Currently available intrinsics are only ld2/st2, which don't support interleaving factor > 2.
This patch teaches the LV to use ld2/st2 recursively to support high
interleaving factors.
If the VectorizedTree still may generate poisonous value, but it is not
the original operand of the reduction op, need to check if Res still the
operand, to generate correct code.
Fixes#114905
As discussed in #94468, this causes switch lookup table entries which
are unreachable to be poison instead of filling them with a value from
one of the reachable cases.
---------
Co-authored-by: DianQK <dianqk@dianqk.net>
After #96127 landed, mshockwave reported that the pass was no longer
threading SPEC2006/perlbench.
After 96127 we started bailing out in `getStateDefMap` and rejecting the
transformation because one of the unpredictable values was coming from
inside the loop. There was no fundamental change in that function except
that we started calling `Loop->contains(IncomingBB)` instead of
`LoopBBs.count(IncomingBB)`. After some analysis I came to the
conclusion that even before 96127 we would reject the transformation if
we provided large enough limits on the path traversal (large enough so
that LoopBBs contained blocks corresponding to that unpredictable
value).
In this patch I changed `getStateDefMap` to not terminate early on
finding an unpredictable value, this is because
`getPathsFromStateDefMap`, later, actually has checks to ensure that the
final list of paths only have predictable values. As a result we can now
partially thread functions like `negative6` in the tests that have some
predictable paths.
This patch does not really have any compile-time impact on the test
suite without `-dfa-early-exit-heuristic=false` (early exit is enabled
by default).
Change-Id: Ie1633b370ed4a0eda8dea52650b40f6f66ef49a3
Adds cost estimation for the variants of the permutations of the scalar
values, used in gather nodes. Currently, SLP just unconditionally emits
shuffles for the reused buildvectors, but in some cases better to leave
them as buildvectors rather than shuffles, if the cost of such
buildvectors is better.
X86, AVX512, -O3+LTO
Metric: size..text
Program size..text
results results0 diff
test-suite :: External/SPEC/CINT2006/445.gobmk/445.gobmk.test 912998.00 913238.00 0.0%
test-suite :: MultiSource/Benchmarks/MiBench/consumer-lame/consumer-lame.test 203070.00 203102.00 0.0%
test-suite :: External/SPEC/CFP2017speed/638.imagick_s/638.imagick_s.test 1396320.00 1396448.00 0.0%
test-suite :: External/SPEC/CFP2017rate/538.imagick_r/538.imagick_r.test 1396320.00 1396448.00 0.0%
test-suite :: MultiSource/Benchmarks/Bullet/bullet.test 309790.00 309678.00 -0.0%
test-suite :: External/SPEC/CFP2017rate/526.blender_r/526.blender_r.test 12477607.00 12470807.00 -0.1%
CINT2006/445.gobmk - extra code vectorized
MiBench/consumer-lame - small variations
CFP2017speed/638.imagick_s
CFP2017rate/538.imagick_r - extra vectorized code
Benchmarks/Bullet - extra code vectorized
CFP2017rate/526.blender_r - extra vector code
RISC-V, sifive-p670, -O3+LTO
CFP2006/433.milc - regressions, should be fixed by https://github.com/llvm/llvm-project/pull/115173
CFP2006/453.povray - extra vectorized code
CFP2017rate/508.namd_r - better vector code
CFP2017rate/510.parest_r - extra vectorized code
SPEC/CFP2017rate - extra/better vector code
CFP2017rate/526.blender_r - extra vectorized code
CFP2017rate/538.imagick_r - extra vectorized code
CINT2006/403.gcc - extra vectorized code
CINT2006/445.gobmk - extra vectorized code
CINT2006/464.h264ref - extra vectorized code
CINT2006/483.xalancbmk - small variations
CINT2017rate/525.x264_r - better vectorization
Reviewers: RKSimon
Reviewed By: RKSimon
Pull Request: https://github.com/llvm/llvm-project/pull/115201
If several reduced values are combined and the first reduced value is
just the original reduced value of the bool logical op, need to freeze
it to prevent the propagation of the poison value.
Fixes#114905
This re-lands the reverted #92418
When the VF is small enough so that dividing the VF by the scaling
factor results in 1, the reduction phi execution thinks the VF is scalar
and sets the reduction's output as a scalar value, tripping assertions
expecting a vector value. The latest commit in this PR fixes that by
using `State.VF` in the scalar check, rather than the divided VF.
---------
Co-authored-by: Nicholas Guy <nicholas.guy@arm.com>
If the operands of the icmp instructions has reduced bitwidth after
MinBitwidth analysis, need to drop samesign flag to preserve correctness
of the transformation.
Fixes#120823
We already canonicalized an undef base vector to the RHS to improve further folding, this extends this to improve the shuffle cost estimate of the single src shuffle
foldShuffleOfShuffles currently only folds unary shuffles to ensure we don't end up with a merged shuffle with more than 2 sources, but this prevented cases where both shuffles were sharing sources.
This patch generalizes the merge process to find up to 2 sources as it merges with the inner shuffles, it also moves the undef/poison handling stages into the merge loop as well.
Fixes#120764
Bail out negative length while inferring initializes attr. Otherwise it
causes an assertion error:
`Attribute 'initializes' does not support unordered ranges`
This patch makes the MemProf undrifting process a little more lenient.
Consider an inlined call hierarchy:
foo -> bar -> ::new
If bar tail-calls ::new, the profile appears to indicate that foo
directly calls ::new. This is a problem because the perceived call
hierarchy in the profile looks different from what we can obtain from
the inline stack in the IR.
Recall that undrifting works by constructing and comparing a list of
direct calls from the profile and that from the IR. This patch
modifies the construction of the latter. Specifically, if foo calls
bar in the IR, but bar is missing the profile, we pretend that foo
directly calls some heap allocation function. We apply this
transformation only in the inline stack leading to some heap
allocation function.
Optionally unconditionally hint allocations as cold or not cold during
the cloning step if the percentage of bytes allocated is at least that
of the given threshold. This is similar to PR120301 which supports this
during matching, but enables the same behavior during cloning, to reduce
the false positives that can be addressed by cloning at the cost of
carrying the additional size metadata/summary.
I've only fixed up the tests where I was able to use a simple sed script
to replace the text. Even after this patch lands, there are still over
50 tests that need updating in X86/CostModel!
This fixes some regressions from recent changes to vector combine in
#120216. It allows shuffleToIdentity to look through fp casts as other
casts, and makes sure mismatching vector types in splats and casts do
not block the transform, as only the lanes should matter.
This PR is motivated by a mismatch we discovered between compilation
results with vs. without `-g3`. We noticed this when compiling SPEC2017
testcases. The specific instance we saw is fixed in this PR by modifying
a guard (see below), but it is likely similar instances exist elsewhere
in the codebase.
The specific case fixed in this PR manifests itself in the `SimplifyCFG`
pass doing different things depending on whether DebugInfo is generated
or not. At the end of this comment, there is reduced example code that
shows the behavior in question.
The differing behavior has two root causes:
1. Commit https://github.com/llvm/llvm-project/commit/c07e19b adds loop
metadata including debug locations to loops that otherwise would not
have loop metadata
2. Commit https://github.com/llvm/llvm-project/commit/ac28efa6c100 adds
a guard to a simplification action in `SImplifyCFG` that prevents it
from simplifying away loop metadata
So, the change in 2. does not consider that when compiling with debug
symbols, loops that otherwise would not have metadata that needs
preserving, now have debug locations in their loop metadata. Thus, with
`-g3`, `SimplifyCFG` behaves differently than without it.
The larger issue is that while debug info is not supposed to influence
the final compilation result, commits like 1. blur the line between what
is and is not debug info, and not all optimization passes account for
this.
This PR does not address that and rather just modifies this particular
guard in order to restore equivalent behavior between debug and
non-debug builds in this one instance.
---
Here is a reduced version of a file from `f526.blender_r` that showcases
the behavior in question:
```C
struct LinkNode;
typedef struct LinkNode {
struct LinkNode *next;
void *link;
} LinkNode;
void do_projectpaint_thread_ph_v_state() {
int *ps = do_projectpaint_thread_ph_v_state;
LinkNode *node;
while (do_projectpaint_thread_ph_v_state)
for (node = ps; node; node = node->next)
;
}
```
Compiling this with and without DebugInfo, and then disassembling the
results, leads to different outcomes (tested on SystemZ and X86). The
reason for this is that the `SimplifyCFG` pass does different things in
either case.
insertelt DestVec, (fneg (extractelt SrcVec, Index)), Index
-> shuffle DestVec, (shuffle (fneg SrcVec), poison, SrcMask), Mask
Original combining left the combine between vectors of different lengths as a TODO. this commit do that. (see
#[baab4aa1ba])
processShuffleMasks can now correctly handle 2 src shuffles, so we can use the existing SK_PermuteSingleSrc splitting cost logic to handle SK_PermuteTwoSrc as well and correctly recognise the number of active subvectors per legalised shuffle.
processShuffleMasks can now correctly handle 2 src shuffles, so we can use the existing SK_PermuteSingleSrc splitting cost logic to handle SK_PermuteTwoSrc as well and correctly recognise the number of active subvectors per legalised shuffle.
Add tests for horizontal add patterns with missing/undemanded elements - which typically prevents folding to the (add (shuffle a, b),(shuffle a, b)) optimal pattern
Following on from https://github.com/llvm/llvm-project/pull/94499, this
patch adds support to the Loop Vectorizer to emit the partial reduction
intrinsics where they may be beneficial for the target.
---------
Co-authored-by: Samuel Tebbs <samuel.tebbs@arm.com>
This patch undrifts source locations in MemProfRecord before readMemprof
starts the matching process.
The thoery of operation is as follows:
1. Collect the lists of direct calls, one from the IR and the other
from the profile.
2. Compute the correspondence (called undrift map in the patch)
between the two lists with longestCommonSequence.
3. Apply the undrift map just before readMemprof consumes
MemProfRecord.
The new function gated by a flag that is off by default.
Don't unnecessarily clone for a caller that wasn't matched to a call
instruction.
This necessitated updated a couple of tests that were either
unnecessarily cloning or unnecessarily processing an allocation and
hinting it not cold.
Currently the addUsersInExitBlocks incorrectly assumes exit phis only
have a single operand, which may not be the case for loops with early
exits when they share a common exit block.
Also further relax the assertion in fixupIVUsers to allow exit values if
they come from theloop latch/middle.block.
PR: https://github.com/llvm/llvm-project/pull/120260
