[SYCL][libclc][CUDA] Add --ffast-math support #5801

pgorlani · 2022-03-14T09:45:56Z

This patch allows the --ffast-math compiler flag to substitute the regular genfloatf math built-ins with their ::native versions.

Moreover, this patch completes the support of natives built-ins for libclc/ptx-nvidiacl connecting them with the __nv_fast functions present in libdevice. If a fast function is not available in libdevice the corresponding nvvm intrinsic is used.

Tests in intel/llvm-test-suite#919

This patch allows the `--ffast-math` compiler flag to substitute the regular `genfloatf` math built-ins with their `::native` versions. Moreover, this patch completes the support of natives built-ins for `libclc/ptx-nvidiacl` connecting them with the `__nv_fast` functions present in libdevice. If a fast function is not available in libdevice the corresponding `nvvm` intrinsic is used.

sycl/include/CL/sycl/builtins.hpp

mlychkov

llvm changes LGTM.

smanna12

FE changes LGTM.

s-kanaev · 2022-03-14T11:56:27Z

/verify with intel/llvm-test-suite#919

bader

Recovering my approve.
@pgorlani, to get this PR merged, you should ping @intel/llvm-reviewers-runtime team and investigate why new test added by intel/llvm-test-suite#919 doesn't pass.

pgorlani · 2022-03-29T10:16:58Z

Hi @bader, thanks for the prompt answer!

and investigate why new test added by intel/llvm-test-suite#919 doesn't pass.

Are you referring to the outcome of /verify with intel/llvm#5801? How can I see those results?

If not, the test introduced by intel/llvm-test-suite#919 is expected to fail if it is run on a branch that doesn't support the --ffast-math flag, i.e., it passes just in case #5801 is applied.

bader · 2022-03-29T11:03:39Z

and investigate why new test added by intel/llvm-test-suite#919 doesn't pass.

Are you referring to the outcome of /verify with intel/llvm#5801? How can I see those results?

If not, the test introduced by intel/llvm-test-suite#919 is expected to fail if it is run on a branch that doesn't support the --ffast-math flag, i.e., it passes just in case #5801 is applied.

I see that there were two runs:

From this PR - [SYCL][libclc][CUDA] Add --ffast-math support #5801 (comment) - logs can be found here or click on "Details" for Jenkins/verify-test-suite-logs.
From [SYCL] Add --ffast-math tests llvm-test-suite#919 - [SYCL] Add --ffast-math tests llvm-test-suite#919 (comment) - logs can be found here or click on "Details" for Jenkins/verify-test-suite-logs.

smanna12

FE changes LGTM

pgorlani · 2022-03-30T13:17:52Z

Thank you so much for the info, @bader.

I had missed the failings in Windows. I pushed a commit that should fix them.

mlychkov

LLVM changes LGTM.

llvm/lib/Target/NVPTX/NVPTXIntrinsics.td

llvm/include/llvm/IR/IntrinsicsNVVM.td

Remove empty line. Co-authored-by: Mikhail Lychkov <[email protected]>

smanna12

FE changes LGTM

pgorlani · 2022-03-31T08:13:43Z

/verify with intel/llvm-test-suite#919

pgorlani · 2022-03-31T10:35:19Z

All the tests passed.
Jenkins/Pre-ci-logs failed for Assert/assert_in_simultaneously_multiple_tus.cpp, which is now disabled.

Ping @intel/llvm-reviewers-runtime

pgorlani · 2022-04-25T12:41:17Z

Ping @v-klochkov

pgorlani · 2022-04-25T13:15:57Z

Thanks to everybody!

Would it be possible to merge this PR?

This patch adds tests for intel/llvm#5801

This patch adds tests for intel#5801

pgorlani requested review from a team and bader as code owners March 14, 2022 09:45

pgorlani requested a review from v-klochkov March 14, 2022 09:45

pgorlani mentioned this pull request Mar 14, 2022

[SYCL] Add --ffast-math tests intel/llvm-test-suite#919

Merged

bader approved these changes Mar 14, 2022

View reviewed changes

Fznamznon previously approved these changes Mar 14, 2022

View reviewed changes

bader reviewed Mar 14, 2022

View reviewed changes

sycl/include/CL/sycl/builtins.hpp Show resolved Hide resolved

mlychkov previously approved these changes Mar 14, 2022

View reviewed changes

smanna12 previously approved these changes Mar 14, 2022

View reviewed changes

pgorlani requested a review from bader March 29, 2022 08:59

bader previously approved these changes Mar 29, 2022

View reviewed changes

add fast-math flag in clang-cl

ffad8ef

pgorlani dismissed stale reviews from bader, smanna12, mlychkov, and Fznamznon via ffad8ef March 30, 2022 13:14

pgorlani requested a review from a team as a code owner March 30, 2022 13:14

bader requested review from Fznamznon, bader, mlychkov and smanna12 March 30, 2022 13:16

bader previously approved these changes Mar 30, 2022

View reviewed changes

smanna12 previously approved these changes Mar 30, 2022

View reviewed changes

mlychkov previously approved these changes Mar 30, 2022

View reviewed changes

llvm/lib/Target/NVPTX/NVPTXIntrinsics.td Outdated Show resolved Hide resolved

mlychkov reviewed Mar 30, 2022

View reviewed changes

llvm/include/llvm/IR/IntrinsicsNVVM.td Outdated Show resolved Hide resolved

Update llvm/lib/Target/NVPTX/NVPTXIntrinsics.td

d2d9055

Remove empty line. Co-authored-by: Mikhail Lychkov <[email protected]>

pgorlani dismissed stale reviews from mlychkov, smanna12, and bader via d2d9055 March 30, 2022 13:29

Update llvm/include/llvm/IR/IntrinsicsNVVM.td

74ad3d1

Remove empty line. Co-authored-by: Mikhail Lychkov <[email protected]>

smanna12 approved these changes Mar 30, 2022

View reviewed changes

mlychkov approved these changes Mar 30, 2022

View reviewed changes

bader approved these changes Mar 30, 2022

View reviewed changes

mdtoguchi approved these changes Mar 30, 2022

View reviewed changes

Fznamznon approved these changes Mar 30, 2022

View reviewed changes

v-klochkov approved these changes Apr 25, 2022

View reviewed changes

againull self-requested a review April 25, 2022 15:01

againull approved these changes Apr 25, 2022

View reviewed changes

againull merged commit 0f0c5d1 into intel:sycl Apr 25, 2022

jingwan2 mentioned this pull request Apr 27, 2022

[CUDA] sycl::exp(double) compile fail with -ffast-math on CUDA backend with patch https://github.com/intel/llvm/pull/5801 #6068

Closed

steffenlarsen pushed a commit to intel/llvm-test-suite that referenced this pull request Jun 24, 2022

[SYCL] Add --ffast-math tests (#919)

773b61e

This patch adds tests for intel/llvm#5801

JackAKirk mentioned this pull request Sep 15, 2022

[SYCL] Fix marray math function impls #6038

Merged

aelovikov-intel pushed a commit to aelovikov-intel/llvm that referenced this pull request Mar 27, 2023

[SYCL] Add --ffast-math tests (intel/llvm-test-suite#919)

7506be1

This patch adds tests for intel#5801

[SYCL][libclc][CUDA] Add --ffast-math support #5801

[SYCL][libclc][CUDA] Add --ffast-math support #5801

Uh oh!

Conversation

pgorlani commented Mar 14, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

mlychkov left a comment

Choose a reason for hiding this comment

Uh oh!

smanna12 left a comment

Choose a reason for hiding this comment

Uh oh!

s-kanaev commented Mar 14, 2022

Uh oh!

bader left a comment

Choose a reason for hiding this comment

Uh oh!

pgorlani commented Mar 29, 2022

Uh oh!

bader commented Mar 29, 2022

Uh oh!

smanna12 left a comment

Choose a reason for hiding this comment

Uh oh!

pgorlani commented Mar 30, 2022

Uh oh!

mlychkov left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

smanna12 left a comment

Choose a reason for hiding this comment

Uh oh!

pgorlani commented Mar 31, 2022

Uh oh!

pgorlani commented Mar 31, 2022

Uh oh!

pgorlani commented Apr 25, 2022

Uh oh!

pgorlani commented Apr 25, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants

pgorlani commented Mar 14, 2022 •

edited

Loading