[Question] Performance vs HPX library

[pfeatherstone]

Hi. I've just noticed there's been a fresh release of HPX and had completely forgotten about it. Have you benchmarked against it?
Also, would it be worth opening up github discussions?
Cheers

[tzcnt]

I'd love to benchmark against HPX... and would still like to try against libdispatch / Grand Central Dispatch as you previously suggested :)

However I have a lot of plates spinning already so it would have to happen when the whim strikes me. I'd happily accept contributions to https://github.com/tzcnt/runtime-benchmarks if you want to take a stab at this yourself? I've found that most of the benchmarks can be fairly easily cloned, especially if you start with a framework that offers a similar API, which is easier to do now that there are already 8 implemented.

I can also open up discussions if you want to talk about other things, but for comparative benchmarks, feel free to open an issue on the runtime-benchmarks repo instead.

[tzcnt]

I've implemented the HPX benchmarks. Unfortunately the results were quite poor.
tzcnt/runtime-benchmarks#4

New results have been posted to the main pages:
https://github.com/tzcnt/runtime-benchmarks
https://fleetcode.com/runtime-benchmarks/

But I only ran them for the EPYC server so far because it will take too long to run on my other machines.

I opened an issue in the HPX repo to see if there is a way to optimize the implementations better. TheHPXProject/hpx#6793