Data parallelism is a very useful approach to parallelism because we can describe what to compute rather than how to compute. It’s especially nice for parallel computing since specifying how to compute often leads to subtle concurrency bugs. However, each user’s each program might have specific requ…

[ANN] FoldsThreads.jl: A zoo of pluggable thread-based data-parallel execution mechanisms

oschulz February 5, 2021, 7:47pm 6

Awesome stuff, @tkf !

1 Like

Topic		Replies	Views
Comparing Folds/FLoop Executors General Usage threads , floops , folds	0	350	December 14, 2022
FLoops ThreadedEx scheduling General Usage package	2	700	September 16, 2022
Overhead of `Threads.@threads` Performance question , multithreading	30	5354	March 13, 2021
[ANN] Folds.jl: threaded, distributed, and GPU-based high-level data-parallel interface for Julia Package Announcements package , announcement , parallel , multithreading	11	2058	February 7, 2021
More threads, slower code, even if not spawning them Performance	19	831	January 29, 2022