From f731cf25731572839a98ec0570023388c04735b7 Mon Sep 17 00:00:00 2001 From: saienduri Date: Tue, 29 Oct 2024 22:38:05 -0500 Subject: [PATCH] update readme --- README.md | 3 +++ 1 file changed, 3 insertions(+) diff --git a/README.md b/README.md index ec5ff55..e2ff214 100644 --- a/README.md +++ b/README.md @@ -14,6 +14,9 @@ This repo is configured with a custom Github runner donated by AMD. You can queu The main things you need to run your own benchmark 1. In `kernels/` create a new file that must start with the name `test_`. This is because we use `pytest` to discover your kernel 2. If you want your benchmark results to persist in a Github Artifact, we recommend using the builtin Triton `benchmark.run(save_path="./perf-artifacts/your_kernel", show_plots=True, print_data=True)` +3. In your PR, if you don't want to run testing on all the kernels, you can specify a specific kernel you want +to test by adding a line like the following to your PR description: "ci-exactly: " as seen +in this PR: [Example](https://github.com/gpu-mode/amd-cluster/pull/3) Have fun! We intend for this to be a social repo, if you have any other requests for things we could do better please let us know!