Viktor Klang (√)
Viktor Klang (√)
Thanks @ktoso! Looks like some nice improvements basically across the board. I am looking at some more potential improvements. Will ping you as soon as I have something to demo....
There are some main things I want to try to tackle: 1) Investigate possibility to cheaply remive Links when the linked promise is completed 2) Investigate possibility of not having...
@ktoso I just pushed a new commit to the wip-optimizations-√ branch, not super happy with the code style, but it seems a bit more performant yet.
@ktoso Absolutely. Fixed. I just pushed a new commit which is pretty cool. Interested to see how it performs on your machine. (https://github.com/viktorklang/scala-futures/commit/a28f2e84b2b1e9fc40c511e16614a645d2651bcf)
@ktoso There's a weird performance regression in firstCompletedOf: ~~~ [info] Benchmark (pool) (recursion) (threads) Mode Cnt Score Error Units [info] FirstCompletedOfBenchmark.improved_post fjp 192 1 thrpt 20 0.499 ± 0.014 ops/ms...
@ktoso Found it. And fixed it. ~~~ [info] Benchmark (pool) (recursion) (threads) Mode Cnt Score Error Units [info] FirstCompletedOfBenchmark.improved_post fjp 192 1 thrpt 20 18.858 ± 0.224 ops/ms [info] FirstCompletedOfBenchmark.improved_pre...
@ktoso Pushed another update.
Looks like that only runs 2 of the benches. Hmmm, I wonder why. Check out master, go to the "benches" project, and run something like: jmh:runMain scala.future.BenchRunner -p threads=1 -p...
:+1: Would be great!
Why do you need to add JAXB annotations on things that are code-representations of a wire protocol?