Skip to content

[rocm7.0_internal_testing] Prevent static initialization of at::cuda::warp_size()#2293

Merged
jithunnair-amd merged 3 commits intorocm7.0_internal_testingfrom
rocm7.0_IT_fix_warpsize
Jun 25, 2025
Merged

[rocm7.0_internal_testing] Prevent static initialization of at::cuda::warp_size()#2293
jithunnair-amd merged 3 commits intorocm7.0_internal_testingfrom
rocm7.0_IT_fix_warpsize

Conversation

@ethanwee1
Copy link

@ethanwee1 ethanwee1 commented Jun 25, 2025 •
edited by jeffdaily
Loading

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

Error

#24 437.7   what():  HIP error: no ROCm-capable device is detected
#24 437.7 HIP kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
#24 437.7 For debugging consider passing AMD_SERIALIZE_KERNEL=3
#24 437.7 Device-side assertions were explicitly omitted for this error check; the error probably arose while initializing the DSA handlers.
#24 437.7 Exception raised from c10_hip_check_implementation at /pytorch/c10/hip/HIPException.cpp:44 (most recent call first):
#24 437.7 frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x88 (0x7f272de18738 in /usr/local/lib/python3.10/dist-packages/torch/lib/libc10.so)
#24 437.7 frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) + 0x55 (0x7f272ddb42ed in /usr/local/lib/python3.10/dist-packages/torch/lib/libc10.so)
...
#24 437.7 frame #7: at::cuda::getCurrentDeviceProperties() + 0x9 (0x7f270b5874e9 in /usr/local/lib/python3.10/dist-packages/torch/lib/libtorch_hip.so)
#24 437.7 frame #8: at::cuda::warp_size() + 0x9 (0x7f270b587509 in /usr/local/lib/python3.10/dist-packages/torch/lib/libtorch_hip.so)
#24 437.7 frame #9: <unknown function> + 0x81ac8b (0x7f2709c27c8b in /usr/local/lib/python3.10/dist-packages/torch/lib/libtorch_hip.so)

Explanation

80cca70 created a static global variable that used at::cuda::warp_size() to initialize its value, which needs GPUs to be visible to query device properties. However, GPUs are not present on CPU-only build systems.

Solution

Convert static variable into a static function, thus preventing static initialization.

Validation

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:

root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=

@rocm-repo-management-api
Copy link

rocm-repo-management-api bot commented Jun 25, 2025 •
edited
Loading

Jenkins build for 9991022d48d5480423fc3dc1d3b0fb93cdaa638a commit finished as FAILURE
Links: Blue Ocean view / Build artifacts

@rocm-repo-management-api
Copy link

rocm-repo-management-api bot commented Jun 25, 2025

Jenkins build for 9991022d48d5480423fc3dc1d3b0fb93cdaa638a commit is in progress
Links: Blue Ocean view / Build artifacts

jeffdaily
jeffdaily approved these changes Jun 25, 2025
View reviewed changes
Copy link
Collaborator

@jeffdaily jeffdaily left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I like it. Please submit upstream PR as well.

@jeffdaily
Copy link
Collaborator

jeffdaily commented Jun 25, 2025 •
edited
Loading

I like it. Please submit upstream PR as well.

Actually, can you upstream this in combination with 80cca70.

@jeffdaily jeffdaily marked this pull request as ready for review June 25, 2025 20:43
@jithunnair-amd jithunnair-amd changed the title [Rocm7.0_internal_testing] Fix warpsize for 80cca7006d94df97ee932fd5903ed20c08c2eb34 [rocm7.0_internal_testing] Fix warpsize for 80cca7006d94df97ee932fd5903ed20c08c2eb34 Jun 25, 2025
@jithunnair-amd jithunnair-amd changed the title [rocm7.0_internal_testing] Fix warpsize for 80cca7006d94df97ee932fd5903ed20c08c2eb34 [rocm7.0_internal_testing] Prevent static initialization of at::cuda::warp_size() Jun 25, 2025
@jithunnair-amd jithunnair-amd merged commit 944be5a into rocm7.0_internal_testing Jun 25, 2025
0 of 3 checks passed
@jithunnair-amd jithunnair-amd deleted the rocm7.0_IT_fix_warpsize branch June 25, 2025 22:10
xinyazhang pushed a commit that referenced this pull request Jul 2, 2025
@ethanwee1 @xinyazhang
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
8a8b0c4 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
xinyazhang pushed a commit that referenced this pull request Jul 2, 2025
@ethanwee1 @xinyazhang
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
ed63a00 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
xinyazhang pushed a commit that referenced this pull request Jul 2, 2025
@ethanwee1 @xinyazhang
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
e88dd95 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
xinyazhang pushed a commit that referenced this pull request Jul 2, 2025
@ethanwee1 @xinyazhang
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
fd2a043 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
xinyazhang pushed a commit that referenced this pull request Jul 2, 2025
@ethanwee1 @xinyazhang
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
20afd57 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
xinyazhang pushed a commit that referenced this pull request Jul 2, 2025
@ethanwee1 @xinyazhang
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
3c1c9e8 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
xinyazhang pushed a commit that referenced this pull request Jul 2, 2025
@ethanwee1 @xinyazhang
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
bc6edec 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
jithunnair-amd pushed a commit that referenced this pull request Jul 2, 2025
@ethanwee1 @jithunnair-amd
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
a1ad153 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
jithunnair-amd pushed a commit that referenced this pull request Jul 2, 2025
@ethanwee1 @jithunnair-amd
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
6b7dbfa 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
jithunnair-amd pushed a commit that referenced this pull request Jul 2, 2025
@ethanwee1 @jithunnair-amd
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
4d586cd 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
@xinyazhang xinyazhang mentioned this pull request Jul 7, 2025
Merged
xinyazhang pushed a commit that referenced this pull request Jul 8, 2025
@ethanwee1 @xinyazhang
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
d6b4612 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
ethanwee1 added a commit that referenced this pull request Sep 10, 2025
@ethanwee1
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
8fe3525 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
ethanwee1 added a commit that referenced this pull request Sep 10, 2025
@ethanwee1
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
31649d8 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
ethanwee1 added a commit that referenced this pull request Sep 11, 2025
@ethanwee1
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
be6fe86 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
ethanwee1 added a commit that referenced this pull request Sep 11, 2025
@ethanwee1
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
960cc08 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
jithunnair-amd pushed a commit that referenced this pull request Sep 15, 2025
@ethanwee1 @jithunnair-amd
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
7df2fc2 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
jithunnair-amd pushed a commit that referenced this pull request Sep 15, 2025
@ethanwee1 @jithunnair-amd
[rocm7.0_internal_testing] Prevent static initialization of at::cuda:…
87a4931 
…:warp_size() (#2293)

Fixes SWDEV-540240, SWDEV-540309, SWDEV-539989

```
...
```

80cca70
created a static global variable that used `at::cuda::warp_size()` to
initialize its value, which needs GPUs to be visible to query device
properties. However, GPUs are not present on CPU-only build systems.

Convert static variable into a static function, thus preventing static
initialization.

http://rocm-ci.amd.com/job/pyt_whl_docker_mainline/1461/artifact/build_artifacts.txt/*view*/

Ran microbenchmark to confirm basic functionality:
```
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10158218145370483
Throughput [img/sec] : 630.0317544289736=
```
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@jeffdaily jeffdaily jeffdaily approved these changes

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@ethanwee1 @jeffdaily @jithunnair-amd