Skip to content

Update cuda_cc flags for newer ELPA versions #4186

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
yqshao wants to merge 2 commits into
base: develop
Choose a base branch
from
Open

Update cuda_cc flags for newer ELPA versions #4186

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
4fbb4e6
Update cuda_cc flags for newer ELPA versions
yqshao Jul 23, 2026
816a60f
Bring back and reword some comments
yqshao Jul 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
24 changes: 15 additions & 9 deletions easybuild/easyblocks/e/elpa.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -200,17 +200,23 @@ def configure_step(self):
'cuda_compute_capabilities easyconfig parameter or via '
'--cuda-compute-capabilities')

# ELPA's --with-NVIDIA-GPU-compute-capability only accepts a single architecture
if len(cuda_cc) != 1:
raise EasyBuildError('ELPA currently only supports specifying one CUDA architecture when '
# ELPA's --with-NVIDIA-GPU-compute-capability only accepts a single architecture until 2025.06.002
if LooseVersion(self.version) < LooseVersion('2025.06.002') and len(cuda_cc) != 1:
raise EasyBuildError('ELPA prior to 2025.06.002 only supports specifying one CUDA architecture when '
'building. You specified cuda-compute-capabilities: %s', cuda_cc)
cuda_cc = cuda_cc[0]
cuda_cc_string = cuda_cc.replace('.', '')
self.cfg.update('configopts', '--with-NVIDIA-GPU-compute-capability=sm_%s' % cuda_cc_string)

cuda_cc_string = ','.join(['sm_' + x.replace('.', '') for x in cuda_cc])
self.cfg.update('configopts', '--with-NVIDIA-GPU-compute-capability=%s' % cuda_cc_string)
self.log.info("Enabling nvidia GPU support for compute capability: %s", cuda_cc_string)
# There is a dedicated kernel for sm80, but only from version 2021.11.001 onwards
# Trying to use these kernels for GPUs newer than sm80 will fail ELPA configure
if float(cuda_cc) == 8.0 and LooseVersion(self.version) >= LooseVersion('2021.11.001'):

# There are dedicated kernels for sm80, but only from version 2021.11.001 onwards
# Trying to use these kernels for GPUs newer than sm80 will fail ELPA configure until 2025.06.002
# Since 2025.06.002 the max compute capability is used to determine whether sm80 kernel may be used
if LooseVersion(self.version) >= LooseVersion('2025.06.002') and max(map(float, cuda_cc)) >= 8.0:
# The configuration option has been renamed since 2023.05.001,
# though the old option is accepted at least until 2025.06.002
self.cfg.update('configopts', '--enable-nvidia-sm80-gpu-kernels')
elif LooseVersion(self.version) >= LooseVersion('2021.06.002') and float(cuda_cc[0]) == 8.0:
self.cfg.update('configopts', '--enable-nvidia-sm80-gpu')

# From v2022.05.001 onwards, the config complains if CPP is not set, resulting in non-zero exit of configure
Expand Down