[llvm] [NVPTX] split long-running wmma.py test into smaller chunks. (PR #72331)
via llvm-commits
llvm-commits at lists.llvm.org
Tue Nov 14 16:58:07 PST 2023
github-actions[bot] wrote:
<!--LLVM CODE FORMAT COMMENT: {darker}-->
:warning: Python code formatter, darker found issues in your code. :warning:
<details>
<summary>
You can test this locally with the following command:
</summary>
``````````bash
darker --check --diff -r beb702c0ad69244da6caedca9a0f3c6ee42e5bae..8185966f0a32764f6b63b0acf60c93f06e598b0c llvm/test/CodeGen/NVPTX/wmma-ptx60-sm70.py llvm/test/CodeGen/NVPTX/wmma-ptx61-sm70.py llvm/test/CodeGen/NVPTX/wmma-ptx63-sm72.py llvm/test/CodeGen/NVPTX/wmma-ptx63-sm75.py llvm/test/CodeGen/NVPTX/wmma-ptx64-sm70.py llvm/test/CodeGen/NVPTX/wmma-ptx65-sm75.py llvm/test/CodeGen/NVPTX/wmma-ptx71-sm80.py llvm/test/CodeGen/NVPTX/wmma.py
``````````
</details>
<details>
<summary>
View the diff from darker here.
</summary>
``````````diff
--- wmma-ptx60-sm70.py 2023-11-15 00:52:43.000000 +0000
+++ wmma-ptx60-sm70.py 2023-11-15 00:57:58.884025 +0000
@@ -10,6 +10,7 @@
# RUN: llc < %t-ptx60-sm_70.ll -march=nvptx64 -mcpu=sm_70 -mattr=+ptx60 \
# RUN: | %ptxas-verify -arch=sm_70 \
# RUN: %}
import wmma
+
wmma.main()
--- wmma-ptx61-sm70.py 2023-11-15 00:52:43.000000 +0000
+++ wmma-ptx61-sm70.py 2023-11-15 00:57:58.891749 +0000
@@ -10,6 +10,7 @@
# RUN: llc < %t-ptx61-sm_70.ll -march=nvptx64 -mcpu=sm_70 -mattr=+ptx61 \
# RUN: | %ptxas-verify -arch=sm_70 \
# RUN: %}
import wmma
+
wmma.main()
--- wmma-ptx63-sm72.py 2023-11-15 00:52:43.000000 +0000
+++ wmma-ptx63-sm72.py 2023-11-15 00:57:58.899677 +0000
@@ -10,6 +10,7 @@
# RUN: llc < %t-ptx63-sm_72.ll -march=nvptx64 -mcpu=sm_72 -mattr=+ptx63 \
# RUN: | %ptxas-verify -arch=sm_72 \
# RUN: %}
import wmma
+
wmma.main()
--- wmma-ptx63-sm75.py 2023-11-15 00:52:43.000000 +0000
+++ wmma-ptx63-sm75.py 2023-11-15 00:57:58.907089 +0000
@@ -10,6 +10,7 @@
# RUN: llc < %t-ptx63-sm_75.ll -march=nvptx64 -mcpu=sm_75 -mattr=+ptx63 \
# RUN: | %ptxas-verify -arch=sm_75 \
# RUN: %}
import wmma
+
wmma.main()
--- wmma-ptx64-sm70.py 2023-11-15 00:52:43.000000 +0000
+++ wmma-ptx64-sm70.py 2023-11-15 00:57:58.914502 +0000
@@ -10,6 +10,7 @@
# RUN: llc < %t-ptx64-sm_70.ll -march=nvptx64 -mcpu=sm_70 -mattr=+ptx64 \
# RUN: | %ptxas-verify -arch=sm_70 \
# RUN: %}
import wmma
+
wmma.main()
--- wmma-ptx65-sm75.py 2023-11-15 00:52:43.000000 +0000
+++ wmma-ptx65-sm75.py 2023-11-15 00:57:58.922164 +0000
@@ -10,6 +10,7 @@
# RUN: llc < %t-ptx65-sm_75.ll -march=nvptx64 -mcpu=sm_75 -mattr=+ptx65 \
# RUN: | %ptxas-verify -arch=sm_75 \
# RUN: %}
import wmma
+
wmma.main()
--- wmma-ptx71-sm80.py 2023-11-15 00:52:43.000000 +0000
+++ wmma-ptx71-sm80.py 2023-11-15 00:57:58.929672 +0000
@@ -10,6 +10,7 @@
# RUN: llc < %t-ptx71-sm_80.ll -march=nvptx64 -mcpu=sm_80 -mattr=+ptx71 \
# RUN: | %ptxas-verify -arch=sm_80 \
# RUN: %}
import wmma
+
wmma.main()
--- wmma.py 2023-11-15 00:52:43.000000 +0000
+++ wmma.py 2023-11-15 00:57:59.180403 +0000
@@ -991,10 +991,11 @@
items += gen_ldmatrix_tests()
items += gen_wmma_mma_tests()
items += gen_mma_tests()
gen_check_unsupported_ops(items)
+
def main():
global ptx_version
global gpu_arch
parser = argparse.ArgumentParser()
parser.add_argument("--ptx", type=int, default=60)
@@ -1004,7 +1005,8 @@
ptx_version = args.ptx
gpu_arch = args.gpu_arch
gen_tests()
+
if __name__ == "__main__":
- main()
+ main()
``````````
</details>
https://github.com/llvm/llvm-project/pull/72331
More information about the llvm-commits
mailing list