You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm trying out pixi to try and make the build process no longer rely on custom linux modules and shell scripts that are unique to our HPC cluster environment.
Get tasks working for the current build system (with the nvidia module already loaded if necessary) and commit that first to make it easier to test with the current system
Replace env.set.local0 with pixi tasks except for loading nvidia-sdk module
Replace nvidia-sdk module in pixi task
Keep athena compilers working with pixi installed compilers
Add openmpi
Try pixi verbose option
Remove extraneous environmental variables
Get srun command without module load
Swap globally installed nvc++ for nvcc, then pixi installed nvcc using CXX
Remove prepended environmental variables to fall back to conda installed libraries to see if pixi can build SlaterGPU on top of its conda dependencies
Test whether you can write to a variable and use it all under the env section in pixi.toml for a task
Deal with libcint dependency being hardcoded to a prebuilt filepath on our HPC -> look into rattler-build and or pixi build as a way to build libcint automatically
The text was updated successfully, but these errors were encountered:
I'm trying out pixi to try and make the build process no longer rely on custom linux modules and shell scripts that are unique to our HPC cluster environment.
The text was updated successfully, but these errors were encountered: