Setup JMH #71
Setup JMH #71
Conversation
narangvivek10
added
improvement
|
/ok to test e0f7b03 |
|
/ok to test 61fc3b7 |
| */ | ||
| package com.nvidia.cuvs.lucene.benchmarks; | ||
|
|
||
| import static com.nvidia.cuvs.lucene.benchmarks.Utils.cleanup; |
There was a problem hiding this comment.
Rather than having yet another set of benchmarks in here, I'd rather us put these in cuvs-bench so that we have a single way of running and reproducing these. One of the cuvs engineers is currently working to make cuvs-bench more pluggle and generalized so that we can plug these in more easily. We want to avoid having many different codes for running benchmarks sitting around. It becomes confusing to users and introduces more ways for us to find (and have to explain) various deltas between expectations and reality... just altogether easier for us to have a single way to run them and a single source of truth.
There was a problem hiding this comment.
Of course, if these are meant to be treated more like microbenchmarks for profiling / perf tuning, then that's completely different. In that case, I'd put these in a directory called bench/ instead of benchmarks/ to match cuvs, raft and other repositories.
There was a problem hiding this comment.
These are meant to be microbenchmarks. Renaming the directory to bench.
jameslamb
left a comment
jameslamb
left a comment
There was a problem hiding this comment.
Giving this a ci-codeowners approval, the update-version.sh changes look small and non-controversial.
|
/ok to test a87e7f7 |
3 participants
Fixes #41