superbenchmark/tests
guoshzhao f38a9829d0
ModelBenchmarks - Fix early stop logic due to num_steps. (#522)
**Description**
Model benchmarks can stop due to `num_steps` or `duration` config which
will take effect when the value is set greater than 0.
If both are set greater than 0, the earliest condition reached will
work.
2023-04-28 13:15:47 +08:00
..
analyzer Release - SuperBench v0.6.0 (#409) 2022-09-06 18:06:05 +08:00
ansible Release - SuperBench v0.4.0 (#278) 2021-12-30 16:24:00 +08:00
benchmarks ModelBenchmarks - Fix early stop logic due to num_steps. (#522) 2023-04-28 13:15:47 +08:00
cli Release - SuperBench v0.7.0 (#468) 2023-01-28 11:07:06 +08:00
common Runner - Generate host groups file in mpi mode (#458) 2023-01-04 19:49:14 +08:00
data Adding HPL benchmark (#482) 2023-03-21 16:44:08 +00:00
executor Config - Update benchmark naming to support annotations (#284) 2022-01-25 09:54:58 +00:00
helper Release - SuperBench v0.4.0 (#278) 2021-12-30 16:24:00 +08:00
monitor Release - SuperBench v0.8.0 (#517) 2023-04-14 12:57:55 +00:00
runner Runner: Support `topo-aware` and `k-batch` pattern in 'mpi' mode (#437) 2023-01-03 10:28:35 +00:00
__init__.py Benchmarks: Revise Test - Revise benchmark test util to support pytorch multi-GPU test (#54) 2021-04-14 16:39:53 +08:00