Page MenuHomePhabricator

dstuttard (David Stuttard)
User

Projects

User does not belong to any projects.

User Details

User Since
Jan 25 2017, 7:29 AM (277 w, 5 d)

Recent Activity

Thu, May 19

dstuttard accepted D125976: [AMDGPU] Allow multiple uses of the same literal in SOP2/SOPC.

LGTM

Thu, May 19, 8:30 AM · Restricted Project, Restricted Project

Mar 29 2022

dstuttard added inline comments to D122653: [AMDGPU] Document the intended semantics of llvm.amdgcn.s.buffer.load.
Mar 29 2022, 6:38 AM · Restricted Project, Restricted Project

Mar 15 2022

dstuttard updated the diff for D121691: [AMDGPU] Fix kill flag state for replaced operands.

Pushed an earlier commit by mistake

Mar 15 2022, 6:12 AM · Restricted Project, Restricted Project
dstuttard added reviewers for D121691: [AMDGPU] Fix kill flag state for replaced operands: rampitec, cdevadas.
Mar 15 2022, 6:06 AM · Restricted Project, Restricted Project
dstuttard added a comment to D121691: [AMDGPU] Fix kill flag state for replaced operands.

Picked this issue up in some testing - seems a logical fix.
I have a reproducer, but even the cut-down isn't that clean - do you think this requires a test?

Mar 15 2022, 6:05 AM · Restricted Project, Restricted Project
dstuttard requested review of D121691: [AMDGPU] Fix kill flag state for replaced operands.
Mar 15 2022, 6:01 AM · Restricted Project, Restricted Project

Dec 2 2021

dstuttard committed rG0e8590f06599: [AMDGPU] Add support for in-order bvh in waitcnt pass (authored by dstuttard).
[AMDGPU] Add support for in-order bvh in waitcnt pass
Dec 2 2021, 6:27 AM
dstuttard closed D114794: [AMDGPU] Add support for in-order bvh in waitcnt pass.
Dec 2 2021, 6:26 AM · Restricted Project
dstuttard committed rG17a3385c3538: [AMDGPU] Test for in-order waitcnt insertion for bvh instructions (authored by dstuttard).
[AMDGPU] Test for in-order waitcnt insertion for bvh instructions
Dec 2 2021, 6:26 AM
dstuttard closed D114792: [AMDGPU] Test for in-order waitcnt insertion for bvh instructions.
Dec 2 2021, 6:26 AM · Restricted Project

Nov 30 2021

dstuttard updated the diff for D114794: [AMDGPU] Add support for in-order bvh in waitcnt pass.

Spelling mistake (and rebase for test changes)

Nov 30 2021, 6:24 AM · Restricted Project
dstuttard updated the diff for D114792: [AMDGPU] Test for in-order waitcnt insertion for bvh instructions.

Updating test in light of review suggestions

Nov 30 2021, 6:22 AM · Restricted Project
dstuttard added inline comments to D114794: [AMDGPU] Add support for in-order bvh in waitcnt pass.
Nov 30 2021, 6:08 AM · Restricted Project
dstuttard added inline comments to D114792: [AMDGPU] Test for in-order waitcnt insertion for bvh instructions.
Nov 30 2021, 6:08 AM · Restricted Project
dstuttard added reviewers for D114794: [AMDGPU] Add support for in-order bvh in waitcnt pass: foad, rampitec.
Nov 30 2021, 6:05 AM · Restricted Project
dstuttard updated the diff for D114792: [AMDGPU] Test for in-order waitcnt insertion for bvh instructions.

Updating with link to change for which this is a pre-commit

Nov 30 2021, 6:03 AM · Restricted Project
dstuttard requested review of D114794: [AMDGPU] Add support for in-order bvh in waitcnt pass.
Nov 30 2021, 6:00 AM · Restricted Project
dstuttard added reviewers for D114792: [AMDGPU] Test for in-order waitcnt insertion for bvh instructions: foad, rampitec.

Pre-commit test

Nov 30 2021, 5:58 AM · Restricted Project
dstuttard requested review of D114792: [AMDGPU] Test for in-order waitcnt insertion for bvh instructions.
Nov 30 2021, 5:54 AM · Restricted Project

Oct 19 2021

dstuttard closed D111979: [llvm-shlib] Fix windows build failed while llvm non-standalone building..

Commited as cf65271e462f0626157a06f2f3a9615b4987c24e

Oct 19 2021, 1:28 AM · Restricted Project
dstuttard committed rGcf65271e462f: [llvm-shlib] Fix windows build failed while llvm non-standalone building. (authored by Qiaojin.Bao <qiaojin.bao@amd.com>).
[llvm-shlib] Fix windows build failed while llvm non-standalone building.
Oct 19 2021, 1:21 AM

Oct 18 2021

dstuttard added reviewers for D111979: [llvm-shlib] Fix windows build failed while llvm non-standalone building.: phosek, MaskRay.
Oct 18 2021, 3:01 AM · Restricted Project
dstuttard requested review of D111979: [llvm-shlib] Fix windows build failed while llvm non-standalone building..
Oct 18 2021, 2:52 AM · Restricted Project

Oct 12 2021

dstuttard added a comment to D111646: [AMDGPU] Enable load clustering in the post-RA scheduler.

Interesting to see that quite a few of the test changes are clustering the loads, but not ending up with a clause (correctly).
Nice result otherwise though.

Oct 12 2021, 8:11 AM · Restricted Project
dstuttard added a comment to D111646: [AMDGPU] Enable load clustering in the post-RA scheduler.

Interesting to see that quite a few of the test changes are clustering the loads, but not ending up with a clause (correctly).
Nice result otherwise though.

Oct 12 2021, 8:02 AM · Restricted Project

Oct 8 2021

dstuttard committed rG69f7d81d0a47: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually… (authored by dstuttard).
[AMDGPU] Set number vgprs used in PS shaders based on input registers actually…
Oct 8 2021, 6:25 AM
dstuttard closed D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used.
Oct 8 2021, 6:25 AM · Restricted Project
dstuttard added inline comments to D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used.
Oct 8 2021, 2:19 AM · Restricted Project
dstuttard added inline comments to D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used.
Oct 8 2021, 1:14 AM · Restricted Project
dstuttard updated the diff for D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used.

Address review comments

Oct 8 2021, 1:01 AM · Restricted Project

Oct 7 2021

dstuttard updated subscribers of D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used.

@critson @foad @Flakebi - are you happy with the changes? Can someone approve?

Oct 7 2021, 1:44 AM · Restricted Project

Sep 28 2021

dstuttard added inline comments to D110579: [AMDGPU] Add a new intrinsic to control fp_trunc rounding mode.
Sep 28 2021, 1:29 AM · Restricted Project, Restricted Project

Sep 27 2021

dstuttard added a comment to D103014: [lit] Attempt for fix tests failing because of 'warning: non-portable path to file'.

@dstuttard, could you please give an example that reproduces the issue?

Sep 27 2021, 4:28 AM · Restricted Project

Sep 23 2021

dstuttard added a comment to D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used.

ping?

Sep 23 2021, 9:06 AM · Restricted Project
dstuttard added a comment to D103014: [lit] Attempt for fix tests failing because of 'warning: non-portable path to file'.

This change has caused some problems for us - resolve doesn't respect the use of subst on windows which we've been using to avoid long file path problems. This is a particular issue for the llvm-lit tests.

Sep 23 2021, 9:01 AM · Restricted Project

Sep 21 2021

dstuttard added inline comments to D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used.
Sep 21 2021, 8:02 AM · Restricted Project
dstuttard updated the diff for D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used.

Rebase and address some of the review comments

Sep 21 2021, 8:01 AM · Restricted Project

Aug 17 2021

dstuttard committed rGebdb0d09a4f4: AMDGPU: During img instruction ret value construction cater for non int values (authored by dstuttard).
AMDGPU: During img instruction ret value construction cater for non int values
Aug 17 2021, 1:09 AM
dstuttard closed D108131: AMDGPU: During img instruction ret value construction cater for non int values.
Aug 17 2021, 1:09 AM · Restricted Project

Aug 16 2021

dstuttard added reviewers for D108131: AMDGPU: During img instruction ret value construction cater for non int values: Joe_Nash, arsenm.

We ran across this issue a while back for graphics, but I'd forgotten to upstream it.

Aug 16 2021, 7:47 AM · Restricted Project
dstuttard requested review of D108131: AMDGPU: During img instruction ret value construction cater for non int values.
Aug 16 2021, 7:45 AM · Restricted Project

Jul 20 2021

dstuttard accepted D106288: [AMDGPU] Disable LDS lowering for GFX shaders.

Thanks Stas - LGTM

Jul 20 2021, 2:30 AM · Restricted Project

Jul 14 2021

dstuttard added a comment to D81070: [AMDGPU] Fixup use of StackPtrOffsetReg when not initialized.

I'm revisiting this - (sorry for the lengthy hiatus)

Jul 14 2021, 3:57 AM · Restricted Project

Jul 9 2021

dstuttard added a comment to D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used.

Do you think that I should add any other reviewers?

Jul 9 2021, 4:27 AM · Restricted Project

Jul 5 2021

dstuttard committed rG83cb9632a13d: [DAGCombiner] Add support for mulhi const folding in DAGCombiner (authored by dstuttard).
[DAGCombiner] Add support for mulhi const folding in DAGCombiner
Jul 5 2021, 4:08 AM
dstuttard closed D103323: [DAGCombiner] Add support for mulhi const folding in DAGCombiner.
Jul 5 2021, 4:08 AM · Restricted Project
dstuttard committed rG4b125b23ba95: [DAGCombiner] Pre-commit test to demonstrate mulhi const folding (authored by dstuttard).
[DAGCombiner] Pre-commit test to demonstrate mulhi const folding
Jul 5 2021, 3:35 AM
dstuttard closed D105424: [DAGCombiner] Pre-commit test to demonstrate mulhi const folding.
Jul 5 2021, 3:35 AM · Restricted Project
dstuttard added inline comments to D103323: [DAGCombiner] Add support for mulhi const folding in DAGCombiner.
Jul 5 2021, 3:28 AM · Restricted Project
dstuttard updated the diff for D103323: [DAGCombiner] Add support for mulhi const folding in DAGCombiner.

Pre-committed test and rebased on top

Jul 5 2021, 3:27 AM · Restricted Project
dstuttard added reviewers for D105424: [DAGCombiner] Pre-commit test to demonstrate mulhi const folding: RKSimon, foad.
Jul 5 2021, 3:26 AM · Restricted Project
dstuttard requested review of D105424: [DAGCombiner] Pre-commit test to demonstrate mulhi const folding.
Jul 5 2021, 3:24 AM · Restricted Project
dstuttard committed rGb8173c317812: [AMDGPU] Stop mulhi from doing 24 bit mul for uniform values (authored by dstuttard).
[AMDGPU] Stop mulhi from doing 24 bit mul for uniform values
Jul 5 2021, 2:34 AM
dstuttard closed D103321: [AMDGPU] Stop mulhi from doing 24 bit mul for uniform values.
Jul 5 2021, 2:34 AM · Restricted Project
dstuttard added inline comments to D103323: [DAGCombiner] Add support for mulhi const folding in DAGCombiner.
Jul 5 2021, 1:50 AM · Restricted Project
dstuttard added a comment to D103323: [DAGCombiner] Add support for mulhi const folding in DAGCombiner.

@dstuttard Please can you rebase? rGcc38f8939da4aec85e7d0ef4de412e30d4de5a14 should give you vector coverage

Jul 5 2021, 1:48 AM · Restricted Project
dstuttard updated the diff for D103323: [DAGCombiner] Add support for mulhi const folding in DAGCombiner.

Updated existing test based on feedback

Jul 5 2021, 1:47 AM · Restricted Project

Jun 21 2021

dstuttard updated the diff for D103321: [AMDGPU] Stop mulhi from doing 24 bit mul for uniform values.

Adding mul_uint24-amdgcn.ll gfx9 variant

Jun 21 2021, 7:07 AM · Restricted Project
dstuttard added inline comments to D103321: [AMDGPU] Stop mulhi from doing 24 bit mul for uniform values.
Jun 21 2021, 6:07 AM · Restricted Project

Jun 18 2021

dstuttard added inline comments to D103321: [AMDGPU] Stop mulhi from doing 24 bit mul for uniform values.
Jun 18 2021, 9:17 AM · Restricted Project
dstuttard added inline comments to D103321: [AMDGPU] Stop mulhi from doing 24 bit mul for uniform values.
Jun 18 2021, 3:39 AM · Restricted Project
dstuttard updated the diff for D103323: [DAGCombiner] Add support for mulhi const folding in DAGCombiner.

Add test

Jun 18 2021, 3:22 AM · Restricted Project

Jun 17 2021

dstuttard updated the diff for D103323: [DAGCombiner] Add support for mulhi const folding in DAGCombiner.

Updating for review comments.

Jun 17 2021, 1:14 AM · Restricted Project

Jun 16 2021

dstuttard added inline comments to D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used.
Jun 16 2021, 2:56 AM · Restricted Project
dstuttard updated the diff for D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used.

Some modifications based on review feedback

Jun 16 2021, 2:56 AM · Restricted Project

May 28 2021

dstuttard updated the diff for D103323: [DAGCombiner] Add support for mulhi const folding in DAGCombiner.

Thanks for reviews.
Made suggested changes.

May 28 2021, 9:14 AM · Restricted Project
dstuttard added reviewers for D103323: [DAGCombiner] Add support for mulhi const folding in DAGCombiner: RKSimon, MatzeB, foad.
May 28 2021, 8:43 AM · Restricted Project
dstuttard requested review of D103323: [DAGCombiner] Add support for mulhi const folding in DAGCombiner.
May 28 2021, 8:42 AM · Restricted Project
dstuttard added a comment to D103321: [AMDGPU] Stop mulhi from doing 24 bit mul for uniform values.

Wasn't sure about the new HasSMulHi - any thoughts?

May 28 2021, 8:38 AM · Restricted Project
dstuttard added a reviewer for D103321: [AMDGPU] Stop mulhi from doing 24 bit mul for uniform values: foad.
May 28 2021, 8:37 AM · Restricted Project
dstuttard retitled D103321: [AMDGPU] Stop mulhi from doing 24 bit mul for uniform values from Stop mulhi from doing 24 bit mul for uniform values to [AMDGPU] Stop mulhi from doing 24 bit mul for uniform values.
May 28 2021, 8:35 AM · Restricted Project
dstuttard requested review of D103321: [AMDGPU] Stop mulhi from doing 24 bit mul for uniform values.
May 28 2021, 8:34 AM · Restricted Project

May 18 2021

dstuttard added reviewers for D101633: [AMDGPU] Set number vgprs used in PS shaders based on input registers actually used: tpr, foad.
May 18 2021, 3:27 AM · Restricted Project

May 14 2021

dstuttard committed rG31b62aa162b4: [AMDGPU] Fix codegen of image intrinsics for g16 and a16 (authored by dstuttard).
[AMDGPU] Fix codegen of image intrinsics for g16 and a16
May 14 2021, 1:31 AM
dstuttard closed D102066: [AMDGPU] Fix codegen of image intrinsics for g16 and a16.
May 14 2021, 1:31 AM · Restricted Project
dstuttard committed rG72d570ca085c: [AMDGPU][AsmParser/Disassembler] Correct A16 and G16 handling (authored by dstuttard).
[AMDGPU][AsmParser/Disassembler] Correct A16 and G16 handling
May 14 2021, 1:28 AM
dstuttard closed D102231: [AMDGPU][AsmParser/Disassembler] Correct A16 and G16 handling.
May 14 2021, 1:28 AM · Restricted Project

May 13 2021

dstuttard added inline comments to D102066: [AMDGPU] Fix codegen of image intrinsics for g16 and a16.
May 13 2021, 9:04 AM · Restricted Project
dstuttard updated the diff for D102066: [AMDGPU] Fix codegen of image intrinsics for g16 and a16.

Updates based on review comments

May 13 2021, 9:03 AM · Restricted Project
dstuttard added inline comments to D102066: [AMDGPU] Fix codegen of image intrinsics for g16 and a16.
May 13 2021, 9:02 AM · Restricted Project
dstuttard added inline comments to D102231: [AMDGPU][AsmParser/Disassembler] Correct A16 and G16 handling.
May 13 2021, 6:51 AM · Restricted Project
dstuttard updated the diff for D102066: [AMDGPU] Fix codegen of image intrinsics for g16 and a16.

Moved part of this change into D102231 (required for that change).

May 13 2021, 6:49 AM · Restricted Project
dstuttard updated the diff for D102231: [AMDGPU][AsmParser/Disassembler] Correct A16 and G16 handling.

Moved change from D102066 required for this change (which will be committed first)

May 13 2021, 6:48 AM · Restricted Project
dstuttard added a comment to D102111: [AMDGPU] Update SCC defs to VCC when uses are changed to VCC.

I've got a case where we see this issue as well.

May 13 2021, 2:07 AM · Restricted Project

May 12 2021

dstuttard updated the diff for D102231: [AMDGPU][AsmParser/Disassembler] Correct A16 and G16 handling.

Made suggested changes

May 12 2021, 2:32 AM · Restricted Project

May 11 2021

dstuttard added inline comments to D102231: [AMDGPU][AsmParser/Disassembler] Correct A16 and G16 handling.
May 11 2021, 8:54 AM · Restricted Project
dstuttard updated the diff for D102231: [AMDGPU][AsmParser/Disassembler] Correct A16 and G16 handling.

Fixed sense of NoG16 argument passed in from asmparser and disassembler

May 11 2021, 6:18 AM · Restricted Project
dstuttard added a comment to D102066: [AMDGPU] Fix codegen of image intrinsics for g16 and a16.

See new patch with globalisel implementation

May 11 2021, 5:26 AM · Restricted Project
dstuttard updated the diff for D102066: [AMDGPU] Fix codegen of image intrinsics for g16 and a16.

Addressed review comments
Added support for global-isel. Note: global-isel produces less optimal code
presently. Will require some additional work to add extra combines

May 11 2021, 5:25 AM · Restricted Project
dstuttard added reviewers for D102231: [AMDGPU][AsmParser/Disassembler] Correct A16 and G16 handling: dp, foad, Flakebi.

This is a combination of the 2 previous patches - both asmparser and disassembler. I've also unified the code from SIInstrInfo.cpp into a single function since the code was duplicated 3 times before.

May 11 2021, 5:20 AM · Restricted Project
dstuttard abandoned D101619: [AMDGPU][AsmParser] Adjust img instruction address field if a16 present.

I've combined this with other changes into D102231 which was easier than trying to update this one

May 11 2021, 5:17 AM · Restricted Project
dstuttard abandoned D101620: [AMDGPU][Disassembler] Adjust img instruction address field if a16 present.

I've combined this with some other changes into D102231 - easier than trying to separate all the various bits

May 11 2021, 5:17 AM · Restricted Project
dstuttard requested review of D102231: [AMDGPU][AsmParser/Disassembler] Correct A16 and G16 handling.
May 11 2021, 5:16 AM · Restricted Project
dstuttard updated the diff for D101619: [AMDGPU][AsmParser] Adjust img instruction address field if a16 present.

a16 and g16 no longer linked

May 11 2021, 4:29 AM · Restricted Project

May 7 2021

dstuttard committed rG606d4e806192: AMDGPU: Correct const_index_stride for wave 32 for PAL ABI (authored by dstuttard).
AMDGPU: Correct const_index_stride for wave 32 for PAL ABI
May 7 2021, 5:44 AM
dstuttard added a comment to D101830: AMDGPU: Correct const_index_stride for wave 32 for PAL ABI.

Yes, thanks. Already reverted.

May 7 2021, 5:12 AM · Restricted Project
dstuttard added a reverting change for rG442de0c1adf3: AMDGPU: Correct const_index_stride for wave 32 for PAL ABI: rG793b4b26039e: Revert "AMDGPU: Correct const_index_stride for wave 32 for PAL ABI".
May 7 2021, 4:50 AM
dstuttard committed rG793b4b26039e: Revert "AMDGPU: Correct const_index_stride for wave 32 for PAL ABI" (authored by dstuttard).
Revert "AMDGPU: Correct const_index_stride for wave 32 for PAL ABI"
May 7 2021, 4:50 AM
dstuttard added a reverting change for D101830: AMDGPU: Correct const_index_stride for wave 32 for PAL ABI: rG793b4b26039e: Revert "AMDGPU: Correct const_index_stride for wave 32 for PAL ABI".
May 7 2021, 4:49 AM · Restricted Project
dstuttard added a comment to D102066: [AMDGPU] Fix codegen of image intrinsics for g16 and a16.

D101619, D101620 are related.

May 7 2021, 4:43 AM · Restricted Project