Faster minmax by using fold
#968
Conversation
Do less comparisons by considering pairs is smart but this was done by `loop { it.next(), it.next() }` while `it.fold` is generally faster.
This results in a lot of changes but most of it is merely moving code and indentation corrections.
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #968 +/- ##
==========================================
+ Coverage 94.38% 94.67% +0.28%
==========================================
Files 48 49 +1
Lines 6665 7232 +567
==========================================
+ Hits 6291 6847 +556
- Misses 374 385 +11 ☔ View full report in Codecov by Sentry. |
There was a problem hiding this comment.
Thanks. I think I get the idea. fold pairwise and see if there's one excess element that has to be processed manually.
Do you have some numbers that show that the proposed implementation indeed is faster? (My gut thought: For a slide, the existing implementation can just call next twice, whereas your proposal has to match acc twice, even though that seems unneeded.)
I'll approve this so that you can merge if you want. But I'd be happy about benchmarks that substantiate that we're taking the right direction.
No description provided.