Update benchmarks #6106
Update benchmarks #6106
Conversation
This is a smoke test to make sure that benchmarks are currently runnable.
|
@pchiusano: You were the last to update the old benchmarks and @pchiusano/misc-benchmarks is your library, so I figure you’re the one who knows whether replacing the old ones is the right thing to do or not. If it’s not, they’ll need to be updated to get them compiling again. |
Makes running benchmarks a bit easier; runs with a higher scheduling priority; and, if used with `sudo`, has an even higher scheduling priority.
| @@ -0,0 +1,9 @@ | |||
| ``` ucm :hide | |||
| > pull @pchiusano/misc-benchmarks | |||
There was a problem hiding this comment.
If there's a particular release of this that we want to pull that would probably be good; if there isn't a release, then let me or Paul know.
There was a problem hiding this comment.
There are currently no releases of @pchiusano/misc-benchmarks.
There was a problem hiding this comment.
@pchiusano Can you create a release of https://share.unison-lang.org/@pchiusano/misc-benchmarks and then tag @sellout
(Unless there's some reason why we don't care about those benchmarks.)
And also if you're in there is it appropriate and easy to change Value.serialize_v4 usages to v5, if appropriate?
There was a problem hiding this comment.
We are trying to move to ghc 9.10 in order to update other parts of the toolchain, but trying to make sure we don't lose some important optimization without realizing it. Here's benchmark outputs Greg found in their tentative ghc 9.10 update branch.
2 participants
Overview
This gets the documented benchmark process working again. It also runs the benchmarks in CI to ensure that they continue to work.
Implementation approach and notes
The benchmarks in benchmarks.md are apparently obsolete, with the current ones being pulled from Unison Share.
Test coverage
This adds coverage of the benchmark suite.
NB: The CI integration is not meant to measure the benchmarks, only to ensure that they continue to compile.