Make EESSI-extend support accelerator installations
#27
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ocaisa
commented
Jul 3, 2025
| Strict installation path checking is enforced by EESSI for EESSI and site | ||
| installations involving accelerators. In these cases, if you wish to create an | ||
| accelerator installation you must set the environement variable | ||
| EESSI_ACCELERATOR_INSTALL (and load/reload this module). |
There was a problem hiding this comment.
This new environment variable has an impact on the build scripts, it needs to be set in the scenario where we expect to do an accelerator installation
| @@ -29,8 +29,8 @@ fi | |||
| if [ ! -z ${EESSI_SOFTWARE_SUBDIR_OVERRIDE} ]; then | |||
| INPUT="export EESSI_SOFTWARE_SUBDIR_OVERRIDE=${EESSI_SOFTWARE_SUBDIR_OVERRIDE}; ${INPUT}" | |||
| fi | |||
| if [ ! -z ${EESSI_ACCELERATOR_TARGET} ]; then | |||
| INPUT="export EESSI_ACCELERATOR_TARGET=${EESSI_ACCELERATOR_TARGET}; ${INPUT}" | |||
| if [ ! -z ${EESSI_ACCELERATOR_TARGET_OVERRIDE} ]; then | |||
There was a problem hiding this comment.
@trz42 This is why I was asking about where these environment variables get set, this should be using the override mechanism
There was a problem hiding this comment.
ocaisa
commented
Jul 3, 2025
| module load EESSI-extend/${{matrix.EESSI_VERSION}}-easybuild | ||
| check_env_var "EASYBUILD_INSTALLPATH" "$EESSI_SOFTWARE_PATH" # installation path should be the same unless we ask for an explicit GPU installation | ||
| check_env_var "EASYBUILD_CUDA_COMPUTE_CAPABILITIES" "$STORED_CUDA_CC" | ||
| export EESSI_ACCELERATOR_INSTALL=1 |
There was a problem hiding this comment.
It is this variable (EESSI_ACCELERATOR_INSTALL) and EESSI_ACCELERATOR_TARGET_OVERRIDE that need to be set by the bot in order to configure EESSI-extend correctly for a GPU installation
ocaisa
commented
Jul 4, 2025
| if (eessi_accelerator_target ~= nil) then | ||
| cuda_compute_capability = string.match(eessi_accelerator_target, "^nvidia/cc([0-9][0-9])$") | ||
| if (cuda_compute_capability ~= nil) then | ||
| easybuild_installpath = pathJoin(easybuild_installpath, 'accel', eessi_accelerator_target) |
There was a problem hiding this comment.
This was actually wrong, archdetect returns paths like accel/nvidia/cc80 (see https://github.com/EESSI/software-layer-scripts/blob/main/tests/archdetect/nvidia-smi/1xa100.output), but we were consistent in our error (the build script set EESSI_ACCELERATOR_TARGET incorrectly rather than set EESSI_ACCELERATOR_TARGET_OVERRIDE which would have affected the behaviour of archdetect).
ocaisa
commented
Jul 4, 2025
| export EESSI_ACCELERATOR_TARGET=$(cfg_get_value "architecture" "accelerator") | ||
| echo "bot/build.sh: EESSI_ACCELERATOR_TARGET='${EESSI_ACCELERATOR_TARGET}'" | ||
| export EESSI_ACCELERATOR_TARGET_OVERRIDE="accel/$(cfg_get_value architecture accelerator)" | ||
| echo "bot/build.sh: EESSI_ACCELERATOR_TARGET_OVERRIDE='${EESSI_ACCELERATOR_TARGET_OVERRIDE}'" |
There was a problem hiding this comment.
ocaisa
commented
Jul 4, 2025
| if [[ -n "$EESSI_ACCELERATOR_TARGET_OVERRIDE" && -z "$EESSI_ACCELERATOR_TARGET" ]]; then | ||
| fatal_error "EESSI module should've set EESSI_ACCELERATOR_TARGET when EESSI_ACCELERATOR_TARGET_OVERRIDE exported." >&2 | ||
| elif [[ -n "$EESSI_ACCELERATOR_TARGET_OVERRIDE" ]]; then | ||
| export EESSI_ACCELERATOR_INSTALL=1 |
There was a problem hiding this comment.
|
bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws arch:x86_64/amd/zen2 |
|
New job on instance
|
|
bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws arch:x86_64/amd/zen2 |
|
New job on instance
|
ocaisa
commented
Jul 4, 2025
ocaisa
commented
Jul 4, 2025
ocaisa
commented
Jul 4, 2025
|
bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws arch:x86_64/amd/zen2 |
|
New job on instance
|
…scripts into eessi-extend-cuda
|
bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws arch:x86_64/amd/zen2 |
|
New job on instance
|
bedroge
reviewed
Jul 4, 2025
Co-authored-by: Bob Dröge <b.e.droge@rug.nl>
|
bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws arch:x86_64/amd/zen2 |
|
New job on instance
|
|
bot: build repo:eessi.io-2023.06-software instance:eessi-bot-mc-aws arch:x86_64/amd/zen3 |
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
|
New job on instance
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.