libfork icon indicating copy to clipboard operation
libfork copied to clipboard

Performance, benchmarks, ...

Open dumblob opened this issue 2 years ago • 2 comments

Hi, I've just discovered this potential gem. But to be sure the features reflect in practical setting, we'd need to test it first.

Do you have any such performance benchmarks (even if not rigorous...)?

I'd be mostly interested in direct comparison with Nim's Weave which I myself consider as state-of-the-art library. So ideally the benchmarks could be aligned with Weave's ones :wink:.

Thoughts?

dumblob avatar Sep 15 '21 08:09 dumblob

This and zero dynamic allocation are my top priorities, unfortunately I'm away from my desk for the next 3 weeks. I welcome any PR's

ConorWilliams avatar Sep 15 '21 14:09 ConorWilliams

Take your time - I'm also very busy, so don't expect any such well thought-through PR from me now :wink:.

dumblob avatar Sep 15 '21 18:09 dumblob

I just noticed a lot is going on in v2 branch.

Do you have any writeup about the goals and ideas for v2 (and why actually v2 and not improving v1 :wink:)?

dumblob avatar Feb 17 '23 15:02 dumblob

Key changes are removing the global scheduler, allowing custom schedulers, allowing tasks to be run inline and overhaul of cmake+docs+benchmarks. Its semver so v2 as API break is required.

ConorWilliams avatar Feb 19 '23 10:02 ConorWilliams

Sounds great! Will keep an eye on the direction here. Feel free to ping me once there is something to publish :wink:.

dumblob avatar Feb 19 '23 21:02 dumblob

Version 2 has released with benchmarks

ConorWilliams avatar Feb 27 '23 12:02 ConorWilliams