LLVM/project 7b09ee8llvm/test/Transforms/SLPVectorizer/AArch64 accelerate-vector-functions-inseltpoison.ll aarch64-vector-functions.ll, llvm/test/Transforms/SLPVectorizer/RISCV math-function.ll

[𝘀𝗽𝗿] initial version

Created using spr 1.3.7
DeltaFile
+352-572llvm/test/Transforms/SLPVectorizer/RISCV/math-function.ll
+90-134llvm/test/Transforms/SLPVectorizer/AArch64/accelerate-vector-functions-inseltpoison.ll
+86-130llvm/test/Transforms/SLPVectorizer/AArch64/aarch64-vector-functions.ll
+6-126llvm/test/Transforms/SLPVectorizer/X86/powi.ll
+17-42llvm/test/Transforms/SLPVectorizer/X86/call.ll
+17-41llvm/test/Transforms/SLPVectorizer/X86/intrinsic.ll
+568-1,0459 files not shown
+666-1,22915 files

LLVM/project 678a6c7llvm/lib/Target/DirectX DXILOpLowering.cpp DXIL.td, llvm/test/CodeGen/DirectX SampleGrad.ll

[DirectX] Implement lowering of SampleGrad. Refactor sample lowering functions (#200871)

Fixes #192549

This PR builds atop #199745 by adding a helper `lowerSampleOp` function
to refactor `lowerSampleBias` and implement `lowerSampleGrad`.

The `lowerSampleGrad` implementation is very similar to
`lowerSampleBias`, just with ddx and ddy arguments instead of a bias.
Unlike SampleBias, SampleGrad is usable in all shader stages because it
has explicit gradient/derivative arguments.

Assisted-by: GitHub Copilot
DeltaFile
+338-0llvm/test/CodeGen/DirectX/SampleGrad.ll
+59-16llvm/lib/Target/DirectX/DXILOpLowering.cpp
+17-0llvm/lib/Target/DirectX/DXIL.td
+414-163 files

LLVM/project 91edd87llvm/lib/CodeGen/SelectionDAG LegalizeDAG.cpp, llvm/lib/Target/PowerPC PPCISelLowering.cpp

[Legalizer] Add support for promoting integers for s/ucmp (#198554)

Instead of manually doing promotion in LowerUCMP of platforms like
PowerPC, we should have this logic in the legalizer.
DeltaFile
+8-11llvm/lib/Target/PowerPC/PPCISelLowering.cpp
+15-0llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp
+2-6llvm/test/CodeGen/PowerPC/ucmp.ll
+25-173 files

LLVM/project 56a553bclang-tools-extra/clang-doc BitcodeReader.cpp Serialize.cpp, clang-tools-extra/clang-doc/tool ClangDocMain.cpp

[clang-doc] Wrap per thread arenas in an accessor for BUILD_SHARED

It seems like for BUILD_SHARED builds of the toolchain on Windows,
specifically aarch64-windows-gnu hosts, the use of the `thread_local`
variables in Representation.cpp causes an issue at link time due to
non-explicit export. Instead, just wrap them in an accessor function,
which should solve the issue in a cross platform way.

Fixes #200915
DeltaFile
+25-24clang-tools-extra/clang-doc/BitcodeReader.cpp
+26-23clang-tools-extra/clang-doc/Serialize.cpp
+24-16clang-tools-extra/clang-doc/Representation.cpp
+10-9clang-tools-extra/clang-doc/Representation.h
+4-2clang-tools-extra/unittests/clang-doc/MergeTest.cpp
+3-2clang-tools-extra/clang-doc/tool/ClangDocMain.cpp
+92-761 files not shown
+93-777 files

LLVM/project c6c2ad5llvm/lib/CodeGen/SelectionDAG DAGCombiner.cpp, llvm/test/CodeGen/X86 aext-and-trunc.ll aext-and-trunc-avx512.ll

[DAG] Fold any-extend(and(trunc(x), C)) -> and(x, C) (#200052)

Fixes #195575

Fix a missed optimization in `DAGCombiner::visitANY_EXTEND` where the
pattern `any-extend(and(trunc(x), C))` was not being folded into `and(x,
C)` on X86, causing a redundant `movzbl` instruction to be emitted after
a small-mask AND.
DeltaFile
+57-17llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+47-0llvm/test/CodeGen/X86/aext-and-trunc.ll
+41-0llvm/test/CodeGen/X86/aext-and-trunc-avx512.ll
+15-15llvm/test/CodeGen/X86/masked_load.ll
+2-2llvm/test/CodeGen/X86/unfold-masked-merge-scalar-constmask-innerouter.ll
+2-2llvm/test/CodeGen/X86/and-with-overflow.ll
+164-3610 files not shown
+179-5116 files

LLVM/project 788018aclang/docs LanguageExtensions.rst, clang/include/clang/Options Options.td

Enable driver changes for fexec-charset
DeltaFile
+14-6clang/lib/Driver/ToolChains/Clang.cpp
+14-4clang/include/clang/Options/Options.td
+11-3clang/test/Driver/clang_f_opts.c
+10-0llvm/lib/Support/TextEncoding.cpp
+4-3clang/test/Driver/cl-options.c
+3-3clang/docs/LanguageExtensions.rst
+56-193 files not shown
+60-199 files

LLVM/project 0d7da31clang/include/clang/Basic TargetInfo.h, clang/lib/AST ASTContext.cpp

convert to exec-charset inside getPredefinedStringLiteralFromCache, test __builtin_FILE()
DeltaFile
+10-0clang/lib/AST/ASTContext.cpp
+3-2clang/lib/Lex/TextEncoding.cpp
+4-0clang/test/CodeGen/systemz-charset.cpp
+3-0clang/lib/Basic/TargetInfo.cpp
+2-0clang/include/clang/Basic/TargetInfo.h
+22-25 files

LLVM/project a844a5dclang/include/clang/Options Options.td, clang/lib/Driver/ToolChains Clang.cpp

address comments
DeltaFile
+2-2clang/include/clang/Options/Options.td
+1-1clang/lib/Driver/ToolChains/Clang.cpp
+3-32 files

LLVM/project cb0be65clang/lib/AST ASTContext.cpp

Convert the key before cache lookup to prevent encoding differences
DeltaFile
+9-9clang/lib/AST/ASTContext.cpp
+9-91 files

LLVM/project becdbc5clang/lib/Sema SemaChecking.cpp

do not convert character by character
DeltaFile
+6-4clang/lib/Sema/SemaChecking.cpp
+6-41 files

LLVM/project b67412bclang/include/clang/Lex TextEncoding.h, clang/lib/Lex TextEncoding.cpp

fix CI
DeltaFile
+2-2clang/lib/Lex/TextEncoding.cpp
+1-0clang/include/clang/Lex/TextEncoding.h
+3-22 files

LLVM/project 060a216clang/lib/AST PrintfFormatString.cpp FormatString.cpp, clang/lib/Sema SemaChecking.cpp

Add format string handling
DeltaFile
+58-31clang/lib/AST/PrintfFormatString.cpp
+46-40clang/lib/AST/FormatString.cpp
+33-21clang/lib/Sema/SemaChecking.cpp
+25-11clang/lib/AST/FormatStringParsing.h
+15-8clang/lib/AST/ScanfFormatString.cpp
+19-0llvm/lib/Support/TextEncoding.cpp
+196-1117 files not shown
+233-12013 files

LLVM/project ae86b34clang/test/CodeGen systemz-charset.c

fix CI
DeltaFile
+2-0clang/test/CodeGen/systemz-charset.c
+2-01 files

LLVM/project 134330bclang/include/clang/AST Expr.h, clang/lib/AST Expr.cpp

fix CI
DeltaFile
+14-0clang/lib/AST/Expr.cpp
+6-0clang/include/clang/AST/Expr.h
+3-2clang/lib/Sema/SemaExpr.cpp
+0-1clang/lib/Parse/ParseDecl.cpp
+23-34 files

LLVM/project 3d7c262clang/include/clang/Sema Sema.h

Remove old include
DeltaFile
+0-1clang/include/clang/Sema/Sema.h
+0-11 files

LLVM/project a5f97edclang/include/clang/Sema Sema.h

Fix build failure
DeltaFile
+1-0clang/include/clang/Sema/Sema.h
+1-01 files

LLVM/project 83538ebclang/lib/Parse ParseDecl.cpp ParseExpr.cpp, clang/lib/Sema SemaExpr.cpp

add ParserConversionAction, do not translate unevaluated strings
DeltaFile
+15-0clang/test/CodeGen/systemz-charset.c
+6-6clang/lib/Sema/SemaExpr.cpp
+10-0clang/lib/Parse/ParseDecl.cpp
+8-0clang/test/CodeGen/systemz-charset-diag.cpp
+3-3clang/lib/Parse/ParseExpr.cpp
+4-0clang/lib/Parse/Parser.cpp
+46-93 files not shown
+52-109 files

LLVM/project 1a2d25aclang/include/clang/Lex LiteralSupport.h, clang/lib/Sema SemaExpr.cpp

change default of StringLiteralParser to NoConversion instead of CA_ToExecEncoding
DeltaFile
+2-1clang/lib/Sema/SemaExpr.cpp
+1-1clang/include/clang/Lex/LiteralSupport.h
+3-22 files

LLVM/project 95453e6flang/lib/Semantics check-omp-structure.cpp

format
DeltaFile
+2-1flang/lib/Semantics/check-omp-structure.cpp
+2-11 files

LLVM/project 0f4b1e1llvm/lib/Target/AArch64 AArch64ISelLowering.cpp, llvm/test/CodeGen/AArch64 sve-fixed-length-trunc-stores.ll

[AArch64] Address issue reported in PR#196029 (#199122)

For certain types of truncating stores, the lowering action is set to
custom although no custom lowering exists for them.

This patch addresses issue reported in PR #196029 by removing the custom lowering entry.
DeltaFile
+26-0llvm/test/CodeGen/AArch64/sve-fixed-length-trunc-stores.ll
+0-1llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
+26-12 files

LLVM/project a4d841dclang/tools/cir-translate cir-translate.cpp CMakeLists.txt

handle cir-translate
DeltaFile
+3-3clang/tools/cir-translate/cir-translate.cpp
+1-1clang/tools/cir-translate/CMakeLists.txt
+4-42 files

LLVM/project abf77feclang/test/CodeGen scoped-atomic-ops.c, clang/test/CodeGenCUDA atomic-options.hip

Merge branch 'users/hvdijk/dxilprettyprinter-ir-printing' into users/hvdijk/directx-delay-converting-debug-info
DeltaFile
+1,547-723llvm/test/CodeGen/NVPTX/atomicrmw-sm60.ll
+1,521-697llvm/test/CodeGen/NVPTX/atomicrmw-sm70.ll
+568-852clang/test/CodeGen/scoped-atomic-ops.c
+867-373llvm/test/CodeGen/NVPTX/atomicrmw-sm90.ll
+144-216clang/test/CodeGenCUDA/atomic-options.hip
+256-63llvm/test/Transforms/LoopVectorize/early_exit_with_stores.ll
+4,903-2,924131 files not shown
+8,662-4,283137 files

LLVM/project de2cf58clang/test/CodeGen scoped-atomic-ops.c, clang/test/CodeGenCUDA atomic-options.hip

Merge branch 'main' into users/hvdijk/dxilprettyprinter-ir-printing
DeltaFile
+1,547-723llvm/test/CodeGen/NVPTX/atomicrmw-sm60.ll
+1,521-697llvm/test/CodeGen/NVPTX/atomicrmw-sm70.ll
+568-852clang/test/CodeGen/scoped-atomic-ops.c
+867-373llvm/test/CodeGen/NVPTX/atomicrmw-sm90.ll
+144-216clang/test/CodeGenCUDA/atomic-options.hip
+256-63llvm/test/Transforms/LoopVectorize/early_exit_with_stores.ll
+4,903-2,924131 files not shown
+8,662-4,283137 files

LLVM/project 0cea23aclang/test/Driver hipspv-toolchain.hip

[HIP] Fix another test for --no-lto (#201382)
DeltaFile
+2-1clang/test/Driver/hipspv-toolchain.hip
+2-11 files

LLVM/project 2d249cdllvm/docs GettingInvolved.rst

Remove incorrect gcal link from C/C++ language wg (#201374)

The C and C++ language working group meets on the first and third Wed of
the month, but Google Calendar does not support doing this via a single
event. Instead, we have one event for recurring on the 1st Wed and a
second event for recurring on the 3rd Wed. That means we cannot use a
single gcal link for the event. Instead of listing two links, this
removes the gcal link entirely because the meeting is also listed on the
community calendar itself. This reduces confusion for folks, but it
would be nice to get a replacement link at some point.
DeltaFile
+1-1llvm/docs/GettingInvolved.rst
+1-11 files

LLVM/project 12f60d0llvm/lib/Target/AMDGPU AMDGPUCodeGenPrepare.cpp

[AMDGPU] Remove unneeded early outs in getDivNumBits. NFC. (#201366)
DeltaFile
+0-4llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
+0-41 files

LLVM/project e97e676lldb/test/API/macosx/expedited-thread-pcs TestExpeditedThreadPCs.py

[lldb/test] Fix TestExpeditedThreadPCs on remote-darwin targets (#201275)
DeltaFile
+5-1lldb/test/API/macosx/expedited-thread-pcs/TestExpeditedThreadPCs.py
+5-11 files

LLVM/project 8ae895allvm/lib/Target/NVPTX NVPTXISelLowering.cpp, llvm/test/CodeGen/NVPTX atomicrmw-sm60.ll atomicrmw-sm70.ll

[NVPTX] Respect FTZ flag when lowering atomicrmw fadd. (#200732)

Previously we unconditionally lowered LLVM atomicrmw fadd to PTX
atom.add.  This is incorrect, because it ignores the FTZ behavior of the
LLVM and PTX instructions.
DeltaFile
+1,547-723llvm/test/CodeGen/NVPTX/atomicrmw-sm60.ll
+1,521-697llvm/test/CodeGen/NVPTX/atomicrmw-sm70.ll
+867-373llvm/test/CodeGen/NVPTX/atomicrmw-sm90.ll
+105-0llvm/test/CodeGen/NVPTX/atomicrmw-allow-ftz-atomics.ll
+54-12llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
+26-8llvm/test/CodeGen/NVPTX/atomics.ll
+4,120-1,8131 files not shown
+4,128-1,8147 files

LLVM/project 582d2fdllvm/lib/Transforms/Scalar LoopInterchange.cpp, llvm/test/Transforms/LoopInterchange function-attr.ll call-instructions.ll

[LoopInterchange] Bail out if function that may diverge is called (#201348)

This patch fixes the issue pointed out in
https://github.com/llvm/llvm-project/pull/200828#issuecomment-4593914293.
As demonstrated by the test cases added in #201331, it is not legal to
interchange loops that contain call instructions which may diverge. This
patch adds an additional check and bails out early when we cannot prove
that a call instruction in the loops doesn't diverge.
DeltaFile
+16-40llvm/test/Transforms/LoopInterchange/function-attr.ll
+7-3llvm/test/Transforms/LoopInterchange/call-instructions.ll
+4-2llvm/lib/Transforms/Scalar/LoopInterchange.cpp
+27-453 files

LLVM/project aa8e38fopenmp/module CMakeLists.txt

[OpenMP] FIx omp_lib.mod compilation for the GPU (#201377)
DeltaFile
+6-0openmp/module/CMakeLists.txt
+6-01 files