Closed jgiannuzzi closed 2 weeks ago
[APPROVALNOTIFIER] This PR is APPROVED
This pull-request has been approved by: jgiannuzzi, sanposhiho
The full list of commands accepted by this bot can be found here.
The pull request process is described here
What type of PR is this?
/kind cleanup
What this PR does / why we need it:
The
nodenumber
WASM plugin used in the scheduler_perf benchmark doesn't use advanced techniques like the custom GC and advertising only the implemented endpoints. By adopting these techniques and rebuilding the WASM module, the benchmark suggests that the WASM plugin performance is between the native code and the extender, whilst it was slower than the extender before this change.Which issue(s) this PR fixes:
Special notes for your reviewer:
I have also changed the location of the WASM module in the
AllEnabled
profile, as it was pointing to thenodenumber
example rather than the one we build for the benchmark.Does this PR introduce a user-facing change?
What are the benchmark results of this change?
Running the
scheduler_perf
benchmark 3 times on the same machine and taking the average of the results for scheduling 10000 pods, this results in a ~10% increase, going from 750 pods/s to 833 pods/s.