John Biddiscombe
John Biddiscombe
Just FYI : In the verbs parcelport branch, I have modified the concurrent::unordered_map to make use of a new readers_writers_lock mutex and unique_lock/scoped_lock - this gives us single writer and...
BTW. I did not see the discussion here, but last night I created a branch and made a PR #2629 because I need to use this concurrent map in code...
@SHANTAM29 if you do begin work on it, please get in touch with me via email or other means for a chat. I have been poking around with this recently.
I suspect that cancellation tokens and support for them needs to be kept open. One solution to this problem (in a broader sense) would be to only enable cancellation of...
closeComment: false
This issue look interesting - should it be kept alive?
I was thinking about this recently, but have not worked on it. In the shared-priority scheduler, I reduced the size of chunks of tasks that are stolen and did away...
Discussion on IRC (@sithhell @msimberg) and elsewhere have raised interesting points that need to be itemized and acted on. Our principle performance related problem is fork-join parallelism and being competitive...
The idea of implementing task_regions or fork_join regions to allow us to switch between normal hpx task stealing and fork-join type work for simpler loops is nice, but ultimately, if...
@sithhell - Have you had any thoughts about this - and does the work you are doing with the scheduler go in this kind of direction? Is there another/better way...