Peter Scheibel

My reasoning behind doing this vs. encoding arithmetic into ramble.yaml configs: I think it has an advantage in the following cases: * Users change how they ask for resources: I...

@douglasjacobsen pointed out * this may be better handled with modifiers (and I think @becker33 understood this suggestion and liked it as an alternative as well). * with the approach...

When you ran `benchpark setup ` did you run `source /setup.sh` ? if so, then `SPACK_DISABLE_LOCAL_CONFIG` should be set in your shell. If you see that shell variable, but are...

This is recorded as https://github.com/spack/spack/issues/38664 in the Spack repository (the issue is in-progress)

I think this is actually missing from the Spack `stream` package definition (I see it does not set `-mcmodel` as a Fortran flag). > unless spack populates the fflags with...

Example script generated from `experiments/saxpy/openmp` and `configs/nosite-x86_64` (tweaked to assume slurm, for a more interesting output): ``` #SBATCH -n 8 #SBATCH -N 1 #SBATCH --time 120 cd /test-saxpy-oslic-new/saxpy/openmp/nosite-x86_64/workspace/experiments/saxpy/problem/saxpy_512_1_2 rm -f...

We should be recording this somewhere, but even though Benchpark enables the Spack options for this, I don't see the associated logs being generated, so I'll need to look into...

Also: any flags that don't appear in the log you listed would have to be `package.py` for `stream`, or in your `compilers.yaml` config, so do you have a path forward...

> We should probably define a graded approach for generating these: > only introduce a new hardware specification if one like it indeed does not exist. To be clear, are...