Page MenuHomePhabricator
Feed Advanced Search

Fri, Aug 5

scchan added a comment to D130784: [AMDGPU] Support LDS spilling.

Just briefly glanced through the code and it seems that it doesn't look at whether a kernel has dynamic LDS or not because LDS spilling has to be disabled in that case, can you confirm?

Fri, Aug 5, 9:03 AM · Restricted Project, Restricted Project

Jun 7 2022

scchan added a comment to D127052: [amdgpu][nfc] Allocate kernel-specific LDS struct deterministically.

Hoping this is uncontroversial - got more patches to follow up with if we can land this,

Jun 7 2022, 8:08 AM · Restricted Project, Restricted Project

Apr 8 2022

scchan requested review of D123387: [clang-offload-bundler] fix "no output file" issue with -outputs.
Apr 8 2022, 6:41 AM · Restricted Project, Restricted Project

Apr 4 2022

scchan updated the diff for D120662: [clang-offload-bundler] add -input/-output options.
  • fixed the hip-link-bundle-archive.hip test issue on Windows
  • fixed a couple of formatting issues
Apr 4 2022, 9:50 AM · Restricted Project, Restricted Project

Apr 1 2022

scchan updated the diff for D120662: [clang-offload-bundler] add -input/-output options.

update the failing lit tests to recognize the new switches

Apr 1 2022, 1:29 PM · Restricted Project, Restricted Project

Mar 28 2022

scchan updated the diff for D120662: [clang-offload-bundler] add -input/-output options.

update the clang-offload-bundler-asserts-on test

Mar 28 2022, 2:41 PM · Restricted Project, Restricted Project
scchan updated the diff for D120662: [clang-offload-bundler] add -input/-output options.

fix lit test failures on Windows, rebased

Mar 28 2022, 2:36 PM · Restricted Project, Restricted Project

Feb 28 2022

scchan added a reviewer for D120662: [clang-offload-bundler] add -input/-output options: tra.
Feb 28 2022, 7:45 AM · Restricted Project, Restricted Project
scchan requested review of D120662: [clang-offload-bundler] add -input/-output options.
Feb 28 2022, 7:44 AM · Restricted Project, Restricted Project

Jan 19 2022

scchan added a comment to D117494: AMDGPU: Account for usage HIP-style dynamic LDS.

I am OK for now. We may need keep an eye for the perf regression and be prepared to figure a way to alleviate that.

Jan 19 2022, 7:57 AM · Restricted Project

Jan 18 2022

scchan updated the diff for D117494: AMDGPU: Account for usage HIP-style dynamic LDS.

Check for zero-sized and move the check into the existing loop.

Jan 18 2022, 3:17 PM · Restricted Project

Jan 17 2022

scchan added reviewers for D117494: AMDGPU: Account for usage HIP-style dynamic LDS: arsenm, yaxunl.
Jan 17 2022, 8:17 AM · Restricted Project
scchan requested review of D117494: AMDGPU: Account for usage HIP-style dynamic LDS.
Jan 17 2022, 8:15 AM · Restricted Project

Jul 22 2021

scchan accepted D106571: [HIP] Fix visibility of __hip_fatbin.

LGTM

Jul 22 2021, 10:36 AM · Restricted Project

Jun 30 2021

scchan accepted D104904: [OpenMP][AMDGCN] Initial math headers support.
Jun 30 2021, 7:59 AM · Restricted Project

Jun 25 2021

scchan added inline comments to D104904: [OpenMP][AMDGCN] Initial math headers support.
Jun 25 2021, 6:50 AM · Restricted Project
scchan added a reviewer for D104904: [OpenMP][AMDGCN] Initial math headers support: b-sumner.
Jun 25 2021, 5:41 AM · Restricted Project

Jun 18 2021

scchan updated the diff for D104392: [HIP] Add support functions for C++ polymorphic types.

Minor clean up in the hip-header.hip test

Jun 18 2021, 10:30 AM · Restricted Project
scchan updated the diff for D104392: [HIP] Add support functions for C++ polymorphic types.

Adding test

Jun 18 2021, 9:46 AM · Restricted Project

Jun 16 2021

scchan requested review of D104392: [HIP] Add support functions for C++ polymorphic types.
Jun 16 2021, 8:42 AM · Restricted Project

Aug 6 2020

scchan added a reviewer for D85471: Make clang HIP headers compatible with C++98: yaxunl.
Aug 6 2020, 12:57 PM · Restricted Project
scchan requested review of D85471: Make clang HIP headers compatible with C++98.
Aug 6 2020, 12:56 PM · Restricted Project

Jun 24 2020

scchan added inline comments to D82496: [amdgpu] Add codegen support for HIP dynamic shared memory..
Jun 24 2020, 1:34 PM · Restricted Project

Feb 27 2020

Gerrit Code Review <gerritcr@atlswgit02.amd.com> committed rGfd9b4174403a: Merge "Speed up SAC when erasing unwanted functions by deleting the function… (authored by scchan).
Merge "Speed up SAC when erasing unwanted functions by deleting the function…
Feb 27 2020, 7:05 AM
Gerrit Code Review <gerritcr@atlswgit02.amd.com> committed rG0823d86dceb7: Merge "SelectAcceleratorCode: strip pointer casts when examining a call" into… (authored by scchan).
Merge "SelectAcceleratorCode: strip pointer casts when examining a call" into…
Feb 27 2020, 7:05 AM
scchan committed rG0334a838d6d4: SelectAcceleratorCode: strip pointer casts when examining a call (authored by sameerds).
SelectAcceleratorCode: strip pointer casts when examining a call
Feb 27 2020, 7:04 AM
Gerrit Code Review <gerritcr@atlswgit02.amd.com> committed rG5429b86d8620: Merge "do not traverse callgraph during pass initialization" into amd-common (authored by scchan).
Merge "do not traverse callgraph during pass initialization" into amd-common
Feb 27 2020, 7:04 AM
scchan committed rGb50ff5b4f946: do not traverse callgraph during pass initialization (authored by sameerds).
do not traverse callgraph during pass initialization
Feb 27 2020, 7:04 AM

Aug 13 2018

scchan accepted D50626: AMDGPU: Implement llvm.amdgcn.icmp/fcmp for i16/f16.

LGTM

Aug 13 2018, 7:44 AM
scchan accepted D50624: AMDGPU: Stop producing icmp/fcmp intrinsics with invalid types.

LGTM

Aug 13 2018, 7:32 AM

Jun 22 2018

scchan added a comment to D48493: [HIP] Support flush denorms bitcode.

LGTM

Jun 22 2018, 10:52 AM · Restricted Project, Restricted Project

May 16 2018

scchan accepted D46769: [AMDGPU] Change llvm.debugtrap to be a debug breakpoint that can resume execution..
May 16 2018, 9:16 AM

Mar 21 2017

scchan added a comment to D31210: [AMDGPU] Switch address space mapping by triple environment.

I'm concerned about the default address space to be 64 bit. It would move alloca into generic address space effectively making private address to be 64 bit.
This may have very undesirable performance implications, like address arithmetic can become expensive 64 bit and only be truncated at load or store.
I realize you will use addrspacecast on an alloca's value, though I'm not sure that is sufficient to mitigate performance hit.
I believe such change shall not be made without a good performance comparison with the feature enabled, provided the very likely performance issues.

Did not we want to use this: http://lists.llvm.org/pipermail/llvm-dev/2017-March/111199.html and use non-0 for our allocas?

Our final goal is to let alloca return private pointer. The Clang changes are mostly common whether alloca returns generic pointer or private pointer. Actually to be able to test the above patch we need to get the changes in Clang done first.

Mar 21 2017, 3:43 PM

Mar 1 2017

scchan accepted D29670: [DebugInfo] Emit address space with DW_AT_address_class attribute for pointer and reference types.

LGTM

Mar 1 2017, 7:12 AM