Assume bundle can have more than one entry with the same name,
but at least AlignmentFromAssumptionsPass::extractAlignmentInfo() uses
getOperandBundle("align"), which internally assumes that it isn't the
case, and happily crashes otherwise.
Minimal reduced reproducer: run `opt -alignment-from-assumptions` on
target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"
%0 = type { i64, %1*, i8*, i64, %2, i32, %3*, i8* }
%1 = type opaque
%2 = type { i8, i8, i16 }
%3 = type { i32, i32, i32, i32 }
; Function Attrs: nounwind
define i32 @f(%0* noalias nocapture readonly %arg, %0* noalias %arg1) local_unnamed_addr #0 {
bb:
call void @llvm.assume(i1 true) [ "align"(%0* %arg, i64 8), "align"(%0* %arg1, i64 8) ]
ret i32 0
}
; Function Attrs: nounwind willreturn
declare void @llvm.assume(i1) #1
attributes #0 = { nounwind "reciprocal-estimates"="none" }
attributes #1 = { nounwind willreturn }
This is what we'd have with -mllvm -enable-knowledge-retention
This reverts commit c95ffadb24.
binop i1 (cmp Pred (ext X, Index0), C0), (cmp Pred (ext X, Index1), C1)
-->
vcmp = cmp Pred X, VecC
ext (binop vNi1 vcmp, (shuffle vcmp, Index1)), Index0
This is a larger pattern than the existing extractelement folds because we can't
reasonably vectorize the sub-patterns with constants based on cost model calcs
(it doesn't usually make sense to replace a single extracted scalar op with
constant operand with a vector op).
I salvaged as much of the existing logic as I could, but there might be better
ways to share and reduce code.
The motivating case from PR43745:
https://bugs.llvm.org/show_bug.cgi?id=43745
...is the special case of a 2-way reduction. We tried to get SLP to handle that
particular pattern in D59710, but that caused crashing and regressions.
This patch is more general, but hopefully safer.
The v2f64 test with SSE2 surprised me - the cost model accounting looks like this:
OldCost = 0 (free extract of f64 at index 0) + 1 (extract of f64 at index 1) + 2 (scalar fcmps) + 1 (and of bools) = 4
NewCost = 2 (vector fcmp) + 1 (shuffle) + 1 (vector 'and') + 1 (extract of bool) = 5
Differential Revision: https://reviews.llvm.org/D82474
Summary:
NOTE: There is a mailing list discussion on this: http://lists.llvm.org/pipermail/llvm-dev/2019-December/137632.html
Complemantary to the assumption outliner prototype in D71692, this patch
shows how we could simplify the code emitted for an alignemnt
assumption. The generated code is smaller, less fragile, and it makes it
easier to recognize the additional use as a "assumption use".
As mentioned in D71692 and on the mailing list, we could adopt this
scheme, and similar schemes for other patterns, without adopting the
assumption outlining.
Reviewers: hfinkel, xbolva00, lebedev.ri, nikic, rjmccall, spatel, jdoerfert, sstefan1
Reviewed By: jdoerfert
Subscribers: yamauchi, kuter, fhahn, merge_guards_bot, hiraditya, bollu, rkruppe, cfe-commits, llvm-commits
Tags: #clang, #llvm
Differential Revision: https://reviews.llvm.org/D71739
This is the integer sibling to D81491.
(a[0] + a[1] + a[2] + a[3]) - (b[0] + b[1] + b[2] +b[3]) -->
(a[0] - b[0]) + (a[1] - b[1]) + (a[2] - b[2]) + (a[3] - b[3])
Removing the "experimental" from these intrinsics is likely
not too far away.
This saves creating/destroying a builder every time we
perform some transform.
The tests show instruction ordering diffs resulting from
always inserting at the root instruction now, but those
should be benign.
(a[0] + a[1] + a[2] + a[3]) - (b[0] + b[1] + b[2] +b[3]) -->
(a[0] - b[0]) + (a[1] - b[1]) + (a[2] - b[2]) + (a[3] - b[3])
This should be the last step in solving PR43953:
https://bugs.llvm.org/show_bug.cgi?id=43953
We started emitting reduction intrinsics with:
D80867/ rGe50059f6b6b3
So it's a relatively easy pattern match now to re-order those ops.
Also, I have not seen any complaints for the switch to intrinsics
yet, so I'll propose to remove the "experimental" tag from the
intrinsics soon.
Differential Revision: https://reviews.llvm.org/D81491
scalarizeBinop currently folds
vec_bo((inselt VecC0, V0, Index), (inselt VecC1, V1, Index))
->
inselt(vec_bo(VecC0, VecC1), scl_bo(V0,V1), Index)
This patch extends this to account for cases where one of the vec_bo operands is already all-constant and performs similar cost checks to determine if the scalar binop with a constant still makes sense:
vec_bo((inselt VecC0, V0, Index), VecC1)
->
inselt(vec_bo(VecC0, VecC1), scl_bo(V0,extractelt(V1,Index)), Index)
Fixes PR42174
Differential Revision: https://reviews.llvm.org/D80885
Motivating examples are seen in the PhaseOrdering tests based on:
https://bugs.llvm.org/show_bug.cgi?id=43953#c2 - if we have
intrinsics there, some pass can fold them.
The intrinsics are still named "experimental" at this point, but
if there is no fallout from this patch, that will be a good
indicator that it is safe to finalize them.
Differential Revision: https://reviews.llvm.org/D80867
As discussed in D80236 - this test (like all PhaseOrdering tests?)
was intended to show that there is no difference with the new
pass manager, but the 'opt' command requires extra parameters
to make that happen.
There are 2 known problem patterns shown in the test diffs here:
vector horizontal ops (an x86 specialization) and vector reductions.
SLP has greater ability to match and fold those than vector-combine,
so let SLP have first chance at that.
This is a quick fix while we continue to improve vector-combine and
possibly canonicalize to reduction intrinsics.
In the longer term, we should improve matching of these patterns
because if they were created in the "bad" forms shown here, then we
would miss optimizing them.
I'm not sure what is happening with alias analysis on the addsub test.
The old pass manager now shows an extra line for that, and we see an
improvement that comes from SLP vectorizing a store. I don't know
what's missing with the new pass manager to make that happen.
Strangely, I can't reproduce the behavior if I compile from C++ with
clang and invoke the new PM with "-fexperimental-new-pass-manager".
Differential Revision: https://reviews.llvm.org/D80236
This eliminates a use of 'B', so it can enable follow-on transforms
as well as improve analysis/codegen.
The PhaseOrdering test was added for D61726, and that shows
the limits of instcombine vs. real reassociation. We would
need to run some form of CSE to collapse that further.
The intermediate variable naming here is intentional because
there's a test at llvm/test/Bitcode/value-with-long-name.ll
that would break with the usual nameless value. I'm not sure
how to improve that test to be more robust.
The naming may also be helpful to debug regressions if this
change exposes weaknesses in the reassociation pass for example.
This is split off from D79799 - where I was proposing to fully iterate
over a function until there are no more transforms. I suspect we are
still going to want to do something like that eventually.
But we can achieve the same gains much more efficiently on the current
set of regression tests just by reversing the order that we visit the
instructions.
This may also reduce the motivation for D79078, but we are still not
getting the optimal pattern for a reduction.
In D74183 clang started emitting alignment for sret parameters
unconditionally. This caused a 1.5% compile-time regression on
tramp3d-v4. The reason is that we now generate many instance of IR like
%ptrint = ptrtoint %class.GuardLayers* %guards_m to i64
%maskedptr = and i64 %ptrint, 3
%maskcond = icmp eq i64 %maskedptr, 0
tail call void @llvm.assume(i1 %maskcond)
to preserve the alignment information during inlining. Based on IR
analysis, these assumptions also regress optimization. The attached
phase ordering test case illustrates two issues: One are instruction
count based optimization heuristics, which are affected by the four
additional instructions of the assumption. The other is blocking of
SROA due to ptrtoint casts (PR45763).
We already encountered the same problem in Rust, where we (unlike
Clang) generally prefer to emit alignment information absolutely
everywhere it is available. We were only able to do this after
hardcoding -preserve-alignment-assumptions-during-inlining=false,
because we were seeing significant optimization and compile-time
regressions otherwise.
This patch disables -preserve-alignment-assumptions-during-inlining
by default, because we should not be punishing people for adding
more alignment annotations.
Once the assume bundle work shakes out and we can represent (and use)
alignment assumptions using assume bundles, it should be possible to
re-enable this with reduced overhead.
Differential Revision: https://reviews.llvm.org/D76886
bitcast (shuf V, MaskC) --> shuf (bitcast V), MaskC'
This is the widen shuffle elements enhancement to D76727.
It builds on the analysis and simplifications in
D77881 and rG6a7e958a423e.
The phase ordering tests show that we can simplify inverse
shuffles across a binop in both directions (widen/narrow or
narrow/widen) now.
There's another potential transform visible in some of the
remaining TODOs - move a bitcasted operand of a shuffle
after the shuffle.
Differential Revision: https://reviews.llvm.org/D78371
As discussed in D76983, that patch can turn a chain of insert/extract
with scalar trunc ops into bitcast+extract and existing instcombine
vector transforms end up creating a shuffle out of that (see the
PhaseOrdering test for an example). Currently, that process requires
at least this sequence: -instcombine -early-cse -instcombine.
Before D76983, the sequence of insert/extract would reach the SLP
vectorizer and become a vector trunc there.
Based on a small sampling of public targets/types, converting the
shuffle to a trunc is better for codegen in most cases (and a
regression of that form is the reason this was noticed). The trunc is
clearly better for IR-level analysis as well.
This means that we can induce "spontaneous vectorization" without
invoking any explicit vectorizer passes (at least a vector cast op
may be created out of scalar casts), but that seems to be the right
choice given that we started with a chain of insert/extract, and the
backend would expand back to that chain if a target does not support
the op.
Differential Revision: https://reviews.llvm.org/D77299
We got some of the potential optimizations with D76727 and D76844.
There are 2 likely enhancements that we could add to -vector-combine
to get most of the remaining cases:
1. Allow bitcasted shuffle mask narrowing (widen the elements).
2. Combine shuffle-of-shuffle into a single shuffle.
This is already partly handled by the x86 backend, but the
tests here show that we still miss some of the potential
combines.
These are versions of a function that regressed with:
rGf2fbdf76d8d0
That particular problem occurs with an instcombine-simplifycfg-instcombine
sequence, but we can show that it exists within instcombine only with
other variations of the pattern.
The initial placement of vector-combine in the opt pipeline revealed phase ordering bugs:
https://bugs.llvm.org/show_bug.cgi?id=45015https://bugs.llvm.org/show_bug.cgi?id=42022
This patch contains a few independent changes:
1. Move the pass up in the pipeline, so it happens just after loop-vectorization.
This is only to keep vectorization passes together in the pipeline at the moment.
I don't have evidence of interaction between these yet.
2. Add an -early-cse pass after -vector-combine to clean up redundant ops. This was
partly proposed as far back as rL219644 (which is why it's effectively being moved
in the old PM code). This is important because the subsequent -instcombine doesn't
work as well without EarlyCSE. With the CSE, -instcombine is able to squash
shuffles together in 1 of the tests (because those are simple "select" shuffles).
3. Remove the -vector-combine pass that was running after SLP. We may want to do that
eventually, but I don't have a test case to support it yet.
Differential Revision: https://reviews.llvm.org/D75145
As discussed in PR41083:
https://bugs.llvm.org/show_bug.cgi?id=41083
...we can assert/crash in EarlyCSE using the current hashing scheme and
instructions with flags.
ValueTracking's matchSelectPattern() may rely on overflow (nsw, etc) or
other flags when detecting patterns such as min/max/abs composed of
compare+select. But the value numbering / hashing mechanism used by
EarlyCSE intersects those flags to allow more CSE.
Several alternatives to solve this are discussed in the bug report.
This patch avoids the issue by doing simple matching of min/max/abs
patterns that never requires instruction flags. We give up some CSE
power because of that, but that is not expected to result in much
actual performance difference because InstCombine will canonicalize
these patterns when possible. It even has this comment for abs/nabs:
/// Canonicalize all these variants to 1 pattern.
/// This makes CSE more likely.
(And this patch adds PhaseOrdering tests to verify that the expected
transforms are still happening in the standard optimization pipelines.
I left this code to use ValueTracking's "flavor" enum values, so we
don't have to change the callers' code. If we decide to go back to
using the ValueTracking call (by changing the hashing algorithm
instead), it should be obvious how to replace this chunk.
Differential Revision: https://reviews.llvm.org/D74285
Test that instcombine and early-cse can cooperate
to reduce sequences of select patterns that are not
composed of the same underlying instructions.
There's a bug in EarlyCSE (PR41083), and we can test
how much a possible fix (D74285) may affect optimization.
GEP index size can be specified in the DataLayout, introduced in D42123. However, there were still places
in which getIndexSizeInBits was used interchangeably with getPointerSizeInBits. This notably caused issues
with Instcombine's visitPtrToInt; but the unit tests was incorrect, so this remained undiscovered.
This fixes the buildbot failures.
Differential Revision: https://reviews.llvm.org/D68328
Patch by Joseph Faulls!
GEP index size can be specified in the DataLayout, introduced in D42123. However, there were still places
in which getIndexSizeInBits was used interchangeably with getPointerSizeInBits. This notably caused issues
with Instcombine's visitPtrToInt; but the unit tests was incorrect, so this remained undiscovered.
Differential Revision: https://reviews.llvm.org/D68328
Patch by Joseph Faulls!
This reapplies: 8ff85ed905
Original commit message:
As a follow-up to my initial mail to llvm-dev here's a first pass at the O1 described there.
This change doesn't include any change to move from selection dag to fast isel
and that will come with other numbers that should help inform that decision.
There also haven't been any real debuggability studies with this pipeline yet,
this is just the initial start done so that people could see it and we could start
tweaking after.
Test updates: Outside of the newpm tests most of the updates are coming from either
optimization passes not run anymore (and without a compelling argument at the moment)
that were largely used for canonicalization in clang.
Original post:
http://lists.llvm.org/pipermail/llvm-dev/2019-April/131494.html
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65410
This reverts commit c9ddb02659.
This change doesn't include any change to move from selection dag to fast isel
and that will come with other numbers that should help inform that decision.
There also haven't been any real debuggability studies with this pipeline yet,
this is just the initial start done so that people could see it and we could start
tweaking after.
Test updates: Outside of the newpm tests most of the updates are coming from either
optimization passes not run anymore (and without a compelling argument at the moment)
that were largely used for canonicalization in clang.
Original post:
http://lists.llvm.org/pipermail/llvm-dev/2019-April/131494.html
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D65410
