer
Subscribers: ckennelly, llvm-commits
Differential Revision: https://reviews.llvm.org/D44769
llvm-svn: 329215
Fixes cases like the new test @nonuniform. In that test, %cc itself
is a uniform value; however, when reading it after the end of the loop in
basic block %if, its value is effectively non-uniform, so the branch is
non-uniform.
This problem was encountered in
https://bugs.freedesktop.org/show_bug.cgi?id=103743; however, this change
in itself is not sufficient to fix that bug, as there is another issue
in the AMDGPU backend.
As discovered after committing an earlier version of this change, this
exposes a subtle interaction between this pass and DivergenceAnalysis:
since we remove and re-create branch instructions, we can no longer rely
on DivergenceAnalysis for branches in subregions that were already
processed by the pass.
Explicitly remove branch instructions from DivergenceAnalysis to
avoid dangling pointers as a matter of defensive programming, and
change how we detect non-uniform subregions.
Change-Id: I32bbffece4a32f686fab54964dae1a5dd72949d4
Differential Revision: https://reviews.llvm.org/D43743
llvm-svn: 329165
This patch teaches SCEV how to prove implications for SCEVUnknown nodes that are Phis.
If we need to prove `Pred` for `LHS, RHS`, and `LHS` is a Phi with possible incoming values
`L1, L2, ..., LN`, then if we prove `Pred` for `(L1, RHS), (L2, RHS), ..., (LN, RHS)` then we can also
prove it for `(LHS, RHS)`. If both `LHS` and `RHS` are Phis from the same block, it is sufficient
to prove the predicate for values that come from the same predecessor block.
The typical case that it handles is that we sometimes need to prove that `Phi(Len, Len - 1) >= 0`
given that `Len > 0`. The new logic was added to `isImpliedViaOperations` and only uses it and
non-recursive reasoning to prove the facts we need, so it should not hurt compile time a lot.
Differential Revision: https://reviews.llvm.org/D44001
Reviewed By: anna
llvm-svn: 329150
Summary:
Currently merge conditional stores can't handle cases where PostBB (the block we need to move the store to) has more than 2 predecessors.
This patch removes that restriction by creating a new block with only the 2 predecessors we care about and an unconditional branch to the original block. This provides a place to put the store.
Reviewers: efriedma, jmolloy, ABataev
Reviewed By: efriedma
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D39760
llvm-svn: 329142
The tests marked with 'FIXME' require loosening the check
in SimplifyAssociativeOrCommutative() to optimize completely;
that's still checking isFast() in Instruction::isAssociative().
llvm-svn: 329121
Summary:
If an alloca need to be stored in the coroutine frame and it has an alignment specified and the alignment does not match the natural alignment of the alloca type. Insert appropriate padding into the coroutine frame to make sure that it gets requested alignment.
For example for a packet type (which natural alignment is 1), but alloca alignment is 8, we may need to insert a padding field with required number of bytes to make sure it is properly aligned.
```
%PackedStruct = type <{ i64 }>
...
%data = alloca %PackedStruct, align 8
```
If the previous field in the coroutine frame had alignment 2, we would have [6 x i8] inserted before %PackedStruct in the coroutine frame:
```
%f.Frame = type { ..., i16, [6 x i8], %PackedStruct }
```
Reviewers: rnk, lewissbaker, modocache
Reviewed By: modocache
Subscribers: EricWF, llvm-commits
Differential Revision: https://reviews.llvm.org/D45221
llvm-svn: 329112
It also updates test/Transforms/LoopInterchange/call-instructions.ll
to use accesses where we can prove dependence after D35430.
Reviewers: sebpop, karthikthecool, blitz.opensource
Reviewed By: sebpop
Differential Revision: https://reviews.llvm.org/D45206
llvm-svn: 329111
Summary:
Folding patterns like:
%vec = shufflevector <4 x i8> %insvec, <4 x i8> undef, <4 x i32> zeroinitializer
%cast = bitcast <4 x i8> %vec to i32
%cond = icmp eq i32 %cast, 0
into:
%ext = extractelement <4 x i8> %insvec, i32 0
%cond = icmp eq i32 %ext, 0
Combined with existing rules, this allows us to fold patterns like:
%insvec = insertelement <4 x i8> undef, i8 %val, i32 0
%vec = shufflevector <4 x i8> %insvec, <4 x i8> undef, <4 x i32> zeroinitializer
%cast = bitcast <4 x i8> %vec to i32
%cond = icmp eq i32 %cast, 0
into:
%cond = icmp eq i8 %val, 0
When we construct a splat vector via a shuffle, and bitcast the vector into an integer type for comparison against an integer constant. Then we can simplify the the comparison to compare the splatted value against the integer constant.
Reviewers: spatel, anna, mkazantsev
Reviewed By: spatel
Subscribers: efriedma, rengolin, llvm-commits
Differential Revision: https://reviews.llvm.org/D44997
llvm-svn: 329087
Summary:
If the load/extractelement/extractvalue instructions are not originally
consecutive, the SLP vectorizer is unable to vectorize them. Patch
allows reordering of such instructions.
Patch does not support reordering of the repeated instruction, this must
be handled in the separate patch.
Reviewers: RKSimon, spatel, hfinkel, mkuper, Ayal, ashahid
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D43776
llvm-svn: 329085
I have taken the opportunity to simplify some tests slightly and move
parts around.
It also brings back a few IR checks for interchangable loops.
Reviewers: karthikthecool, sebpop, grosser
Reviewed By: sebpop
Differential Revision: https://reviews.llvm.org/D45207
llvm-svn: 329081
For Hexagon, peeling loops with small runtime trip count is beneficial for our
benchmarks. We set PeelCount in HexagonTargetInfo.cpp and we use PeelCount set
by the target for computing the desired peel count.
Differential Revision: https://reviews.llvm.org/D44880
llvm-svn: 329042
We use two approaches for determining the minimum bitwidth.
* Demanded bits
* Value tracking
If demanded bits doesn't result in a narrower type, we then try value tracking.
We need this if we want to root SLP trees with the indices of getelementptr
instructions since all the bits of the indices are demanded.
But there is a missing piece though. We need to be able to distinguish "demanded
and shrinkable" from "demanded and not shrinkable". For example, the bits of %i
in
%i = sext i32 %e1 to i64
%gep = getelementptr inbounds i64, i64* %p, i64 %i
are demanded, but we can shrink %i's type to i32 because it won't change the
result of the getelementptr. On the other hand, in
%tmp15 = sext i32 %tmp14 to i64
%tmp16 = insertvalue { i64, i64 } undef, i64 %tmp15, 0
it doesn't make sense to shrink %tmp15 and we can skip the value tracking.
Ideas are from Matthew Simpson!
Differential Revision: https://reviews.llvm.org/D44868
llvm-svn: 329035
Summary:
When attempting to split a coroutine with 'hidden' visibility (for
example, a C++ coroutine that is inlined when compiled with the option
'-fvisibility-inlines-hidden'), LLVM would hit an assertion in
include/llvm/IR/GlobalValue.h:240: "local linkage requires default
visibility". The issue is that the visibility is copied from the source
of the function split in the `CloneFunctionInto` function, but the linkage
is not. To fix, create the new function first with external linkage,
then copy the linkage from the original function *after* `CloneFunctionInto`
is called.
Since `GlobalValue::setLinkage` in turn calls `maybeSetDsoLocal`, the
explicit call to `setDSOLocal` can be removed in CoroSplit.cpp.
Test Plan: check-llvm
Reviewers: GorNishanov, lewissbaker, EricWF, majnemer, rnk
Reviewed By: rnk
Subscribers: llvm-commits, eric_niebler
Differential Revision: https://reviews.llvm.org/D44185
llvm-svn: 329033
Summary:
The cast simplifications that instcombine does here do not make any
attempt to obey the verifier rules for musttail calls. Therefore we have
to disable them.
Reviewers: efriedma, majnemer, pcc
Subscribers: hiraditya, llvm-commits
Differential Revision: https://reviews.llvm.org/D45186
llvm-svn: 329027
Some Function level metadatas, such as function entry count, are not cloned in
DeadArgumentElim. This happens a lot in lto/thinlto because of DeadArgumentElim
after internalization.
This patch clones the metadatas in the original function to the new function.
Differential Revision: https://reviews.llvm.org/D44127
llvm-svn: 328991
Summary:
A recent addition to Coroutines TS (https://wg21.link/p0913) adds a pre-defined coroutine noop_coroutine that does nothing.
To implement this feature, we implemented an llvm.coro.noop intrinsic that returns a coroutine handle to a coroutine that does nothing when resumed or destroyed.
Reviewers: EricWF, modocache, rnk, lewissbaker
Reviewed By: modocache
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D45114
llvm-svn: 328986
Summary:
If the load/extractelement/extractvalue instructions are not originally
consecutive, the SLP vectorizer is unable to vectorize them. Patch
allows reordering of such instructions.
Reviewers: RKSimon, spatel, hfinkel, mkuper, Ayal, ashahid
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D43776
llvm-svn: 328980
Summary:
Adds -import-cutoff=N which will stop importing during the thin link
after N imports. Default is -1 (no limit).
Reviewers: wmi
Subscribers: inglorion, llvm-commits
Differential Revision: https://reviews.llvm.org/D45127
llvm-svn: 328934
If a loop has a loop exiting latch, it can be profitable
to rotate the loop if it leads to the simplification of
a phi node. Perform rotation in these cases even if loop
rotate itself didnt simplify the loop to get there.
Differential Revision: https://reviews.llvm.org/D44199
llvm-svn: 328933
Summary:
Useful to selectively disable importing into specific modules for
debugging/triaging/workarounds.
Reviewers: eraman
Subscribers: inglorion, llvm-commits
Differential Revision: https://reviews.llvm.org/D45062
llvm-svn: 328909
For Hexagon, peeling loops with small runtime trip count is beneficial for our
benchmarks. We set PeelCount in HexagonTargetInfo.cpp and we use PeelCount set
by the target for computing the desired peel count.
Differential Revision: https://reviews.llvm.org/D44880
llvm-svn: 328854
In r312664 (D36404), JumpThreading stopped threading edges into
loop headers. Unfortunately, I observed a significant performance
regression as a result of this change. Upon further investigation,
the problematic pattern looked something like this (after
many high level optimizations):
while (true) {
bool cond = ...;
if (!cond) {
<body>
}
if (cond)
break;
}
Now, naturally we want jump threading to essentially eliminate the
second if check and hook up the edges appropriately. However, the
above mentioned change, prevented it from doing this because it would
have to thread an edge into the loop header.
Upon further investigation, what is happening is that since both branches
are threadable, JumpThreading picks one of them at arbitrarily. In my
case, because of the way that the IR ended up, it tended to pick
the one to the loop header, bailing out immediately after. However,
if it had picked the one to the exit block, everything would have
worked out fine (because the only remaining branch would then be folded,
not thraded which is acceptable).
Thus, to fix this problem, we can simply eliminate loop headers from
consideration as possible threading targets earlier, to make sure that
if there are multiple eligible branches, we can still thread one of
the ones that don't target a loop header.
Patch by Keno Fischer!
Differential Revision: https://reviews.llvm.org/D42260
llvm-svn: 328798
Fixed counter/weight overflow that leads to an assertion. Also fixed the help
string for pgo-emit-branch-prob option.
Differential Revision: https://reviews.llvm.org/D44809
llvm-svn: 328653
As a follow-up to r328480, this updates the logic for the decreasing
safety checks in a similar manner:
- CanBeMax is replaced by CannotBeMaxInLoop which queries
isLoopEntryGuardedByCond on the maximum value.
- SumCanReachMin is replaced by isSafeDecreasingBound which includes
some logic from parseLoopStructure and, again, has been updated to
use isLoopEntryGuardedByCond on the given bounds.
Differential Revision: https://reviews.llvm.org/D44776
llvm-svn: 328613
Currently, `getExact` fails if it sees two exit counts in different blocks. There is
no solid reason to do so, given that we only calculate exact non-taken count
for exiting blocks that dominate latch. Using this fact, we can simply take min
out of all exits of all blocks to get the exact taken count.
This patch makes the calculation more optimistic with enforcing our assumption
with asserts. It allows us to calculate exact backedge taken count in trivial loops
like
for (int i = 0; i < 100; i++) {
if (i > 50) break;
. . .
}
Differential Revision: https://reviews.llvm.org/D44676
Reviewed By: fhahn
llvm-svn: 328611
This patch teaches `computeConstantDifference` handle calculation of constant
difference between `(X + C1)` and `(X + C2)` which is `(C2 - C1)`.
Differential Revision: https://reviews.llvm.org/D43759
Reviewed By: anna
llvm-svn: 328609
MemorySSAUpdater::getPreviousDefRecursive is a recursive algorithm, for
each block, it computes the previous definition for each predecessor,
then takes those definitions and combines them. But currently it doesn't
remember results which it already computed; this means it can visit the
same block multiple times, which adds up to exponential time overall.
To fix this, this patch adds a cache. If we computed the result for a
block already, we don't need to visit it again because we'll come up
with the same result. Well, unless we RAUW a MemoryPHI; in that case,
the TrackingVH will be updated automatically.
This matches the original source paper for this algorithm.
The testcase isn't really a test for the bug, but it adds coverage for
the case where tryRemoveTrivialPhi erases an existing PHI node. (It's
hard to write a good regression test for a performance issue.)
Differential Revision: https://reviews.llvm.org/D44715
llvm-svn: 328577
This change brings performance of zlib up by 10%. The example below is from a
hot loop in longest_match() from zlib.
do.body:
%cur_match.addr.0 = phi i32 [ %cur_match, %entry ], [ %2, %do.cond ]
%idx.ext = zext i32 %cur_match.addr.0 to i64
%add.ptr = getelementptr inbounds i8, i8* %win, i64 %idx.ext
%add.ptr2 = getelementptr inbounds i8, i8* %add.ptr, i64 %idx.ext1
%add.ptr3 = getelementptr inbounds i8, i8* %add.ptr2, i64 -1
In this example %idx.ext1 is a loop invariant. It will be moved above the use of
loop induction variable %idx.ext such that it can be hoisted out of the loop by
LICM. The operands that have dependences carried by the loop will be sinked down
in the GEP chain. This patch will produce the following output:
do.body:
%cur_match.addr.0 = phi i32 [ %cur_match, %entry ], [ %2, %do.cond ]
%idx.ext = zext i32 %cur_match.addr.0 to i64
%add.ptr = getelementptr inbounds i8, i8* %win, i64 %idx.ext1
%add.ptr2 = getelementptr inbounds i8, i8* %add.ptr, i64 -1
%add.ptr3 = getelementptr inbounds i8, i8* %add.ptr2, i64 %idx.ext
llvm-svn: 328539
This replaces a large chunk of code that was looking for compound
patterns that include these sub-patterns. Existing tests ensure that
all of the previous examples are still folded as expected.
We still need to loosen the FMF check.
llvm-svn: 328502
Current logic of loop SCEV invalidation in Loop Unroller implicitly relies on
fact that exit count of outer loops cannot rely on exiting blocks of
inner loops, which is true in current implementation of backedge taken count
calculation but is wrong in general. As result, when we only forget the loop that
we have just unrolled, we may still have cached data for its outer loops (in particular,
exit counts) which keeps references on blocks of inner loop that could have been
changed or even deleted.
The attached test demonstrates a situaton when after unrolling of innermost loop
the outermost loop contains a dangling pointer on non-existant block. The problem
shows up when we apply patch https://reviews.llvm.org/D44677 that makes SCEV
smarter about exit count calculation. I am not sure if the bug exists without this patch,
it appears that now it is accidentally correct just because in practice exact backedge
taken count for outer loops with complex control flow inside is never calculated.
But when SCEV learns to do so, this problem shows up.
This patch replaces existing logic of SCEV loop invalidation with a correct one, which
happens to be invalidation of outermost loop (which also leads to invalidation of all
loops inside of it). It is the only way to ensure that no outer loop keeps dangling pointers
on removed blocks, or just outdated information that has changed after unrolling.
Differential Revision: https://reviews.llvm.org/D44818
Reviewed By: samparker
llvm-svn: 328483
CanBeMin is currently used which will report true for any unknown
values, but often a check is performed outside the loop which covers
this situation:
for (int i = 0; i < N; ++i)
...
if (N > 0)
for (int i = 0; i < N; ++i)
...
So I've add 'LoopGuardedAgainstMin' which reports whether N is
greater than the minimum value which then allows loop with a variable
loop count to be optimised. I've also moved the increasing bound
checking into its own function and replaced SumCanReachMax is another
isLoopEntryGuardedByCond function.
llvm-svn: 328480
We have thorough coverage of predicates and scalar types,
so we just need a sampling of vector tests to show that
things are working or not with vectors types.
llvm-svn: 328449
Summary:
This was motivated by absence of PrunEH functionality in new PM.
It was decided that a proper way to do PruneEH is to add NoUnwind inference
into PostOrderFunctionAttrs and then perform normal SimplifyCFG on top.
This change generalizes attribute handling implemented for (a removal of)
Convergent attribute, by introducing a generic builder-like class
AttributeInferer
It registers all the attribute inference requests, storing per-attribute
predicates into a vector, and then goes through an SCC Node, scanning all
the instructions for not breaking attribute assumptions.
The main idea is that as soon all the instructions from all the functions
of SCC Node conform to attribute assumptions then we are free to infer
the attribute as set for all the functions of SCC Node.
It handles two distinct cases of attributes:
- those that might break due to derefinement of the function code
for these attributes we are allowed to apply inference only if all the
functions are "exact definitions". Example - NoUnwind.
- those that do not care about derefinement
for these attributes we are allowed to apply inference as soon as we see
any function definition. Example - removal of Convergent attribute.
Also in this commit:
* Converted all the FunctionAttrs tests to use FileCheck and added new-PM
invocations to them
* FunctionAttrs/convergent.ll test demonstrates a difference in behavior between
new and old PM implementations. Marked with FIXME.
* PruneEH tests were converted to new-PM as well, using function-attrs+simplify-cfg
combo as intended
* some of "other" tests were updated since function-attrs now infers 'nounwind'
even for old PM pipeline
* -disable-nounwind-inference hidden option added as a possible workaround for a supposedly
rare case when nounwind being inferred by default presents a problem
Reviewers: chandlerc, jlebar
Reviewed By: jlebar
Subscribers: eraman, llvm-commits
Differential Revision: https://reviews.llvm.org/D44415
llvm-svn: 328377
When building the SLP tree, we look for reuse among the vectorized tree
entries. However, each gather sequence is represented by a unique tree entry,
even though the sequence may be identical to another one. This means, for
example, that a gather sequence with two uses will be counted twice when
computing the cost of the tree. We should only count the cost of the definition
of a gather sequence rather than its uses. During code generation, the
redundant gather sequences are emitted, but we optimize them away with CSE. So
it looks like this problem just affects the cost model.
Differential Revision: https://reviews.llvm.org/D44742
llvm-svn: 328316
For comparisons with parameters, we can use the ParamState lattice
elements which also provide constant range information. This improves
the code for PR33253 further and gets us closer to use
ValueLatticeElement for all values.
Also, as we are using the range information in the solver directly, we
do not need tryToReplaceWithConstantRange afterwards anymore.
Reviewers: dberlin, mssimpso, davide, efriedma
Reviewed By: mssimpso
Differential Revision: https://reviews.llvm.org/D43762
llvm-svn: 328307
Loop peeling also has an impact on the induction variables, so we should
benefit from induction variable simplification after peeling too.
Reviewers: sanjoy, bogner, mzolotukhin, efriedma
Reviewed By: efriedma
Differential Revision: https://reviews.llvm.org/D43878
llvm-svn: 328301
Summary:
Revert r325687 workaround for PR36032 since
a fix was committed in r326154.
Reviewers: sbaranga
Differential Revision: http://reviews.llvm.org/D44768
From: Evgeny Stupachenko <evstupac@gmail.com>
<evgeny.v.stupachenko@intel.com>
llvm-svn: 328257
Summary:
When building with libFuzzer, converting control flow to selects or
obscuring the original operands of CMPs reduces the effectiveness of
libFuzzer's heuristics.
This patch provides an attribute to disable or modify certain optimizations
for optimal fuzzing signal.
Provides a less aggressive alternative to https://reviews.llvm.org/D44057.
Reviewers: vitalybuka, davide, arsenm, hfinkel
Reviewed By: vitalybuka
Subscribers: junbuml, mehdi_amini, wdng, javed.absar, hiraditya, llvm-commits, kcc
Differential Revision: https://reviews.llvm.org/D44232
llvm-svn: 328214
Summary:
LoopPredication is not profitable when the loop is known to always exit
through some block other than the latch block.
A coarse grained latch check can cause loop predication to predicate the
loop, and unconditionally deoptimize.
However, without predicating the loop, the guard may never fail within the
loop during the dynamic execution because the non-latch loop termination
condition exits the loop before the latch condition causes the loop to
exit.
We teach LP about this using BranchProfileInfo pass.
Reviewers: apilipenko, skatkov, mkazantsev, reames
Reviewed by: skatkov
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D44667
llvm-svn: 328210
There are at least 3 problems:
1. We're distributing across large patterns, but fail to do that for the minimal patterns.
2. We're not checking uses, so we may create more instructions than we eliminate.
3. We should be able to do these transforms with less than full 'fast' fmuls.
llvm-svn: 328152
Summary:
This change is part of step five in the series of changes to remove alignment argument from
memcpy/memmove/memset in favour of alignment attributes. In particular, this changes the
MemCpyOpt pass to cease using:
1) The old getAlignment() API of MemoryIntrinsic in favour of getting source & dest specific
alignments through the new API.
2) The old IRBuilder CreateMemCpy/CreateMemMove single-alignment APIs in favour of the new
API that allows setting source and destination alignments independently.
We also add a few tests to fill gaps in the testing of this pass.
Steps:
Step 1) Remove alignment parameter and create alignment parameter attributes for
memcpy/memmove/memset. ( rL322965, rC322964, rL322963 )
Step 2) Expand the IRBuilder API to allow creation of memcpy/memmove with differing
source and dest alignments. ( rL323597 )
Step 3) Update Clang to use the new IRBuilder API. ( rC323617 )
Step 4) Update Polly to use the new IRBuilder API. ( rL323618 )
Step 5) Update LLVM passes that create memcpy/memmove calls to use the new IRBuilder API,
and those that use use MemIntrinsicInst::[get|set]Alignment() to use [get|set]DestAlignment()
and [get|set]SourceAlignment() instead. ( rL323886, rL323891, rL324148, rL324273, rL324278,
rL324384, rL324395, rL324402, rL324626, rL324642, rL324653, rL324654, rL324773, rL324774,
rL324781, rL324784, rL324955, rL324960, rL325816, rL327398, rL327421 )
Step 6) Remove the single-alignment IRBuilder API for memcpy/memmove, and the
MemIntrinsicInst::[get|set]Alignment() methods.
Reference
http://lists.llvm.org/pipermail/llvm-dev/2015-August/089384.htmlhttp://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151109/312083.html
llvm-svn: 328097
Summary:
If the operands of a udiv/urem can be proved to fit within a smaller
power-of-two-sized type, reduce the width of the udiv/urem.
Backed out for causing performance regressions. Re-landing
because we've determined that these regressions were noise.
Original Differential Revision: https://reviews.llvm.org/D44102
llvm-svn: 328096
The inline assembly generated for the ARC autorelease elision marker
must have a funclet token if it's emitted inside a funclet, otherwise
the inline assembly (and all subsequent code in the funclet) will be
marked unreachable by WinEHPrepare.
Note that this only applies for the non-O0 case, since at O0, clang
emits the autorelease elision marker itself rather than deferring to the
backend. The fix for clang is handled in a separate change.
Differential Revision: https://reviews.llvm.org/D44641
llvm-svn: 328042
Summary: Fix a bug in entry block shuffled to middle of the chain.
Reviewers: davide, courbet
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D44642
llvm-svn: 327971
Summary:
It turned out to be error-prone to expect the callers to handle that - better to
leave the decision to this routine and make the required data to be explicitly
passed to the function.
This handles the case that was missed in the r322473 and fixes the assert
mentioned in PR36524.
Reviewers: dorit, mssimpso, Ayal, dcaballe
Reviewed By: dcaballe
Subscribers: Ka-Ka, hiraditya, dneilson, hsaito, llvm-commits
Differential Revision: https://reviews.llvm.org/D43812
llvm-svn: 327960
This is complicated by -0.0 and nan. This is based on the DAG patterns
as shown in D44091. I'm hoping that we can just remove those DAG folds
and always rely on IR canonicalization to handle the matching to fabs.
We would still need to delete the broken code from DAGCombiner to fix
PR36600:
https://bugs.llvm.org/show_bug.cgi?id=36600
Differential Revision: https://reviews.llvm.org/D44550
llvm-svn: 327858
This is re-land of https://reviews.llvm.org/rL327362 with a fix
and regression test.
The crash was due to it is possible that for found MDL loop,
LHS or RHS may contain an invariant unknown SCEV which
does not dominate the MDL. Please see regression
test for an example.
Reviewers: sanjoy, mkazantsev, reames
Reviewed By: mkazantsev
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D44553
llvm-svn: 327822
LICM deletes trivially dead instructions which it won't attempt to sink.
Attempt to salvage debug values which reference these instructions.
llvm-svn: 327800
As shown in the code comment, we don't need all of 'fast',
but we do need reassoc + nsz + nnan.
Differential Revision: https://reviews.llvm.org/D43765
llvm-svn: 327796
This builds on the work from https://reviews.llvm.org/D44287. It turned out supporting fcmp was much easier than I realized, so let's do that now.
As an aside, our -O3 handling of a floating point IVs leaves a lot to be desired. We do convert the float IV to an integer IV, but do so late enough that many other optimizations are missed (e.g. we don't vectorize).
Differential Revision: https://reviews.llvm.org/D44542
llvm-svn: 327722
JumpThreading iterates over F until the IR quiesces. Transforming
unreachable BBs increases compile time and it is also possible to
never stabilize causing JumpThreading to hang. An older attempt at
fixing this problem was D3991 where removeUnreachableBlocks(F)
was called before JumpThreading began. This has a few drawbacks:
* expensive - the routine attempts to fix up the IR to identify
additional BBs that can be removed along with unreachable BBs.
* aggressive - does not identify and preserve the shape of the IR.
At a minimum it does not preserve loop hierarchies.
* invasive - altering reachable blocks it may disrupt IR shapes
that could have otherwise been JumpThreaded.
This patch avoids removeUnreachableBlocks(F) and instead tracks
unreachable BBs in a SmallPtrSet using DominatorTree to validate the
initial state of all BBs. We then rely on subsequent passes to identify
and remove these unreachable blocks from F.
Reviewers: dberlin, sebpop, kuhar, dinesh.d
Reviewed by: sebpop, kuhar
Subscribers: hiraditya, uabelho, llvm-commits
Differential Revision: https://reviews.llvm.org/D44177
llvm-svn: 327713
This patch provides an implementation of getArithmeticReductionCost for
AArch64. We can specialize the cost of add reductions since they are computed
using the 'addv' instruction.
Differential Revision: https://reviews.llvm.org/D44490
llvm-svn: 327702
If the loop body contains conditions of the form IndVar < #constant, we
can remove the checks by peeling off #constant iterations.
This improves codegen for PR34364.
Reviewers: mkuper, mkazantsev, efriedma
Reviewed By: mkazantsev
Differential Revision: https://reviews.llvm.org/D43876
llvm-svn: 327671
It is common to have conditional exits within a loop which are known not to be taken on some iterations, but not necessarily all. This patches extends our reasoning around guaranteed to execute (used when establishing whether it's safe to dereference a location from the preheader) to handle the case where an exit is known not to be taken on the first iteration and the instruction of interest *is* known to be taken on the first iteration.
This case comes up in two major ways:
* If we have a range check which we've been unable to eliminate, we frequently know that it doesn't fail on the first iteration.
* Pass ordering. We may have a check which will be eliminated through some sequence of other passes, but depending on the exact pass sequence we might never actually do so or we might miss other optimizations from passes run before the check is finally eliminated.
The initial version (here) is implemented via InstSimplify. At the moment, it catches a few cases, but misses a lot too. I added test cases for missing cases in InstSimplify which I'll follow up on separately. Longer term, we should probably wire SCEV through to here to get much smarter loop aware simplification of the first iteration predicate.
Differential Revision: https://reviews.llvm.org/D44287
llvm-svn: 327664
If we've already established an invariant scope with an earlier generation, we don't want to hide it in the scoped hash table with one with a later generation. I noticed this when working on the invariant-load handling, but it also applies to the invariant.start case as well.
Without this change, my previous patch for invariant-load regresses some cases, so I'm pushing this without waiting for review. This is why you don't make last minute tweaks to patches to catch "obvious cases" after it's already been reviewed. Bad Philip!
llvm-svn: 327655
This is a follow up to https://reviews.llvm.org/D43716 which rewrites the invariant load handling using the new infrastructure. It's slightly more powerful, but only in somewhat minor ways for the moment. It's not clear that DSE of stores to invariant locations is actually interesting since why would your IR have such a construct to start with?
Note: The submitted version is slightly different than the reviewed one. I realized the scope could start for an invariant load which was proven redundant and removed. Added a test case to illustrate that as well.
Differential Revision: https://reviews.llvm.org/D44497
llvm-svn: 327646
This matcher implementation appears to be slightly more efficient than
the generic constant check that it is replacing because every use was
for matching FP patterns, but the previous code would check int and
pointer type nulls too.
llvm-svn: 327627
From the LangRef definition for frem:
"The value produced is the floating-point remainder of the two operands.
This is the same output as a libm ‘fmod‘ function, but without any
possibility of setting errno. The remainder has the same sign as the
dividend. This instruction is assumed to execute in the default
floating-point environment."
llvm-svn: 327626
When hoisting common code from the "then" and "else" branches of a condition
to before the "if", the HoistThenElseCodeToIf routine will attempt to merge
the debug location associated with the two original copies of the hoisted
instruction.
This is a problem in the special case where the hoisted instruction is a
debug info intrinsic, since for those the debug location is considered
part of the intrinsic and attempting to modify it may resut in invalid
IR. This is the underlying cause of PR36410.
This patch fixes the problem by handling debug info intrinsics specially:
instead of hoisting one copy and merging the two locations, the code now
simply hoists both copies, each with its original location intact. Note
that this is still only done in the case where both original copies are
otherwise (i.e. apart from location metadata) identical.
Reviewed By: aprantl
Differential Revision: https://reviews.llvm.org/D44312
llvm-svn: 327622
There are two nontrivial details here:
* Loop structure update interface is quite different with new pass manager,
so the code to add new loops was factored out
* BranchProbabilityInfo is not a loop analysis, so it can not be just getResult'ed from
within the loop pass. It cant even be queried through getCachedResult as LoopCanonicalization
sequence (e.g. LoopSimplify) might invalidate BPI results.
Complete solution for BPI will likely take some time to discuss and figure out,
so for now this was partially solved by making BPI optional in IRCE
(skipping a couple of profitability checks if it is absent).
Most of the IRCE tests got their corresponding new-pass-manager variant enabled.
Only two of them depend on BPI, both marked with TODO, to be turned on when BPI
starts being available for loop passes.
Reviewers: chandlerc, mkazantsev, sanjoy, asbirlea
Reviewed By: mkazantsev
Differential Revision: https://reviews.llvm.org/D43795
llvm-svn: 327619
Summary:
Before this patch call graph is like this in the LoopUnrollPass:
tryToUnrollLoop
ApproximateLoopSize
collectEphemeralValues
/* Use collected ephemeral values */
computeUnrollCount
analyzeLoopUnrollCost
/* Bail out from the analysis if loop contains CallInst */
This patch moves collection of the ephemeral values to the tryToUnrollLoop
function and passes the collected values into both ApproximateLoopsize (as
before) and additionally starts using them in analyzeLoopUnrollCost:
tryToUnrollLoop
collectEphemeralValues
ApproximateLoopSize(EphValues)
/* Use EphValues */
computeUnrollCount(EphValues)
analyzeLoopUnrollCost(EphValues)
/* Ignore ephemeral values - they don't contribute to the final cost */
/* Bail out from the analysis if loop contains CallInst */
Reviewers: mzolotukhin, evstupac, sanjoy
Reviewed By: evstupac
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D43931
llvm-svn: 327617
These should all be folded. The vector tests need to have
m_AnyZero updated to ignore undef elements, but we need to
be careful not to return the existing value in that case
and unintentionally propagate undef.
llvm-svn: 327585
If we have an invariant.start with no corresponding invariant.end, then the memory location becomes invariant indefinitely after the invariant.start. As a result, anything dominated by the start is guaranteed to see the value the memory location had when the invariant.start executed.
This patch adds an AvailableInvariants table which tracks the generation a particular memory location became invariant and then uses that information to allow value forwarding that would otherwise be disallowed by potentially aliasing stores. (Reminder: In EarlyCSE everything clobbers everything by default.)
This should be compatible with the MemorySSA variant, but design is generational. We can and should add first class support for invariant.start within MemorySSA at a later time. I took a quick look at doing so, but probably need some input from a MemorySSA expert.
Differential Revision: https://reviews.llvm.org/D43716
llvm-svn: 327577
As shown in:
https://bugs.llvm.org/show_bug.cgi?id=27151
...the existing fold could miscompile when X is NaN.
The fold was also dependent on 'ninf' but that's not necessary.
From IEEE-754 (with default rounding which we can assume for these opcodes):
"When the sum of two operands with opposite signs (or the difference of two
operands with like signs) is exactly zero, the sign of that sum (or difference)
shall be +0...However, x + x = x − (−x) retains the same sign as x even when
x is zero."
llvm-svn: 327575
Summary:
This pattern came up in PR36682 / D44390
https://bugs.llvm.org/show_bug.cgi?id=36682https://reviews.llvm.org/D44390https://godbolt.org/g/oKvT5H
Looking at the IR pattern in question, as per [[ https://github.com/rutgers-apl/alive-nj | alive-nj ]], for all the type combinations i checked
(input: `i16`, `i32`, `i64`; intermediate: `half`/`i16`, `float`/`i32`, `double`/`i64`)
for the following `icmp` comparisons the `uitofp`+`bitcast`+`icmp` can be evaluated to a boolean:
* `slt 0`
* `sgt -1`
I did not check vectors, but i'm guessing it's the same there.
{F5889242}
Thus all these cases are in the testcase (along with the vector variant with additional `undef` element in the middle).
There are no negative patterns here (unless alive-nj lied/is broken), all of these should be optimized.
Reviewers: spatel, majnemer, efriedma, arsenm
Reviewed By: spatel
Subscribers: wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D44421
llvm-svn: 327535
Summary:
This pattern came up in PR36682 / D44390
https://bugs.llvm.org/show_bug.cgi?id=36682https://reviews.llvm.org/D44390https://godbolt.org/g/oKvT5H
Looking at the IR pattern in question, as per [[ https://github.com/rutgers-apl/alive-nj | alive-nj ]], for all the type combinations i checked
(input: `i16`, `i32`, `i64`; intermediate: `half`/`i16`, `float`/`i32`, `double`/`i64`)
for the following `icmp` comparisons the `uitofp`+`bitcast` can be dropped:
* `eq 0`
* `ne 0`
I did not check vectors, but i'm guessing it's the same there.
{F5889189}
Thus all these cases are in the testcase (along with the vector variant with additional `undef` element in the middle).
There are no negative patterns here (unless alive-nj lied/is broken), all of these should be optimized.
Generated with
{F5889196}
Reviewers: spatel, majnemer, efriedma, arsenm
Reviewed By: spatel
Subscribers: wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D44416
llvm-svn: 327534
Summary:
It is possible for LVI to encounter instructions that are not in valid
SSA form and reference themselves. One example is the following:
%tmp4 = and i1 %tmp4, undef
Before this patch LVI would recurse until running out of stack memory
and crashed. This patch marks these self-referential instructions as
Overdefined and aborts analysis on the instruction.
Fixes https://bugs.llvm.org/show_bug.cgi?id=33357
Reviewers: craig.topper, anna, efriedma, dberlin, sebpop, kuhar
Reviewed by: dberlin
Subscribers: uabelho, spatel, a.elovikov, fhahn, eli.friedman, mzolotukhin, spop, evandro, davide, llvm-commits
Differential Revision: https://reviews.llvm.org/D34135
llvm-svn: 327432
This was supposed to be an NFC refactoring that will eventually allow
eliminating the isFast() predicate, but there's a rare possibility
that we would pessimize the code as shown in the test case because
we failed to check 'hasOneUse()' properly. This version also removes
an inefficiency of the old code; we would look for:
(X * C) * C1 --> X * (C * C1)
...but that pattern is always handled by
SimplifyAssociativeOrCommutative().
llvm-svn: 327404
Summary:
This change is part of step five in the series of changes to remove alignment argument from
memcpy/memmove/memset in favour of alignment attributes. In particular, this changes the
SROA pass to cease using the old getAlignment() & setAlignment() APIs of MemoryIntrinsic in
favour of getting source & dest specific alignments through the new API. This allows us
to enhance visitMemTransferInst to be more aggressive setting the alignment in memcpy
calls that it creates, as well as to only change the alignment of a memcpy/memmove
argument that it replaces.
Steps:
Step 1) Remove alignment parameter and create alignment parameter attributes for
memcpy/memmove/memset. ( rL322965, rC322964, rL322963 )
Step 2) Expand the IRBuilder API to allow creation of memcpy/memmove with differing
source and dest alignments. ( rL323597 )
Step 3) Update Clang to use the new IRBuilder API. ( rC323617 )
Step 4) Update Polly to use the new IRBuilder API. ( rL323618 )
Step 5) Update LLVM passes that create memcpy/memmove calls to use the new IRBuilder API,
and those that use use MemIntrinsicInst::[get|set]Alignment() to use [get|set]DestAlignment()
and [get|set]SourceAlignment() instead. ( rL323886, rL323891, rL324148, rL324273, rL324278,
rL324384, rL324395, rL324402, rL324626, rL324642, rL324653, rL324654, rL324773, rL324774,
rL324781, rL324784, rL324955, rL324960, rL325816 )
Step 6) Remove the single-alignment IRBuilder API for memcpy/memmove, and the
MemIntrinsicInst::[get|set]Alignment() methods.
Reference
http://lists.llvm.org/pipermail/llvm-dev/2015-August/089384.htmlhttp://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20151109/312083.html
Reviewers: chandlerc, bollu, efriedma
Reviewed By: efriedma
Subscribers: efriedma, eraman, llvm-commits
Differential Revision: https://reviews.llvm.org/D42974
llvm-svn: 327398
It is a revert of rL327362 which causes build bot failures with assert like
Assertion `isAvailableAtLoopEntry(RHS, L) && "RHS is not available at Loop Entry"' failed.
llvm-svn: 327363
IsKnownPredicate is updated to implement the following algorithm
proposed by @sanjoy and @mkazantsev :
isKnownPredicate(Pred, LHS, RHS) {
Collect set S all loops on which either LHS or RHS depend.
If S is non-empty
a. Let PD be the element of S which is dominated by all other elements of S
b. Let E(LHS) be value of LHS on entry of PD.
To get E(LHS), we should just take LHS and replace all AddRecs that
are attached to PD on with their entry values.
Define E(RHS) in the same way.
c. Let B(LHS) be value of L on backedge of PD.
To get B(LHS), we should just take LHS and replace all AddRecs that
are attached to PD on with their backedge values.
Define B(RHS) in the same way.
d. Note that E(LHS) and E(RHS) are automatically available on entry of PD,
so we can assert on that.
e. Return true if isLoopEntryGuardedByCond(Pred, E(LHS), E(RHS)) &&
isLoopBackedgeGuardedByCond(Pred, B(LHS), B(RHS))
Return true if Pred, L, R is known from ranges, splitting etc.
}
This is follow-up for https://reviews.llvm.org/D42417.
Reviewers: sanjoy, mkazantsev, reames
Reviewed By: sanjoy, mkazantsev
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D43507
llvm-svn: 327362
Summary:
This change fixes PR36483. The bug was originally introduced by a change
that marked non-prevailing symbols dead. This broke LowerTypeTests
handling of available_externally functions, which are non-prevailing.
LowerTypeTests uses liveness information to avoid emitting thunks for
unused functions.
Marking available_externally functions dead is incorrect, the functions
are used though the function definitions are not. This change keeps them
live, and lets the EliminateAvailableExternally/GlobalDCE passes remove
them later instead.
(Reland with a suspected fix for a unit test failure I haven't been able
to reproduce locally)
Reviewers: pcc, tejohnson
Reviewed By: tejohnson
Subscribers: grimar, mehdi_amini, inglorion, eraman, llvm-commits
Differential Revision: https://reviews.llvm.org/D43690
llvm-svn: 327360
In the case that the CallInst that is being moved has an associated
operand bundle which is a funclet, the move will construct an invalid
instruction. The new site will have a different token and needs to be
reassociated with the new instruction.
Unfortunately, there is no way to alter the bundle after the
construction of the instruction. Replace the call instruction cloning
with a custom helper to clone the instruction and reassociate the
funclet token.
llvm-svn: 327336
This is the FP equivalent of D42818. Use it for the few cases in InstSimplify
with -0.0 folds (that's the only current use of m_NegZero()).
Differential Revision: https://reviews.llvm.org/D43792
llvm-svn: 327307
Summary:
This pattern came up in PR36682:
https://bugs.llvm.org/show_bug.cgi?id=36682https://godbolt.org/g/LhuD9A
Tests for proposed fix in D44367.
Looking at the IR pattern in question, as per [[ https://github.com/rutgers-apl/alive-nj | alive-nj ]], for all the type combinations i checked
(input: `i16`, `i32`, `i64`; intermediate: `half`/`i16`, `float`/`i32`, `double`/`i64`)
for the following `icmp` comparisons the `sitofp`+`bitcast` can be dropped:
* `eq 0`
* `ne 0`
* `slt 0`
* `sle 0`
* `sge 0`
* `sgt 0`
* `slt 1`
* `sge 1`
* `sle -1`
* `sgt -1`
I did not check vectors, but i'm guessing it's the same there.
{F5887419}
Thus all these cases are in the testcase (along with the vector variant with additional `undef` element in the middle).
There are no negative patterns here (unless alive-nj lied/is broken), all of these should be optimized.
Generated with {F5887551}
Reviewers: spatel, majnemer, efriedma, arsenm
Reviewed By: spatel
Subscribers: nlopes, wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D44390
llvm-svn: 327301
Craig Topper