Page MenuHomePhabricator

stuij (Ties Stuij)
User

Projects

User does not belong to any projects.

User Details

User Since
Apr 24 2018, 8:11 AM (135 w, 3 d)

Recent Activity

Oct 27 2020

stuij added a comment to D88566: be more specific when testing for no fuse-ld warnings.

If this cannot be reproduced with the OSS LLVM, I am not sure you should adjust such a test.

Oct 27 2020, 9:46 AM · Restricted Project
stuij abandoned D88566: be more specific when testing for no fuse-ld warnings.

Abandoned because lack of reaction for such an unimportant issue.

Oct 27 2020, 6:32 AM · Restricted Project

Oct 20 2020

stuij added a comment to D88566: be more specific when testing for no fuse-ld warnings.

ping

Oct 20 2020, 5:39 AM · Restricted Project

Oct 13 2020

stuij committed rG208987844ffa: [ARM] Follow AACPS standard for volatile bit-fields access width (authored by stuij).
[ARM] Follow AACPS standard for volatile bit-fields access width
Oct 13 2020, 2:32 AM
stuij closed D72932: [ARM] Follow AACPS standard for volatile bit-fields access width.
Oct 13 2020, 2:32 AM · Restricted Project, Restricted Project

Oct 12 2020

stuij updated the diff for D72932: [ARM] Follow AACPS standard for volatile bit-fields access width.

removed clang/test/CodeGen/lit.local.cfg.py

Oct 12 2020, 7:45 AM · Restricted Project, Restricted Project
stuij updated the diff for D72932: [ARM] Follow AACPS standard for volatile bit-fields access width.

addressed review comment to hardwire non-MS target platforms

Oct 12 2020, 7:40 AM · Restricted Project, Restricted Project

Oct 7 2020

stuij added inline comments to D88566: be more specific when testing for no fuse-ld warnings.
Oct 7 2020, 10:22 AM · Restricted Project
stuij added a comment to D88566: be more specific when testing for no fuse-ld warnings.

Hi @MaskRay. Yes, so we're seeing a warning specific to our Armcompiler toolchain, so I'm guessing that isn't relevant to OSS LLVM:
armclang: warning: '--target=x86_64-unknown-linux' is not supported.

Oct 7 2020, 10:21 AM · Restricted Project
stuij updated the diff for D72932: [ARM] Follow AACPS standard for volatile bit-fields access width.

After committing this patch, clang/test/CodeGen/volatile.c failed on Arm/AArch64 buildbot hosts. The reason for this is that %itanium_abi_triple, a run line Lit target triple substitution at the top of the file, is filled in with the host arch triple. For example: aarch64-unknown-linux-gnu. I've amended the tests to take into account the changes to code generation in this patch.

Oct 7 2020, 4:47 AM · Restricted Project, Restricted Project
stuij reopened D72932: [ARM] Follow AACPS standard for volatile bit-fields access width.

Reopening as this commit made clang/test/CodeGen/volatile.c fail on Arm/AArch64 buildbot hosts.

Oct 7 2020, 4:46 AM · Restricted Project, Restricted Project

Sep 30 2020

stuij updated the summary of D88566: be more specific when testing for no fuse-ld warnings.
Sep 30 2020, 6:38 AM · Restricted Project
stuij updated the summary of D88566: be more specific when testing for no fuse-ld warnings.
Sep 30 2020, 6:38 AM · Restricted Project
stuij updated the summary of D88566: be more specific when testing for no fuse-ld warnings.
Sep 30 2020, 6:38 AM · Restricted Project
stuij updated the diff for D88566: be more specific when testing for no fuse-ld warnings.

slight change in commit message

Sep 30 2020, 6:35 AM · Restricted Project
stuij requested review of D88566: be more specific when testing for no fuse-ld warnings.
Sep 30 2020, 6:29 AM · Restricted Project

Sep 11 2020

stuij added a comment to D87463: [clang][aarch64] Fix mangling of bfloat16 neon vectors.

I'm not sure what the procedure is to get patches into LLVM 11, but I think there's still time. Could you try to get this in there as well @c-rhodes ?

Sep 11 2020, 3:27 AM · Restricted Project

Sep 10 2020

stuij accepted D87463: [clang][aarch64] Fix mangling of bfloat16 neon vectors.

Thanks, this one crept in during mass-renaming.

Sep 10 2020, 4:47 PM · Restricted Project

Sep 8 2020

stuij committed rGd6f3f612318f: Revert "[ARM] Follow AACPS standard for volatile bit-fields access width" (authored by stuij).
Revert "[ARM] Follow AACPS standard for volatile bit-fields access width"
Sep 8 2020, 10:47 AM
stuij added a reverting change for rG514df1b2bb1e: [ARM] Follow AACPS standard for volatile bit-fields access width: rGd6f3f612318f: Revert "[ARM] Follow AACPS standard for volatile bit-fields access width".
Sep 8 2020, 10:47 AM
stuij added a reverting change for D72932: [ARM] Follow AACPS standard for volatile bit-fields access width: rGd6f3f612318f: Revert "[ARM] Follow AACPS standard for volatile bit-fields access width".
Sep 8 2020, 10:46 AM · Restricted Project, Restricted Project
stuij committed rG514df1b2bb1e: [ARM] Follow AACPS standard for volatile bit-fields access width (authored by stuij).
[ARM] Follow AACPS standard for volatile bit-fields access width
Sep 8 2020, 9:50 AM
stuij closed D72932: [ARM] Follow AACPS standard for volatile bit-fields access width.
Sep 8 2020, 9:50 AM · Restricted Project, Restricted Project
stuij added inline comments to D72932: [ARM] Follow AACPS standard for volatile bit-fields access width.
Sep 8 2020, 7:55 AM · Restricted Project, Restricted Project
stuij updated the diff for D72932: [ARM] Follow AACPS standard for volatile bit-fields access width.

addressed review comment

Sep 8 2020, 7:53 AM · Restricted Project, Restricted Project
stuij commandeered D72932: [ARM] Follow AACPS standard for volatile bit-fields access width.

Commandeering as I've made some changes to the patch.

Sep 8 2020, 7:52 AM · Restricted Project, Restricted Project

Aug 28 2020

stuij added a comment to D72932: [ARM] Follow AACPS standard for volatile bit-fields access width.

@ostannard: pinging on behalf of @dnsampaio. The changes still apply cleanly.

Aug 28 2020, 7:16 AM · Restricted Project, Restricted Project
stuij accepted D86088: Reduce the number of memory allocations when displaying a warning about clobbering reserved registers (NFC)..

Looks harmless :o) Thanks, LGTM.

Aug 28 2020, 6:21 AM · Restricted Project
stuij committed rGd678e14c55be: [AArch64][CodeGen] Restrict bfloat vector operations to what's actually… (authored by stuij).
[AArch64][CodeGen] Restrict bfloat vector operations to what's actually…
Aug 28 2020, 3:45 AM
stuij closed D85101: [AArch64][CodeGen] Restrict bfloat vector operations to what's actually supported.
Aug 28 2020, 3:44 AM · Restricted Project

Aug 27 2020

stuij updated the diff for D85101: [AArch64][CodeGen] Restrict bfloat vector operations to what's actually supported.

addressing review comments

Aug 27 2020, 7:56 AM · Restricted Project

Aug 3 2020

stuij added a comment to D85101: [AArch64][CodeGen] Restrict bfloat vector operations to what's actually supported.

Quick question - what is the expected behaviour? Do we just never expect to see an bf16 add, and if we do it's a fatal error? Or is some form of automatic promotion expected to happen?

Aug 3 2020, 5:18 AM · Restricted Project

Aug 2 2020

stuij added reviewers for D85101: [AArch64][CodeGen] Restrict bfloat vector operations to what's actually supported: dmgreen, dnsampaio, chill.
Aug 2 2020, 3:59 PM · Restricted Project
stuij requested review of D85101: [AArch64][CodeGen] Restrict bfloat vector operations to what's actually supported.
Aug 2 2020, 3:47 PM · Restricted Project

Jul 8 2020

stuij committed rG26a22478cdfe: [CodeGen] Don't combine extract + concat vectors with non-legal types (authored by stuij).
[CodeGen] Don't combine extract + concat vectors with non-legal types
Jul 8 2020, 7:30 AM
stuij closed D83231: [CodeGen] Don't combine extract + concat vectors with non-legal types.
Jul 8 2020, 7:30 AM · Restricted Project
stuij added inline comments to D83231: [CodeGen] Don't combine extract + concat vectors with non-legal types.
Jul 8 2020, 4:26 AM · Restricted Project
stuij updated the diff for D83231: [CodeGen] Don't combine extract + concat vectors with non-legal types.

addressed review comment

Jul 8 2020, 4:23 AM · Restricted Project

Jul 6 2020

stuij updated the summary of D83231: [CodeGen] Don't combine extract + concat vectors with non-legal types.
Jul 6 2020, 9:17 AM · Restricted Project
stuij updated the summary of D83231: [CodeGen] Don't combine extract + concat vectors with non-legal types.
Jul 6 2020, 8:22 AM · Restricted Project
stuij updated the summary of D83231: [CodeGen] Don't combine extract + concat vectors with non-legal types.
Jul 6 2020, 8:21 AM · Restricted Project
stuij added reviewers for D83231: [CodeGen] Don't combine extract + concat vectors with non-legal types: miyuki, arsenm, dmgreen.
Jul 6 2020, 8:19 AM · Restricted Project
Herald added a project to D83231: [CodeGen] Don't combine extract + concat vectors with non-legal types: Restricted Project.
Jul 6 2020, 8:11 AM · Restricted Project
stuij added inline comments to D79709: [AArch64][BFloat] basic AArch64 bfloat support.
Jul 6 2020, 7:06 AM · Restricted Project

Jun 23 2020

stuij accepted D80928: [BFloat] Add convert/copy instrinsic support.

LGTM

Jun 23 2020, 4:45 AM · Restricted Project, Restricted Project

Jun 22 2020

stuij accepted D82206: [ARM][BFloat] Implement bf16 get/set_lane without casts to i16 vectors.

LGTM. Thanks!

Jun 22 2020, 10:12 AM · Restricted Project

Jun 18 2020

stuij committed rG9dda41e43327: [BFloat] Move LLVMBFloatTypeKind to the end of the enum (authored by stuij).
[BFloat] Move LLVMBFloatTypeKind to the end of the enum
Jun 18 2020, 6:04 PM
stuij closed D82135: [BFloat] Move LLVMBFloatTypeKind to the end of the enum.
Jun 18 2020, 6:04 PM · Restricted Project
stuij added inline comments to D78190: Add Bfloat IR type.
Jun 18 2020, 4:59 PM · Restricted Project, Restricted Project
stuij added a reviewer for D82135: [BFloat] Move LLVMBFloatTypeKind to the end of the enum: tstellar.
Jun 18 2020, 4:59 PM · Restricted Project
stuij created D82135: [BFloat] Move LLVMBFloatTypeKind to the end of the enum.
Jun 18 2020, 4:59 PM · Restricted Project
stuij committed rG035795659b60: [ARM][bfloat] Do not coerce bfloat arguments and returns to integers (authored by stuij).
[ARM][bfloat] Do not coerce bfloat arguments and returns to integers
Jun 18 2020, 10:54 AM
stuij closed D81837: [ARM][bfloat] Removing lowering of bfloat arguments and returns from Clang's CodeGen.
Jun 18 2020, 10:54 AM · Restricted Project
stuij added inline comments to D78190: Add Bfloat IR type.
Jun 18 2020, 9:45 AM · Restricted Project, Restricted Project
stuij accepted D81304: [llvm][SveEmitter] Emit the bfloat version of `svld1ro`..

LGTM

Jun 18 2020, 9:45 AM · Restricted Project
stuij added inline comments to D81373: [ARM] Basic bfloat support.
Jun 18 2020, 8:46 AM · Restricted Project
stuij added inline comments to D81373: [ARM] Basic bfloat support.
Jun 18 2020, 8:06 AM · Restricted Project
stuij added inline comments to D81304: [llvm][SveEmitter] Emit the bfloat version of `svld1ro`..
Jun 18 2020, 5:24 AM · Restricted Project

Jun 16 2020

stuij added a comment to D81451: [ARM][Clang] Removing lowering of half-precision FP arguments and returns from Clang's CodeGen.

Hi @stuij,

The changes to the backend only handle the half (f16) type itself, not vectors that have it as their base type.

From what I've checked on the AAPCS, the rules for handling those types are a bit different and they would require their own handling in the backend's calling convention lowering.
I haven't looked into the backend's handling of those types in detail, but I believe a similar approach to the one taken for f16 would be possible for the vector types as well.

Jun 16 2020, 11:00 AM · Restricted Project
stuij updated the diff for D81837: [ARM][bfloat] Removing lowering of bfloat arguments and returns from Clang's CodeGen.

We only remove clang-side lowering for non-vector bfloat types.

Jun 16 2020, 10:26 AM · Restricted Project

Jun 15 2020

stuij accepted D81779: [SelectionDAG] Add MVT::bf16 to getConstantFP().

LGTM. Thanks :)

Jun 15 2020, 6:14 PM · Restricted Project
stuij added inline comments to D81304: [llvm][SveEmitter] Emit the bfloat version of `svld1ro`..
Jun 15 2020, 4:35 PM · Restricted Project
stuij added a comment to D81451: [ARM][Clang] Removing lowering of half-precision FP arguments and returns from Clang's CodeGen.

I would have expected changes for neon as well.

Jun 15 2020, 5:56 AM · Restricted Project
stuij updated the diff for D81837: [ARM][bfloat] Removing lowering of bfloat arguments and returns from Clang's CodeGen.

adjust arm bfloat mangle tests for these changes

Jun 15 2020, 5:24 AM · Restricted Project
stuij added reviewers for D81837: [ARM][bfloat] Removing lowering of bfloat arguments and returns from Clang's CodeGen: pratlucas, LukeGeeson, dnsampaio, ostannard.
Jun 15 2020, 4:51 AM · Restricted Project
stuij created D81837: [ARM][bfloat] Removing lowering of bfloat arguments and returns from Clang's CodeGen.
Jun 15 2020, 4:19 AM · Restricted Project

Jun 12 2020

stuij accepted D80716: [AArch64]: BFloat Load/Store Intrinsics&CodeGen.

LGTM. Thanks!

Jun 12 2020, 7:31 AM · Restricted Project, Restricted Project

Jun 11 2020

stuij accepted D80752: [AArch64]: BFloat MatMul Intrinsics&CodeGen.

LGTM. Thanks!

Jun 11 2020, 6:35 AM · Restricted Project, Restricted Project
stuij added inline comments to D80716: [AArch64]: BFloat Load/Store Intrinsics&CodeGen.
Jun 11 2020, 4:50 AM · Restricted Project, Restricted Project

Jun 7 2020

stuij committed rG5945e9799e77: [clang][BFloat] Add reinterpret cast intrinsics (authored by stuij).
[clang][BFloat] Add reinterpret cast intrinsics
Jun 7 2020, 6:53 AM
stuij closed D79869: [clang][BFloat] Add reinterpret cast intrinsics.
Jun 7 2020, 6:52 AM · Restricted Project

Jun 5 2020

stuij added a comment to D81302: Fix convertBFloatAPFloatToAPInt for NaN/Inf values.

Ah, awesome. Thanks :)

Jun 5 2020, 3:43 PM · Restricted Project
stuij committed rG8b137a430636: [clang][BFloat] Add create/set/get/dup intrinsics (authored by stuij).
[clang][BFloat] Add create/set/get/dup intrinsics
Jun 5 2020, 6:38 AM
stuij closed D79710: [clang][BFloat] Add create/set/get/dup intrinsics.
Jun 5 2020, 6:37 AM · Restricted Project
stuij committed rGa6fcf5ca033a: [clang][BFloat] add NEON emitter for bfloat (authored by stuij).
[clang][BFloat] add NEON emitter for bfloat
Jun 5 2020, 6:37 AM
stuij closed D79708: [clang][BFloat] add NEON emitter for bfloat.
Jun 5 2020, 6:36 AM · Restricted Project
stuij added a comment to D81181: [TargetLowering][NFC] More efficient emitPatchpoint()..

Yup, LLVM now builds and check-all's. Thanks for fixing!

Jun 5 2020, 6:33 AM · Restricted Project
stuij added a comment to D81181: [TargetLowering][NFC] More efficient emitPatchpoint()..

ah sorry, I guess that might have been the wrong link. My copy pasting skills are flaky.

Jun 5 2020, 6:01 AM · Restricted Project
stuij committed rG1e447318339a: [ARM] Add poly64_t on AArch32. (authored by stuij).
[ARM] Add poly64_t on AArch32.
Jun 5 2020, 5:29 AM
stuij closed D79711: [ARM] Add poly64_t on AArch32..
Jun 5 2020, 5:29 AM · Restricted Project
stuij added a comment to D81181: [TargetLowering][NFC] More efficient emitPatchpoint()..

Seems like this is breaking on buildbot (and also my local checkout): http://lab.llvm.org:8011/builders/lldb-x64-windows-ninja/builds/16759

Jun 5 2020, 5:27 AM · Restricted Project
stuij committed rGecd682bbf5e6: [ARM] Add __bf16 as new Bfloat16 C Type (authored by stuij).
[ARM] Add __bf16 as new Bfloat16 C Type
Jun 5 2020, 2:44 AM
stuij closed D76077: [ARM] Add __bf16 as new Bfloat16 C Type.
Jun 5 2020, 2:44 AM · Restricted Project

Jun 4 2020

stuij updated the diff for D79708: [clang][BFloat] add NEON emitter for bfloat.

addressed remaining nits

Jun 4 2020, 4:36 PM · Restricted Project
stuij updated the diff for D76077: [ARM] Add __bf16 as new Bfloat16 C Type.

Resolved merge conflicts with head. Fixed minor oversight.

Jun 4 2020, 4:36 PM · Restricted Project
stuij added inline comments to D76077: [ARM] Add __bf16 as new Bfloat16 C Type.
Jun 4 2020, 4:36 PM · Restricted Project
stuij added a comment to D80716: [AArch64]: BFloat Load/Store Intrinsics&CodeGen.

Besides from rebasing to get @pratlucas changes upstream.

@stuij please could you confirm if you are happy with this, so I can merge

Jun 4 2020, 3:30 PM · Restricted Project, Restricted Project
stuij added a comment to D80752: [AArch64]: BFloat MatMul Intrinsics&CodeGen.

For the backend tests, I suggest using -asm-verbose=0 with llc to only print instructions and get rid of // kill: .. and friends. Use update_cc_test_checks.py again to regenerate the testing.

Jun 4 2020, 3:30 PM · Restricted Project, Restricted Project

Jun 3 2020

stuij updated the diff for D76077: [ARM] Add __bf16 as new Bfloat16 C Type.

Addressed review comments. Notably what to do in combination with -mfloat-abi=soft.

Jun 3 2020, 5:07 PM · Restricted Project

Jun 2 2020

stuij added inline comments to D79710: [clang][BFloat] Add create/set/get/dup intrinsics.
Jun 2 2020, 9:52 AM · Restricted Project

Jun 1 2020

stuij added a comment to D76077: [ARM] Add __bf16 as new Bfloat16 C Type.

At the moment when going through the GCC compatibility driver (standard interface), we get __bf16 is not supported on this target.

If this is the current behaviour, and consistent with GCC, that sounds reasonable.

Jun 1 2020, 9:39 AM · Restricted Project
stuij updated the summary of D76077: [ARM] Add __bf16 as new Bfloat16 C Type.
Jun 1 2020, 8:00 AM · Restricted Project
stuij added a comment to D76077: [ARM] Add __bf16 as new Bfloat16 C Type.

Can you summarise where we are? I.e.,

  • float-abi=soft doesn't work. But what is the problem? Are we not simply passing i16s, is that not what we are supposed to do?
Jun 1 2020, 8:00 AM · Restricted Project
stuij added inline comments to D76077: [ARM] Add __bf16 as new Bfloat16 C Type.
Jun 1 2020, 6:55 AM · Restricted Project

May 29 2020

stuij added inline comments to D76077: [ARM] Add __bf16 as new Bfloat16 C Type.
May 29 2020, 8:06 AM · Restricted Project

May 28 2020

stuij requested changes to D80716: [AArch64]: BFloat Load/Store Intrinsics&CodeGen.

We need testing for the backend code.

May 28 2020, 7:02 AM · Restricted Project, Restricted Project

May 27 2020

stuij committed rG29f8056b54ea: [CodeGen] fix typo `def nxv1bf32` -> `def nxv1f32` (authored by stuij).
[CodeGen] fix typo `def nxv1bf32` -> `def nxv1f32`
May 27 2020, 10:18 AM
stuij added inline comments to D79706: [CodeGen][BFloat] Add bfloat MVT type.
May 27 2020, 10:17 AM · Restricted Project
stuij added inline comments to D79706: [CodeGen][BFloat] Add bfloat MVT type.
May 27 2020, 9:43 AM · Restricted Project
stuij committed rG78bd0c0e5e8f: [AArch64][BFloat] add BFloat instruction support for AArch64 (authored by stuij).
[AArch64][BFloat] add BFloat instruction support for AArch64
May 27 2020, 8:06 AM