[Merged by Bors] - chore: remove CoeFun instances where FunLike is available#17911
[Merged by Bors] - chore: remove CoeFun instances where FunLike is available#17911Vierkantor wants to merge 3 commits intomasterfrom
Conversation
During the port we found that `FunLike` is robust enough not to need an extra `CoeFun` shortcut. Let's see if that rule can be consistently applied to the whole of the library. There is still duplication between `FunLike` and `CoeFun` for `Grp`, `Mon`, `CommGrp` and `CommMon`, which will need a more thorough fix. See also #17866. This is the first half of the changes, to see if those work without slowdowns.
|
!bench |
PR summary 7a06803237Import changes for modified filesNo significant changes to the import graph Import changes for all files
|
|
Here are the benchmark results for commit 7a06803. |
|
This PR/issue depends on: |
Indeed, but we also found that the shortcut could often improve performance, as you are now running into! |
|
@eric-wieser Could you please explain what you mean? Because the bench bot says there is no significant performance regression. |
|
I think "often" is compatible with what Anne is finding: removing the shortcut entirely causes a slowdown. Let's just merge the parts that don't cause a slow down, and soon Anne will hopefully have localised the places where removing it really is damaging, and then we can study that. |
|
bors merge |
During the port we found that `FunLike` is robust enough not to need an extra `CoeFun` shortcut. Let's see if that rule can be consistently applied to the whole of the library. There is still duplication between `FunLike` and `CoeFun` for `Grp`, `Mon`, `CommGrp` and `CommMon`, which will need a more thorough fix. See also #17866. I am currently bisecting to figure out exactly why the benchmarks are indicating a slowdown with no obvious cause.
|
Pull request successfully merged into master. Build succeeded: |
During the port we found that
FunLikeis robust enough not to need an extraCoeFunshortcut. Let's see if that rule can be consistently applied to the whole of the library.There is still duplication between
FunLikeandCoeFunforGrp,Mon,CommGrpandCommMon, which will need a more thorough fix. See also #17866.I am currently bisecting to figure out exactly why the benchmarks are indicating a slowdown with no obvious cause.
CoeFuninstances whereFunLikeis available #17900