Justin M Wozniak
Justin M Wozniak
That sounds good.
They will read that as Deep Learn only :) . How about --data-dir ? Will that be a standard flag for all Benchmark invocation? The default will be the current...
Yes, those are fine.
Assign to @rajeeja
module ibm-wml requires a different module naming scheme
Yes, I think that should be possible. For example, on some machines I want to clone the Benchmarks in my home directory (fully backed up, small quota) but keep the...
That was right, the Python installation was not accessible, it works now, thanks.
The root cause was that I put all the Flux dependencies in a Miniconda that was not visible from the compute nodes.
I don't think the flux processes are finding each other yet: I get: ``` $ mpiexec -n 2 --pmi=pmi flux start 'flux resource list' STATE NNODES NCORES NGPUS NODELIST flux...
``` $ mpiexec -n 2 --pmi=pmi flux pmi -v barrier flux-pmi-client: trying 'simple' simple: PMI_FD not found in environ flux-pmi-client: trying 'libpmi2' libpmi2: libpmi2.so: cannot open shared object file: No...