RIKEN Benchmark Status

[dyokelson]

Here is a table of all the RIKEN benchmarks and systems and whether we at LLNL have been able to successfully run them

Definition	Legacy works on LLNL	New works on LLNL	New works on Fugaku
Fugaku System	N/A	N/A	Tested+Merged (PR #451)
QWS	Yes	Yes	Tested+Merged (PR #387)
Genesis	Yes	Yes	PR #534
Salmon	Yes (PR #391)	Not Started	Not Started

Salmon fails during compilation (possibly due to not having fj compiler available on LLNL systems). Steps are included for reproducing error on the PR.

Fugaku system.py generates yaml that looks correct and passes a dry run, this needs to be tested on Fugaku by actually running experiments once the system is defined, steps are listed on the PR.

QWS new experiment.py runs on LLNL CTS system Ruby, passes dry run, steps for testing on Fugaku are included in the PR.

TODO:

• Test Fugaku system.py on Fugaku (with existing benchmark)
• Test QWS experiment.py on Fugaku (with new Fugaku system.py)
• Rewrite Genesis in the experiment.py format
• Get legacy Salmon running on LLNL machines