ARCHER Benchmarks - Initial Application Benchmark Results
The ARCHER CSE team at EPCC recently worked with major ARCHER users to develop an updated ARCHER benchmarking suite to reflect the range of computational work undertaken on the service.
This work led to a proposal for an updated set of ARCHER benchmarks that has recently been published as an ARCHER white paper:
We have recently started to run these new application benchmarks on ARCHER and have now published the initial set of results on the ARCHER website at:
This page will continue to be updated with additional data as soon as it becomes available.
Our plans for these benchmarks over the next few months include:
- Running them across as many different HPC systems and architectures that we can gain access to.
- Performing period runs of the benchmarks to assess the range of performance that users may experience on ARCHER.
- Expanding the results by using different process and thread counts where possible.