benchpark
benchpark copied to clipboard
LLNL
An open collaborative repository for reproducible specifications of HPC benchmarks and cross site benchmarking environments
Updated: - [X] @scheibelp implemented an instantiation of a system which `benchpark setup` can use. - [x] @scheibelp, please implement a way to instantiate an experiment which `benchpark setup` can...
Adding GPCNet benchmarks, network_test and network_load_test. FOMs are TBD. Potentially closes #322.
Progress on sandia microbenchmarks. So far, both mpi_overhead and base msgrate workloads work, although I'm not finished setting up the rma and shm variants for msgrate in spack. This is...
This adds a new system definition you can select with `benchpark system init` to generate a refined version of https://github.com/LLNL/benchpark/tree/develop/configs/LLNL-Sierra-IBM-power9-V100-Infiniband, e.g. `benchpark system init --basedir=test-lassen-create lassen cuda=10-1-243 compiler=xl` The system...
bin/benchpark experiment init --dest laghos Don't believe there's any variants that need to be specified here
Rewriting `qws` as an experiment object.
Adding osu-microbenchmarks experiment.py file.
This is the Kripke PR, for review by Riyaz
- [ ] Variables @dyokelson - [ ] Type of study (strong scaling, weak scaling) @dyokelson - [ ] Process binding; we should use https://github.com/bcumming/affinity @daboehme, can this be used...
The commands listed [here](https://software.llnl.gov/benchpark/2-benchpark-list.html) need to be updated to traverse and return values from any new structures before the old configs are removed.
