https://reviews.llvm.org/D78411 introduced test changes which relied on
the ability to strip debugify metadata even if module-level metadata is
missing. This introduces a more targeted test for that ability.
Summary: The native alignment may generally not be used when lowering a vector.transfer to the underlying load/store operation. This revision fixes the unmasked load/store alignment to match that of the masked path.
Differential Revision: https://reviews.llvm.org/D83684
This is intended to address a compile-time regression from
1eddce4177. A SmallPtrSet was
replaced with a SetVector there, which had an unexpected large
compile-time impact. It turns out that this structure is getting
swapped a lot, and previously this used an optimized std::swap()
specialization for SmallPtrSet. Now it ends up using the default,
triple-move based implementation, which is much more expensive.
This patch (partly) addresses the issue by specializing std::swap()
for SetVector.
Differential Revision: https://reviews.llvm.org/D82230
Original commit c60216db15.
The test can only run on Darwin because of how it was setup, so I'm
enforcing that.
Summary:
Test Plan:
Reviewers:
Subscribers:
Tasks:
Tags:
- Provide default value for `ArrayRef<NamedAttribute> attributes` parameter of
the collective params build method.
- Change the `genSeparateArgParamBuilder` function to not generate build methods
that may be ambiguous with the new collective params build method.
- This change should help eliminate passing empty NamedAttribue ArrayRef when the
collective params build method is used
- Extend op-decl.td unit test to make sure the ambiguous build methods are not
generated.
Differential Revision: https://reviews.llvm.org/D83517
Per the Vulkan's SPIR-V environment spec, "for the OpSRem and OpSMod
instructions, if either operand is negative the result is undefined."
So we cannot directly use spv.SRem/spv.SMod if either operand can be
negative. Emulate it via spv.UMod.
Because the emulation uses spv.SNegate, this commit also defines
spv.SNegate.
Differential Revision: https://reviews.llvm.org/D83679
This allows skipping a test when running the testsuite on macOS under
the Rosetta translation layer.
Differential Revision: https://reviews.llvm.org/D83600
This is in preparation for the 'development' mode advisor. We currently
want to track what the default policy's decision would have been, this
refactoring makes it easier to do that.
Character literals can be formatted using octal or hex escapes for
non-ascii characters. This is so that the program can be unparsed for
either pgf90 or gfortran to compile. But modfiles should not be affected
by that -- they should be consistent.
This changes causes modfiles to always have character literals formatted
with octal escapes.
Differential Revision: https://reviews.llvm.org/D83703
Place the ssa.copy instructions for assumes after the assume,
instead of before it. Both options are valid, but placing them
afterwards prevents assumes from being replaced with assume(true).
This fixes https://bugs.llvm.org/show_bug.cgi?id=37541 in NewGVN
and will avoid a similar issue in SCCP when we handle more
predicate infos.
Differential Revision: https://reviews.llvm.org/D83631
Also compacted the checkpoints (variables) to one file (plus the index).
This reduces the binary model files to just the variables and their
index. The index is very small. The variables are serialized float
arrays. When updated through training, the changes are very likely
unlocalized, so there's very little value in them being anything else
than binary.
This was emitting the raw value for the reg class ID with a comment
for the actual class name. Switch to emitting the qualified enum name
instead, which obviates the need for the comment and also helps keep
the lit tests on the emitter output more stable.
The GlobalISelEmitter is stricter about matching timm instruction
outputs to timm inputs (although in an accidental sort of way that
doesn't hit a proper import failure error). Also, apparently no
intrinsic patterns were importing since the ID enum declaration was
missing.
This patch generalizes the APIs for writing re-entry blocks, trampolines and
stubs to allow their final linked address to differ from the address of
their initial working memory. This will allow these routines to be used with
JITLinkMemoryManagers, which will in turn allow for unification of code paths
for in-process and cross-process lazy JITing.
This will be used by upcoming patches that implement indirection utils
(reentry, reentry trampolines, and stubs) on top of
JITLinkMemoryManager to unify in-process and cross-process lazy
compilation support.
Summary:
This is an experimental ML-based native size estimator, necessary for
computing partial rewards during -Oz inliner policy training. Data
extraction for model training will be provided in a separate patch.
RFC: http://lists.llvm.org/pipermail/llvm-dev/2020-April/140763.html
Reviewers: davidxl, jdoerfert
Subscribers: mgorny, hiraditya, mgrang, arphaman, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D82817
We have this generic transform in IR (instcombine),
but as shown in PR41098:
http://bugs.llvm.org/PR41098
...the pattern may emerge in codegen too.
x86 has a potential refinement/reversal opportunity here,
but that should come later or needs a target hook to
avoid the transform. Converting to bswap is the more
specific form, so we should use it if it is available.
Summary:
The test fails on AIX as it allows open() and read() on a directory. This patch adds `# UNSUPPORTED:
system-aix` to the test to prevent it from running on AIX.
Reviewers: sameerarora101, daltenty, ShuhongL, hubert.reinterpretcast, MaskRay, smeenai, alexshap
Reviewed By: sameerarora101, hubert.reinterpretcast, MaskRay, smeenai
Subscribers: MaskRay, rupprecht, llvm-commits
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D83579
Benjamin Kramer