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
We need a way to differentiate externals that work on different types of hardware.
When I log in to Izar I see (at least) an external that does not run on GPU and I'm about to add an external that runs only on GPU but that will be available on other clusters as well.
Perhaps in the external_packages we need boolean keywords for:
runs_on_cpu
runs_on_gpu
So that we can end up with three variations of externals:
codes that work on CPU but may benefit of GPU acceleration (e.g. Gaussian);
codes that run on CPU only (likely the majority);
codes that depend on GPUs and don't have a CPU-only counterpart (e.g. TeraChem)
This would hopefully interact with platforms/<version>/<cluster>.yaml and take an action based on the gpu filter.
The text was updated successfully, but these errors were encountered:
We need a way to differentiate externals that work on different types of hardware.
When I log in to Izar I see (at least) an external that does not run on GPU and I'm about to add an external that runs only on GPU but that will be available on other clusters as well.
Perhaps in the
external_packages
we need boolean keywords for:So that we can end up with three variations of externals:
This would hopefully interact with
platforms/<version>/<cluster>.yaml
and take an action based on thegpu
filter.The text was updated successfully, but these errors were encountered: