This reverts commit 7d4e8c1f3bbfe976f4871c9cf953f76d771b0eda.
Contrary to the commit description, this does cause large
compile-time regressions (up to 10% on individual files).
This change updates a few of the transformations in foldFMulReassoc to
respect absent fast-math flags in cases where fmul and fdiv, fadd, or fsub
instructions were being folded but the code was only checking for
fast-math flags on the fmul instruction and was transferring flags to
the folded instruction that were not present on the other original
instructions.
This fixes https://github.com/llvm/llvm-project/issues/82857
We can try to vectorize long store sequences, if short ones were
unsuccessful because of the non-profitable vectorization. It should not
increase compile time significantly (stores are sorted already,
complexity is n x log n), but vectorize extra code.
Metric: size..text
Program size..text
results results0 diff
test-suite :: External/SPEC/CINT2006/400.perlbench/400.perlbench.test 1088012.00 1088236.00 0.0%
test-suite :: SingleSource/UnitTests/matrix-types-spec.test 480396.00 480476.00 0.0%
test-suite :: External/SPEC/CINT2017rate/525.x264_r/525.x264_r.test 664613.00 664661.00 0.0%
test-suite :: External/SPEC/CINT2017speed/625.x264_s/625.x264_s.test 664613.00 664661.00 0.0%
test-suite :: External/SPEC/CFP2017rate/510.parest_r/510.parest_r.test 2041105.00 2040961.00 -0.0%
test-suite :: MultiSource/Applications/JM/lencod/lencod.test 836563.00 836387.00 -0.0%
test-suite :: MultiSource/Benchmarks/7zip/7zip-benchmark.test 1035100.00 1032140.00 -0.3%
In all benchmarks extra code gets vectorized
Reviewers: RKSimon
Reviewed By: RKSimon
Pull Request: https://github.com/llvm/llvm-project/pull/88563
Using fp type in the compiler is not the best idea, here it used with
the comparison for equal to 0 and may cause undefined behavior in some
cases.
Reviewers: fhahn
Reviewed By: fhahn
Pull Request: https://github.com/llvm/llvm-project/pull/87241
Need to check that unsigned argument can be safely used in smax/smin
intrinsics by checking if at least single sign bit is cleared, otherwise
its value may be treated as negative instead of positive.
Fixes#76609
This patch does:
- relax the phis constraint in `CanRedirectPredsOfEmptyBBToSucc`
- guarantee the BB has multiple different predecessors to redirect, so
that we can handle the case without phis in BB. Without this change and
phi constraint, we may redirect the CommonPred.
The motivation is consistent with JumpThreading. We always want the
branch to jump more direct to the destination, without passing the
middle block. In this way, we can expose more other optimization
opportunities.
An obivous example proposed by @dtcxzyw is like:
```llvm
define i32 @test(...) {
entry:
br i1 %c, label %do.end, label %if.then
if.then: ; preds = %entry
%call2 = call i32 @dummy()
%tobool3.not = icmp eq i32 %call2, 0
br i1 %tobool3.not, label %do.end, label %return
do.end: ; preds = %entry, %if.then
br label %return
return: ; preds = %if.then, %do.end
%retval.0 = phi i32 [ 0, %do.end ], [ %call2, %if.then ]
ret i32 %retval.0
}
```
`entry` can directly jump to return, without passing `do.end`, and then
the if-else pattern can be simplified further:
```llvm
define i32 @test(...) {
entry:
br i1 %c, label %return, label %if.then
if.then: ; preds = %entry
%call2 = call i32 @dummy()
br label %return
return: ; preds = %if.then
%retval.0 = phi i32 [ 0, %entry ], [ %call2, %if.then ]
ret i32 %retval.0
}
```
When inserting truncs during IV widening, mark the trunc as either nuw
or nsw depending on whether zext or sext widening was used. For
non-negative IVs both nuw and nsw apply.
When build with option -msve-vector-bits=512, the return vaule of
Subtarget->getMinSVEVectorSizeInBits() is 512;
While the MinElts is still 4 for <vscale x 4 x double> in
getNumInterleavedAccesses, so it creates invalid
llvm.aarch64.sve.ld2.sret.nxv4f64, which need be splited.
Unlikely, the related custom spilting is not supported now.
Fix https://github.com/llvm/llvm-project/issues/88247
Also reapply "[llvm] Teach whole program devirtualization about
relative vtables"
This reverts commit 1c604a9780fcfe92a99d539913553f0835b81de3 and
474f5efebed24547e76d022f0c5ffcc9db97ce6f.
The option intended primarily for LIT tests to suppress heuristic based
profitability check and proceed vectorization of a seemingly
unprofitable alternate operation pattern. This allows the vectorizer to
execute path that was the original intent of a test.
This patch restructures the way names for printing VPValues are handled.
It moves the logic to generate names for printing to VPSlotTracker.
VPSlotTracker will now version names of the same underlying value if it
is used by multiple VPValues, by adding a .V suffix to the name.
This fixes cases where at the moment the same name is printed for
different VPValues.
PR: https://github.com/llvm/llvm-project/pull/81411
This reverts commit b9cd48f96acdd07c627ccafbf4386a1f3dcd6c51.
-------------------------------------------------------------
Original commit message:
Adds logic to the IR verifier that checks whether !tbaa.struct nodes are
well-formed. That is, it checks that the operands of !tbaa.struct nodes
are in groups of three, that each group of three operands consists of
two integers and a valid tbaa node, and that the regions described by
the offset and size operands are non-overlapping.
PR: https://github.com/llvm/llvm-project/pull/86709
An example from https://github.com/image-rs/image:
```
define void @test_ult_rhsc(i8 %x) {
%val = add nsw i8 %x, -2
%cmp = icmp ult i8 %val, 11
%cond = select i1 %cmp, i8 %val, i8 6
switch i8 %cond, label %bb1 [
i8 0, label %bb2
i8 10, label %bb3
]
bb1:
call void @func1()
unreachable
bb2:
call void @func2()
unreachable
bb3:
call void @func3()
unreachable
}
```
When `%cmp` evaluates to false, we can prove that the range of `%val` is
[11, umax]. Thus we can safely replace `%cond` with `%val` since both
`switch 6` and `switch %val` go to the default dest `%bb1`.
Alive2: https://alive2.llvm.org/ce/z/uSTj6w
Godbolt: https://godbolt.org/z/MGrG84bzr
This patch will benefit many rust applications and some C/C++
applications (e.g., cvc5).
As the undef can be replaced with a zero value, this is not legal
in the general case. We can only allow poison values. This matches
what the other ValueTracking helpers like computeKnownBits() do.
The large case index out of scope is dead code, but it is still be
created for TableContents in SwitchLookupTable::SwitchLookupTable,
so make sure the table size after growing should not get smaller.
Fix https://github.com/llvm/llvm-project/issues/88607
If not performing gep splits can prevent important optimizations, such
as preventing the element indices / member offsets from being
(partially) folded into load/store instruction immediates.
This ensures that MinVF is a power-of-2, even if ValueTy's width is
not a power-of-2.
This should fix a number of buildbot failures with X86 bootstrapping.
