pearce8
pearce8
Adding the unet experiment using LBANN, also with Caliper timings.
Updated: Below the dashed line, we have the *original* system config design. Things have changed since then, and we need to: - What should the user use as a starter...
Types of experiments - [x] The current Problem 1 experiment should be named "throughput" (varying the size of the problem, keeping resources the same). x-axis label: problem size. Add plots...
Current priorities: - [ ] package.py - test - [ ] application.py - [ ] cuda experiment - a single experiment - [ ] test the 3 above, add output...
@gracenansamba should this PR be closed?
- [ ] Thread pinning - [ ] GPU mapping - [ ] Enabling GPU-aware MPI - [ ] Running bonded interactions on GPUs
Test https://github.com/GoogleCloudPlatform/ramble/discussions/511 and move the pinned Ramble version to v0.5.0 (and corresponding Spack) when all is in the clear.
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...