x86 (x86_64) clusters¶
Example: a standard x86_64 SLURM cluster (e.g. AMD EPYC nodes). Define it in your
registry and submit by its --system name.
Submit¶
labflow submit my_experiment --system cluster-x86 --partition normal \
--nodes 2 --time 2:00:00 --account <account>
Notes¶
- Manylinux wheels generally work; a venv or conda env is usually enough (no need to compile from source as on ARM).
- Tune
-march/-mtuneto the node CPU if you do build native extensions.
See HPC systems registry for the entry schema.