Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add benchmark noise reducing info. #5924

Merged
merged 3 commits into from
Nov 29, 2023
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 21 additions & 0 deletions benchmarks/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,27 @@ The two main benchmarking scripts are
- `result_analyzer.py` to aggregate the benchmark result in CSV form.


## Reducing benchmark noise

It is important to keep the benchmark runs safe from external effects
to reduce noise. Do the following:

Sets the CPU statically to the highest tuneable frequency.
Prevent energy saving features to kick in.

```sudo cpupower frequency-set --governor performance```

Lock GPU clocks to lower frequency to reduce the chance of thermal throttling. Choose
FREQ based on your GPU info. To find out clock frequency on your device run:
`nvidia-smi -q -d CLOCK`, and look for Graphics/SM in Max Clocks section.
Setting the clock a couple hundrend MHz below, or ~80% of max
will most likely prevent thermal throttling effects.

```FREQ=... nvidia-smi --lock-gpu-clocks=$FREQ,$FREQ```

Disable autoboost selecting clock rate based on thermal, and power budget effects.
```CUDA_AUTO_BOOST=0```

## Experiment runner

Run the `experiment_runner.py` from the `pytorch` directory, which should be the
Expand Down