Page MenuHomePhabricator

foad (Jay Foad)
User

Projects

User does not belong to any projects.

User Details

User Since
Oct 29 2014, 9:58 AM (254 w, 6 d)

Recent Activity

Today

foad committed rGc92e51d84bb7: [SDA] Don't stop divergence propagation at the IPD. (authored by foad).
[SDA] Don't stop divergence propagation at the IPD.
Wed, Sep 18, 6:41 AM
foad committed rL372223: [SDA] Don't stop divergence propagation at the IPD..
[SDA] Don't stop divergence propagation at the IPD.
Wed, Sep 18, 6:41 AM
foad closed D65274: [SDA] Don't stop divergence propagation at the IPD..
Wed, Sep 18, 6:41 AM · Restricted Project

Yesterday

foad added a comment to D65274: [SDA] Don't stop divergence propagation at the IPD..

Ping! @simoll are you ready to land this now?

Tue, Sep 17, 9:10 AM · Restricted Project

Mon, Sep 16

foad added a comment to rL371955: AMDGPU/GlobalISel: Remove illegal select tests.

I am still getting this failure in a release build:

FAIL: LLVM :: CodeGen/AMDGPU/GlobalISel/inst-select-load-constant.mir (86 of 1760)
******************** TEST 'LLVM :: CodeGen/AMDGPU/GlobalISel/inst-select-load-constant.mir' FAILED ********************
Script:
--
: 'RUN: at line 2';   /home/jayfoad2/llvm-release/bin/llc -march=amdgcn -mcpu=tahiti -run-pass=instruction-select -verify-machineinstrs  -global-isel-abort=0 -o - /home/jayfoad2/git/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-constant.mir | /home/jayfoad2/llvm-release/bin/FileCheck -check-prefix=GFX6 /home/jayfoad2/git/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-constant.mir
: 'RUN: at line 3';   /home/jayfoad2/llvm-release/bin/llc -march=amdgcn -mcpu=hawaii -run-pass=instruction-select -verify-machineinstrs  -global-isel-abort=0 -o - /home/jayfoad2/git/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-constant.mir | /home/jayfoad2/llvm-release/bin/FileCheck -check-prefix=GFX7 /home/jayfoad2/git/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-constant.mir
: 'RUN: at line 4';   /home/jayfoad2/llvm-release/bin/llc -march=amdgcn -mcpu=fiji -run-pass=instruction-select -verify-machineinstrs -global-isel-abort=0 -o - /home/jayfoad2/git/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-constant.mir | /home/jayfoad2/llvm-release/bin/FileCheck -check-prefix=GFX8 /home/jayfoad2/git/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-constant.mir
: 'RUN: at line 5';   /home/jayfoad2/llvm-release/bin/llc -march=amdgcn -mcpu=gfx1010 -run-pass=instruction-select -verify-machineinstrs -global-isel-abort=0 -o - /home/jayfoad2/git/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-constant.mir | /home/jayfoad2/llvm-release/bin/FileCheck -check-prefix=GFX10 /home/jayfoad2/git/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-load-constant.mir
--
Exit Code: 1
Mon, Sep 16, 2:50 AM

Fri, Sep 13

foad added a comment to D63510: [LiveInterval] Removed bogus empty subrange assert.

The same change has now been committed in rL371792.

Fri, Sep 13, 7:17 AM · Restricted Project

Thu, Sep 12

foad added a comment to D51849: [RegisterCoalescer] Avoid "Use not jointly dominated by defs" in removePartialRedundancy.

Answering my own question: yes S_BUFFER_LOAD_DWORD_IMM needs an extra 0 operand because of a6322941ffc15275e837435eee7aabba0ca15211 and S_ENDPGM also needs a 0 operand as in 7df225dfc25adc8371188dc57f3adf300b0bd697.

Thu, Sep 12, 7:04 AM · Restricted Project
Herald added a project to D51849: [RegisterCoalescer] Avoid "Use not jointly dominated by defs" in removePartialRedundancy: Restricted Project.

If I apply this now, the new test fails with:

*** Bad machine code: Too few operands ***
- function:    _amdgpu_ps_main
- basic block: %bb.6  (0x53b7738)
- instruction: %8:sreg_32_xm0_xexec = S_BUFFER_LOAD_DWORD_IMM undef %9:sreg_128, 2708, 0 :: (dereferenceable invariant load 4)
5 operands expected, but 4 given.

Does this mean that the format of buffer load instructions has changed since you wrote this mir test?

Thu, Sep 12, 6:02 AM · Restricted Project

Fri, Sep 6

foad committed rG6c0204c794ba: [AMDGPU] Mark s_barrier as having side effects but not accessing memory. (authored by foad).
[AMDGPU] Mark s_barrier as having side effects but not accessing memory.
Fri, Sep 6, 3:08 AM
foad committed rL371192: [AMDGPU] Mark s_barrier as having side effects but not accessing memory..
[AMDGPU] Mark s_barrier as having side effects but not accessing memory.
Fri, Sep 6, 3:06 AM
foad closed D67218: [AMDGPU] Mark s_barrier as having side effects but not accessing memory..
Fri, Sep 6, 3:06 AM · Restricted Project
foad added inline comments to D67218: [AMDGPU] Mark s_barrier as having side effects but not accessing memory..
Fri, Sep 6, 3:06 AM · Restricted Project

Thu, Sep 5

foad updated subscribers of rL370667: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in….

Fixed in rL371000, thanks @rnk!

Thu, Sep 5, 3:32 AM
foad created D67218: [AMDGPU] Mark s_barrier as having side effects but not accessing memory..
Thu, Sep 5, 3:22 AM · Restricted Project

Wed, Sep 4

foad added a comment to rL370667: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in….

I have managed to build on Linux with EXPENSIVE_CHECKS but kludged to remove -D_GLIBCXX_DEBUG (because of problems) and I can't reproduce the failure. I have also done the same plus -DLLVM_REVERSE_ITERATION=ON and I still can't reproduce the failure.

Wed, Sep 4, 2:37 AM

Tue, Sep 3

foad added a comment to rL370667: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in….

Thanks. That is odd. I will investigate as soon as I am able to do an EXPENSIVE_CHECKS build.

Tue, Sep 3, 7:47 AM
foad added a comment to rL370667: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in….

@foad This has broken EXPENSIVE_CHECKS builds - can you take a look or should I revert?

Tue, Sep 3, 6:43 AM
foad committed rL370703: Add myself..
Add myself.
Tue, Sep 3, 12:40 AM

Mon, Sep 2

foad committed rG6e18266aa4dd: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in… (authored by foad).
Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in…
Mon, Sep 2, 7:41 AM
foad added a reverting change for D61491: AMDGPU: Be explicit about whether the high-word in SI_PC_ADD_REL_OFFSET is 0: rG6e18266aa4dd: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in….
Mon, Sep 2, 7:41 AM · Restricted Project
foad added a reverting change for D61491: AMDGPU: Be explicit about whether the high-word in SI_PC_ADD_REL_OFFSET is 0: rL370667: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in….
Mon, Sep 2, 7:41 AM · Restricted Project
foad committed rL370667: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in….
Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in…
Mon, Sep 2, 7:41 AM
foad closed D65813: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in SI_PC_ADD_REL_OFFSET is 0".
Mon, Sep 2, 7:40 AM · Restricted Project

Fri, Aug 23

foad committed rGeac23862a85f: [AMDGPU] gfx10 atomic optimizer changes. (authored by foad).
[AMDGPU] gfx10 atomic optimizer changes.
Fri, Aug 23, 3:08 AM
foad committed rL369745: [AMDGPU] gfx10 atomic optimizer changes..
[AMDGPU] gfx10 atomic optimizer changes.
Fri, Aug 23, 3:08 AM
foad closed D65644: [AMDGPU] gfx10 atomic optimizer changes..
Fri, Aug 23, 3:08 AM · Restricted Project
foad added a comment to D65813: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in SI_PC_ADD_REL_OFFSET is 0".

Ping!

Fri, Aug 23, 3:04 AM · Restricted Project

Aug 14 2019

foad added a comment to D65813: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in SI_PC_ADD_REL_OFFSET is 0".

Ping.

Aug 14 2019, 4:25 AM · Restricted Project

Aug 13 2019

foad added a comment to D65644: [AMDGPU] gfx10 atomic optimizer changes..

Ping.

Aug 13 2019, 1:48 AM · Restricted Project

Aug 7 2019

foad committed rG7d4ab7751d6c: [InstCombine] Add a TODO comment (authored by foad).
[InstCombine] Add a TODO comment
Aug 7 2019, 8:19 AM
foad committed rL368176: [InstCombine] Add a TODO comment.
[InstCombine] Add a TODO comment
Aug 7 2019, 8:18 AM
foad committed rG8e8b29583559: [InstCombine] Propagate fast math flags through selects (authored by foad).
[InstCombine] Propagate fast math flags through selects
Aug 7 2019, 8:18 AM
foad committed rL368175: [InstCombine] Propagate fast math flags through selects.
[InstCombine] Propagate fast math flags through selects
Aug 7 2019, 8:18 AM
foad closed D65658: [InstCombine] Propagate fast math flags through selects.
Aug 7 2019, 8:18 AM · Restricted Project
foad added a comment to D65658: [InstCombine] Propagate fast math flags through selects.

The code change seems fine, but the lone test does not provide enough coverage. Please have a look and rebase:
rL368028

Aug 7 2019, 4:08 AM · Restricted Project
foad updated the diff for D65658: [InstCombine] Propagate fast math flags through selects.

Rebase.

Aug 7 2019, 4:06 AM · Restricted Project
foad added a comment to D65813: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in SI_PC_ADD_REL_OFFSET is 0".

Could you please change shouldEmitConstantsToTextSection to check for == AMDPAL instead? This really doesn't make sense for Mesa... and then add a test that explicit sets amdpal in the triple?

Aug 7 2019, 3:23 AM · Restricted Project
foad updated the diff for D65813: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in SI_PC_ADD_REL_OFFSET is 0".

Only put constants in the text section on PAL.

Aug 7 2019, 3:21 AM · Restricted Project

Aug 6 2019

foad created D65813: Partially revert D61491 "AMDGPU: Be explicit about whether the high-word in SI_PC_ADD_REL_OFFSET is 0".
Aug 6 2019, 9:07 AM · Restricted Project
foad accepted D65274: [SDA] Don't stop divergence propagation at the IPD..
Aug 6 2019, 5:42 AM · Restricted Project
foad updated the diff for D65644: [AMDGPU] gfx10 atomic optimizer changes..

Add new hasDPPBroadcasts and hasDPPWavefrontShifts.
Use CreateCall instead of CreateIntrinsic in new helper functions.

Aug 6 2019, 2:31 AM · Restricted Project

Aug 5 2019

foad added inline comments to D65644: [AMDGPU] gfx10 atomic optimizer changes..
Aug 5 2019, 7:59 AM · Restricted Project
foad added a reviewer for D65644: [AMDGPU] gfx10 atomic optimizer changes.: nhaehnle.

I'm happy to spilt this up if the reviewers would like. There are a few NFC changes I could apply first, and/or I could try to split the wave32 changes out from the gfx10 dpp changes.

Aug 5 2019, 3:50 AM · Restricted Project

Aug 2 2019

foad added inline comments to D64713: [InstCombine] X *fast (C ? 1.0 : 0.0) -> C ? X : 0.0.
Aug 2 2019, 7:44 AM · Restricted Project
foad created D65658: [InstCombine] Propagate fast math flags through selects.
Aug 2 2019, 7:44 AM · Restricted Project
foad added a comment to D55153: [ThinLTO] Implement index-based WPD.

This test is failing for me:

Exit Code: 1
Aug 2 2019, 7:27 AM · Restricted Project
foad created D65644: [AMDGPU] gfx10 atomic optimizer changes..
Aug 2 2019, 3:35 AM · Restricted Project
foad committed rGb874b3d3faac: [LV] Fix test failure in a Release build. (authored by foad).
[LV] Fix test failure in a Release build.
Aug 2 2019, 1:36 AM
foad committed rL367666: [LV] Fix test failure in a Release build..
[LV] Fix test failure in a Release build.
Aug 2 2019, 1:33 AM

Jul 31 2019

foad added a comment to D64790: [LLVM][NFC] Adding an Alignment type to LLVM.

This also broke ninja check-llvm-unit on a Release build:

Failing Tests (7):
    LLVM-Unit :: Support/./SupportTests/Alignment.CantConvertUnsetMaybe
    LLVM-Unit :: Support/./SupportTests/Alignment.CompareAlignToUndefMaybeAlign
    LLVM-Unit :: Support/./SupportTests/Alignment.CompareMaybeAlignToZero
    LLVM-Unit :: Support/./SupportTests/Alignment.ComparisonsWithZero
    LLVM-Unit :: Support/./SupportTests/Alignment.Division
    LLVM-Unit :: Support/./SupportTests/Alignment.InvalidCTors
    LLVM-Unit :: Support/./SupportTests/Alignment.Log2

Typical failure:

FAIL: LLVM-Unit :: Support/./SupportTests/Alignment.Division (2696 of 3890)
******************** TEST 'LLVM-Unit :: Support/./SupportTests/Alignment.Division' FAILED ********************
Note: Google Test filter = Alignment.Division
[==========] Running 1 test from 1 test case.
[----------] Global test environment set-up.
[----------] 1 test from Alignment
[ RUN      ] Alignment.Division
/home/jayfoad2/git/llvm-project/llvm/unittests/Support/AlignmentTest.cpp:70: Failure
Death test: Align(Value) / 2
    Result: failed to die.
 Error msg:
[  DEATH   ] 
/home/jayfoad2/git/llvm-project/llvm/unittests/Support/AlignmentTest.cpp:71: Failure
Death test: MaybeAlign(Value) / 2
    Result: failed to die.
 Error msg:
[  DEATH   ] 
/home/jayfoad2/git/llvm-project/llvm/unittests/Support/AlignmentTest.cpp:79: Failure
Death test: Align(8) / 0
    Result: failed to die.
 Error msg:
[  DEATH   ] 
/home/jayfoad2/git/llvm-project/llvm/unittests/Support/AlignmentTest.cpp:80: Failure
Death test: Align(8) / 3
    Result: failed to die.
 Error msg:
[  DEATH   ] 
[  FAILED  ] Alignment.Division (2 ms)
[----------] 1 test from Alignment (2 ms total)
Jul 31 2019, 5:10 AM · Restricted Project

Jul 29 2019

foad committed rG3bdcedbf3d6a: [AMDGPU] Fix typo in error message (authored by foad).
[AMDGPU] Fix typo in error message
Jul 29 2019, 9:20 AM
foad committed rL367235: [AMDGPU] Fix typo in error message.
[AMDGPU] Fix typo in error message
Jul 29 2019, 9:20 AM
foad committed rGdcb75324794e: [DivergenceAnalysis] Add methods for querying divergence at use (authored by foad).
[DivergenceAnalysis] Add methods for querying divergence at use
Jul 29 2019, 7:55 AM
foad committed rG2bd9da8a72d5: [AMDGPU] Add amdgpu_kernel for consistency with other tests (authored by foad).
[AMDGPU] Add amdgpu_kernel for consistency with other tests
Jul 29 2019, 7:55 AM
foad committed rL367221: [AMDGPU] Add amdgpu_kernel for consistency with other tests.
[AMDGPU] Add amdgpu_kernel for consistency with other tests
Jul 29 2019, 7:39 AM
foad committed rL367218: [DivergenceAnalysis] Add methods for querying divergence at use.
[DivergenceAnalysis] Add methods for querying divergence at use
Jul 29 2019, 7:39 AM
foad closed D65141: [DivergenceAnalysis] Add methods for querying divergence at use.
Jul 29 2019, 7:39 AM · Restricted Project

Jul 26 2019

foad added a comment to D65336: [InstSimplify] remove quadratic time looping (PR42771).

Looks OK to me now, thanks.

Jul 26 2019, 1:00 PM · Restricted Project
foad added inline comments to D65336: [InstSimplify] remove quadratic time looping (PR42771).
Jul 26 2019, 10:05 AM · Restricted Project
foad added inline comments to D65336: [InstSimplify] remove quadratic time looping (PR42771).
Jul 26 2019, 9:29 AM · Restricted Project
foad added a comment to D65336: [InstSimplify] remove quadratic time looping (PR42771).

There's a test diff here for a degenerate case with llvm.assume that I didn't step through, but I think that's caused because isInstructionTriviallyDead() may not be true for a function call even if that call has no uses.

Jul 26 2019, 9:14 AM · Restricted Project

Jul 25 2019

foad added a comment to D65274: [SDA] Don't stop divergence propagation at the IPD..

I can confirm that this fixes the original test case that I had, that I reduced into bug 42706, and it doesn't seem to cause any other problems in my other test cases.

Jul 25 2019, 8:28 AM · Restricted Project
foad added a comment to D65141: [DivergenceAnalysis] Add methods for querying divergence at use.

To @arsenm 's question: at least for the SelectionDAG, the key is in SITargetLowering::isSDNodeSourceOfDivergence, which calls isDivergent on a value.

The way the SelectionDAG works, we have in our hands at this point a live value that comes into a basic block via a register. This suggests an isDivergentInBB.

Jul 25 2019, 8:14 AM · Restricted Project

Jul 24 2019

foad added a comment to D65141: [DivergenceAnalysis] Add methods for querying divergence at use.

Does this avoid the need to run LCSSA for isel, assuming the DAG switches to checking the use instead of the def?

Jul 24 2019, 9:07 AM · Restricted Project
foad committed rG565c54320e00: [InstSimplify] Rename SimplifyFPUnOp and SimplifyFPBinOp (authored by foad).
[InstSimplify] Rename SimplifyFPUnOp and SimplifyFPBinOp
Jul 24 2019, 5:51 AM
foad committed rL366902: [InstSimplify] Rename SimplifyFPUnOp and SimplifyFPBinOp.
[InstSimplify] Rename SimplifyFPUnOp and SimplifyFPBinOp
Jul 24 2019, 5:49 AM
foad closed D64902: [InstSimplify] Rename SimplifyFPUnOp and SimplifyFPBinOp.
Jul 24 2019, 5:49 AM · Restricted Project
foad added a comment to D65141: [DivergenceAnalysis] Add methods for querying divergence at use.

It's missing tests though -- I'm not sure if we want to print this additional information in the divergence printer

Jul 24 2019, 4:11 AM · Restricted Project
foad updated the diff for D65141: [DivergenceAnalysis] Add methods for querying divergence at use.

Address review comments.

Jul 24 2019, 4:10 AM · Restricted Project

Jul 23 2019

foad created D65141: [DivergenceAnalysis] Add methods for querying divergence at use.
Jul 23 2019, 6:48 AM · Restricted Project

Jul 22 2019

foad added inline comments to D64902: [InstSimplify] Rename SimplifyFPUnOp and SimplifyFPBinOp.
Jul 22 2019, 8:10 AM · Restricted Project
foad added a comment to D64902: [InstSimplify] Rename SimplifyFPUnOp and SimplifyFPBinOp.

@spatel since no-one else has commented, is this still OK to commit? Should I do anything to warn out-of-tree users about the API name change?

Jul 22 2019, 8:05 AM · Restricted Project
foad added inline comments to D65088: [AMDGPU][RFC] New llvm.amdgcn.ballot intrinsic.
Jul 22 2019, 7:46 AM · Restricted Project
foad updated the diff for D65088: [AMDGPU][RFC] New llvm.amdgcn.ballot intrinsic.

Address review comments.

Jul 22 2019, 7:45 AM · Restricted Project
foad created D65088: [AMDGPU][RFC] New llvm.amdgcn.ballot intrinsic.
Jul 22 2019, 6:38 AM · Restricted Project
foad committed rG298500ae3319: [AMDGPU] Save some work when an atomic op has no uses (authored by foad).
[AMDGPU] Save some work when an atomic op has no uses
Jul 22 2019, 12:24 AM
foad committed rL366667: [AMDGPU] Save some work when an atomic op has no uses.
[AMDGPU] Save some work when an atomic op has no uses
Jul 22 2019, 12:19 AM
foad closed D64981: [AMDGPU] Save some work when an atomic op has no uses.
Jul 22 2019, 12:19 AM · Restricted Project
foad added a comment to rL366585: AMDGPU/GlobalISel: Selection for fminnum/fmaxnum.

The added tests are failing for me, both on NetBSD and on one of my Linux systems:
http://lab.llvm.org:8011/builders/netbsd-amd64/builds/21070/steps/run%20unit%20tests

Jul 22 2019, 12:15 AM

Jul 19 2019

foad added a comment to D64981: [AMDGPU] Save some work when an atomic op has no uses.

Test?

Jul 19 2019, 10:14 AM · Restricted Project
foad added a comment to D64981: [AMDGPU] Save some work when an atomic op has no uses.

Follow-on patches will also improve the generated code when the result is not needed.

Jul 19 2019, 2:06 AM · Restricted Project
foad created D64981: [AMDGPU] Save some work when an atomic op has no uses.
Jul 19 2019, 2:06 AM · Restricted Project
foad committed rG7d06ffff466d: [AMDGPU] Simplify the exclusive scan used for optimized atomics (authored by foad).
[AMDGPU] Simplify the exclusive scan used for optimized atomics
Jul 19 2019, 1:44 AM
foad committed rL366543: [AMDGPU] Simplify the exclusive scan used for optimized atomics.
[AMDGPU] Simplify the exclusive scan used for optimized atomics
Jul 19 2019, 1:40 AM
foad closed D64411: [AMDGPU] Simplify the exclusive scan used for optimized atomics.
Jul 19 2019, 1:40 AM · Restricted Project

Jul 17 2019

foad added inline comments to D64713: [InstCombine] X *fast (C ? 1.0 : 0.0) -> C ? X : 0.0.
Jul 17 2019, 11:07 PM · Restricted Project
foad created D64902: [InstSimplify] Rename SimplifyFPUnOp and SimplifyFPBinOp.
Jul 17 2019, 11:06 PM · Restricted Project
foad added inline comments to D64713: [InstCombine] X *fast (C ? 1.0 : 0.0) -> C ? X : 0.0.
Jul 17 2019, 10:44 PM · Restricted Project
foad added inline comments to D64713: [InstCombine] X *fast (C ? 1.0 : 0.0) -> C ? X : 0.0.
Jul 17 2019, 12:40 PM · Restricted Project
foad committed rG70235c642e66: [AMDGPU] Optimize atomic AND/OR/XOR (authored by foad).
[AMDGPU] Optimize atomic AND/OR/XOR
Jul 17 2019, 6:44 AM
foad committed rL366323: [AMDGPU] Optimize atomic AND/OR/XOR.
[AMDGPU] Optimize atomic AND/OR/XOR
Jul 17 2019, 6:39 AM
foad closed D64809: [AMDGPU] Optimize atomic AND/OR/XOR.
Jul 17 2019, 6:39 AM · Restricted Project

Jul 16 2019

foad added inline comments to D64328: [AMDGPU] Optimize atomic max/min.
Jul 16 2019, 10:53 AM · Restricted Project
foad created D64809: [AMDGPU] Optimize atomic AND/OR/XOR.
Jul 16 2019, 10:53 AM · Restricted Project
foad committed rG17060f0a54b6: [AMDGPU] Optimize atomic max/min (authored by foad).
[AMDGPU] Optimize atomic max/min
Jul 16 2019, 10:47 AM
foad committed rL366235: [AMDGPU] Optimize atomic max/min.
[AMDGPU] Optimize atomic max/min
Jul 16 2019, 10:45 AM
foad closed D64328: [AMDGPU] Optimize atomic max/min.
Jul 16 2019, 10:44 AM · Restricted Project
foad added a comment to D64328: [AMDGPU] Optimize atomic max/min.

Ping?

Jul 16 2019, 9:31 AM · Restricted Project
foad added inline comments to D64713: [InstCombine] X *fast (C ? 1.0 : 0.0) -> C ? X : 0.0.
Jul 16 2019, 1:25 AM · Restricted Project
foad added a comment to rL366121: AMDGPU/GlobalISel: Select G_AND/G_OR/G_XOR.

The new tests are failing in my Release build:

Failing Tests (3):
    LLVM :: CodeGen/AMDGPU/GlobalISel/inst-select-and.mir
    LLVM :: CodeGen/AMDGPU/GlobalISel/inst-select-or.mir
    LLVM :: CodeGen/AMDGPU/GlobalISel/inst-select-xor.mir

The failures all look like:

/home/jayfoad2/git/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-and.mir:120:12: error: WAVE64: expected string not found in input
 ; WAVE64: [[COPY:%[0-9]+]]:sgpr(s32) = COPY $sgpr0
           ^
<stdin>:375:2: note: scanning from here
 %0:sreg_32_xm0 = COPY $sgpr0
 ^
Jul 16 2019, 1:23 AM