Skip to content

Don't assume that a CUDA CC only contains two numbers, we want to also support CC100, CC120, etc #128

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

Merged
bedroge merged 2 commits into from
Nov 18, 2025
Merged

Don't assume that a CUDA CC only contains two numbers, we want to also support CC100, CC120, etc #128

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
8dde4c2
Don't assume that a CUDA CC only contains two numbers, we want to als…
Nov 13, 2025
aae5169
Enable by default the architecture-specific optimizations for 9.0, an…
Nov 13, 2025
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
16 changes: 14 additions & 2 deletions EESSI-extend-easybuild.eb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -96,12 +96,24 @@ sysroot = os.getenv("EESSI_EPREFIX")
-- Check if we have GPU capabilities and configure CUDA compute capabilities
eessi_accelerator_target = os.getenv("EESSI_ACCELERATOR_TARGET")
if (eessi_accelerator_target ~= nil) then
cuda_compute_capability = string.match(eessi_accelerator_target, "^accel/nvidia/cc([0-9][0-9])$")
cuda_compute_capability = string.match(eessi_accelerator_target, "^accel/nvidia/cc([0-9]+)$")
if (cuda_compute_capability ~= nil) then
easybuild_cuda_compute_capabilities = cuda_compute_capability:sub(1, 1) .. "." .. cuda_compute_capability:sub(2, 2)
-- The last digit should be the minor version, insert a dot in the one-but-last position
major_version = cuda_compute_capability:sub(1, #cuda_compute_capability - 1)
minor_version = cuda_compute_capability:sub(#cuda_compute_capability)
easybuild_cuda_compute_capabilities = string.format("%s.%s", major_version, minor_version)
else
LmodError("Incorrect value for $EESSI_ACCELERATOR_TARGET: " .. eessi_accelerator_target)
end

-- If architectures are 9.0, 10.0 or 12.0, enable architecture or family-specific optimizations
if easybuild_cuda_compute_capabilities == '9.0' then
easybuild_cuda_compute_capabilities = '9.0a'
elseif easybuild_cuda_compute_capabilities == '10.0' then
easybuild_cuda_compute_capabilities = '10.0f'
elseif easybuild_cuda_compute_capabilities == '12.0' then
easybuild_cuda_compute_capabilities = '12.0f'
end
end

-- Some environment variables affect behaviour, let's gather them once
Expand Down
4 changes: 2 additions & 2 deletions init/eessi_archdetect.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -165,11 +165,11 @@ accelpath() {
# If EESSI_ACCELERATOR_TARGET_OVERRIDE is set, use it
log "DEBUG" "accelpath: Override variable set as '$EESSI_ACCELERATOR_TARGET_OVERRIDE' "
if [ ! -z $EESSI_ACCELERATOR_TARGET_OVERRIDE ]; then
if [[ "$EESSI_ACCELERATOR_TARGET_OVERRIDE" =~ ^accel/nvidia/cc[0-9][0-9]$ ]]; then
if [[ "$EESSI_ACCELERATOR_TARGET_OVERRIDE" =~ ^accel/nvidia/cc[0-9]+$ ]]; then
echo ${EESSI_ACCELERATOR_TARGET_OVERRIDE}
return 0
else
log "ERROR" "Value of \$EESSI_ACCELERATOR_TARGET_OVERRIDE should match 'accel/nvidia/cc[0-9[0-9]', but it does not: '$EESSI_ACCELERATOR_TARGET_OVERRIDE'"
log "ERROR" "Value of \$EESSI_ACCELERATOR_TARGET_OVERRIDE should match 'accel/nvidia/cc[0-9]+', but it does not: '$EESSI_ACCELERATOR_TARGET_OVERRIDE'"
fi
return 0
fi
Expand Down