Page MenuHomePhabricator

Recent Activity

Today

spatel added a comment to D102002: [PassManager] unify vector passes between regular and LTO pipelines.

Without a lot of evidence that this is better across a range of architectures, I would recommend trying to be conservative, taking things a step at a time.
We have one test where this patch would make it 88% worse. That might be a pathological case but no one is going to put up with on tenth of the speed :-)

Fri, May 7, 4:12 AM · Restricted Project
ABataev added inline comments to D98714: [SLP] Add insertelement instructions to vectorizable tree.
Fri, May 7, 4:09 AM · Restricted Project
sdesmalen added a comment to D102053: [AArch64][SVE] Fix isel failure for FP-extending loads.

Cheers @bsmith and @kmclaughlin!

Fri, May 7, 4:06 AM · Restricted Project
sdesmalen updated the diff for D101945: [LV] Add -scalable-vectorization=<option> flag..
  • Added SK_Undefined to fix the case where 'vectorize.width' is specified, but 'vectorize.scalable.enable' is not. In the previous revision of this patch, it would default to the value of the flag, which when set to 'preferred' would cause loops with #pragma clang loop vectorize_width(4) to be vectorized with VF=vscale x 4. This is now resolved (and tested).
  • Also renamed disableScalableVectorization -> isScalableVectorizationDisabled.
Fri, May 7, 4:04 AM · Restricted Project
OikawaKirie requested review of D102062: [analyzer][ctu] Append ctu-dir to ctu-invocation-list for non-absolute paths.
Fri, May 7, 4:04 AM · Restricted Project
I_Perry requested review of D102061: [flang] [openmp] Add Fortran specific semantic checks for OpenMP Allocate directive..
Fri, May 7, 3:59 AM · Restricted Project
lebedev.ri committed rGbda9ca3e44c1: [NFC][X86][MCA] AMD Zen 3: add tests with non-eliminatible MMX moves (authored by lebedev.ri).
[NFC][X86][MCA] AMD Zen 3: add tests with non-eliminatible MMX moves
Fri, May 7, 3:57 AM
lebedev.ri committed rG7059b28d5d27: [X86] AMD Zen 3: 32/64 -bit GPR register moves are zero-cycle (authored by lebedev.ri).
[X86] AMD Zen 3: 32/64 -bit GPR register moves are zero-cycle
Fri, May 7, 3:56 AM
lebedev.ri committed rG227678089cf6: [NFC][X86][MCA] AMD Zen 3: add tests with eliminatible GPR moves (authored by lebedev.ri).
[NFC][X86][MCA] AMD Zen 3: add tests with eliminatible GPR moves
Fri, May 7, 3:56 AM
ftynse closed D101999: Add a test case for attributes of linalg.init_tensor.

This was landed as 1f109f9d9cddbc90d97b50c154a8474e7e623356

Fri, May 7, 3:53 AM · Restricted Project
anton-afanasyev added inline comments to D98714: [SLP] Add insertelement instructions to vectorizable tree.
Fri, May 7, 3:52 AM · Restricted Project
ikudrin added inline comments to D101996: [LLD] Improve reporting unresolved symbols in shared libraries.
Fri, May 7, 3:52 AM · Restricted Project, lld
Chia-hungDuan updated the diff for D101607: [mlir-reduce] Create MlirReduceLib.

Amend previous uploaded patches which was an accident.

Fri, May 7, 3:52 AM · Restricted Project
ikudrin updated the diff for D101996: [LLD] Improve reporting unresolved symbols in shared libraries.
  • Updated a comment for SharedFile::requiredSymbols
  • Moved the test into allow-shlib-undefined.s
  • Added a test to check that the error is reported for each shared library where the symbol is referenced
Fri, May 7, 3:51 AM · Restricted Project, lld
kmclaughlin accepted D102053: [AArch64][SVE] Fix isel failure for FP-extending loads.

LGTM!

Fri, May 7, 3:51 AM · Restricted Project
StephenTozer committed rG0791f968fee2: [DebugInfo] Fix updateDbgUsersToReg to support DBG_VALUE_LIST (authored by StephenTozer).
[DebugInfo] Fix updateDbgUsersToReg to support DBG_VALUE_LIST
Fri, May 7, 3:51 AM
bsmith accepted D102053: [AArch64][SVE] Fix isel failure for FP-extending loads.

Does the same issue also apply to fixed length codegen?

Existing code (in same function under hasNeon()) already sets this to Expand for all fixed-width vectors.

Fri, May 7, 3:51 AM · Restricted Project
StephenTozer closed D101523: [DebugInfo] Fix updateDbgUsersToReg to support DBG_VALUE_LIST.
Fri, May 7, 3:51 AM · Restricted Project, debug-info
Chia-hungDuan updated the diff for D101046: [mlir-reduce] Reducer refactor..

Added comments

Fri, May 7, 3:50 AM · Restricted Project
sdesmalen added a comment to D102053: [AArch64][SVE] Fix isel failure for FP-extending loads.

Does the same issue also apply to fixed length codegen?

Existing code (in same function under hasNeon()) already sets this to Expand for all fixed-width vectors.

Fri, May 7, 3:46 AM · Restricted Project
dfki-jugr added a comment to D98041: [MLIR] Create memref dialect and move dialect-specific ops from std..

@mehdi_amini Try to fix the issue here: https://reviews.llvm.org/D102060

Fri, May 7, 3:43 AM · Restricted Project
Leporacanthicus added a comment to D102008: [OpenMP]Add support for workshare loop modifier in lowering.

Patch doesn't apply because it is relying on my previous patch. Doh!

Fri, May 7, 3:42 AM · Restricted Project
dfki-jugr added a reviewer for D102060: [WIP][mlir] Resolve memref dependency in canonicalize pass.: mehdi_amini.
Fri, May 7, 3:42 AM · Restricted Project
dfki-jugr requested review of D102060: [WIP][mlir] Resolve memref dependency in canonicalize pass..
Fri, May 7, 3:41 AM · Restricted Project
Chia-hungDuan updated the diff for D101607: [mlir-reduce] Create MlirReduceLib.

Added comment

Fri, May 7, 3:41 AM · Restricted Project
ABataev added inline comments to D98714: [SLP] Add insertelement instructions to vectorizable tree.
Fri, May 7, 3:36 AM · Restricted Project
bsmith added a comment to D102053: [AArch64][SVE] Fix isel failure for FP-extending loads.

Does the same issue also apply to fixed length codegen?

Fri, May 7, 3:35 AM · Restricted Project
ABataev added inline comments to D98714: [SLP] Add insertelement instructions to vectorizable tree.
Fri, May 7, 3:34 AM · Restricted Project
dfki-mako added a comment to D100838: [mlir] Split BufferAliasAnalysis into 2 classes.

I've worked around this issue in my own code now, so what do you think about me just doing this rename? Then there will be no confusion between this pass and the new AliasAnalysis (and their definitions of aliasing). I think that confusion is good to avoid, and since I have a different solution for my own use case, I can avoid creating more confusion by creating another Analysis at this time.

+1 to renaming it to avoid confusion in the future.

Fri, May 7, 3:32 AM · Restricted Project
anton-afanasyev added inline comments to D98714: [SLP] Add insertelement instructions to vectorizable tree.
Fri, May 7, 3:31 AM · Restricted Project
gbreynoo committed rGf0762fc42f0f: [llvm-dwarfdump] Help option output should be consistent with the command guide (authored by gbreynoo).
[llvm-dwarfdump] Help option output should be consistent with the command guide
Fri, May 7, 3:23 AM
gbreynoo closed D101646: [llvm-dwarfdump] Help option output should be consistent with the command guide.
Fri, May 7, 3:23 AM · Restricted Project
gchatelet committed rGe805b7c2d63c: [llvm][NFC] Remove remaining deprecated alignment functions from CodeGen (authored by gchatelet).
[llvm][NFC] Remove remaining deprecated alignment functions from CodeGen
Fri, May 7, 3:23 AM
gchatelet closed D102058: [llvm][NFC] Remove remaining deprecated alignment functions from CodeGen.
Fri, May 7, 3:22 AM · Restricted Project
gchatelet committed rGeb1b26ec1d1a: [llvm][NFC] Remove deprecated TargetFrameLowering and InstrTypes alignment… (authored by gchatelet).
[llvm][NFC] Remove deprecated TargetFrameLowering and InstrTypes alignment…
Fri, May 7, 3:22 AM
gchatelet closed D102056: [llvm][NFC] Remove deprecated TargetFrameLowering and InstrTypes alignment functions.
Fri, May 7, 3:21 AM · Restricted Project
anton-afanasyev added inline comments to D98714: [SLP] Add insertelement instructions to vectorizable tree.
Fri, May 7, 3:20 AM · Restricted Project
Petar.Avramovic added inline comments to D101992: AMDGPU/GlobalISel: Use destination register bank in applyMappingLoad.
Fri, May 7, 3:20 AM · Restricted Project
foad added inline comments to D101992: AMDGPU/GlobalISel: Use destination register bank in applyMappingLoad.
Fri, May 7, 3:15 AM · Restricted Project
OikawaKirie updated the diff for D101763: [analyzer][ctu] Avoid parsing invocation list again and again during on-demand parsing of CTU.

By inspecting the output that is piped to the count, I have a suspicion:

openat(AT_FDCWD, "invocations.yaml", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
Fri, May 7, 3:14 AM · Restricted Project