From: bobbypowers@... Date: 2020-02-29T18:27:27+00:00 Subject: [ruby-core:97314] [Ruby master Feature#16648] improve GC performance by 5% with builtin_prefetch Issue #16648 has been updated by bpowers (Bobby Powers). alanwu (Alan Wu) wrote in #note-1: > I ran the patch on some included GC benchmarks in the repo and it doesn't seem to be a pure win (built-ruby is the patched version): Thanks! I hadn't seen these. I see roughly similar results locally on these benchmarks; I'll dig in to see if I can understand whats happening. ---------------------------------------- Feature #16648: improve GC performance by 5% with builtin_prefetch https://bugs.ruby-lang.org/issues/16648#change-84439 * Author: bpowers (Bobby Powers) * Status: Open * Priority: Normal ---------------------------------------- The mark phase of non-incremental major GC is (I believe) dominated by pointer chasing. One way we can improve that is by prefetching cachelines from memory before they are accessed, to reduce stalls. I did some experimenting, and the following patch reduces the time spent on a full GC from ~ 950 milliseconds to ~ 900 milliseconds, a small but stable improvement. I would love if additional folks have other benchmarks (or could point me at them) to see if this holds up beyond the web service I tested, and whether something like this could be considered for inclusion. I also attempted a more "principled" approach based on an optimization described in the GC handbook: putting a FIFO queue in front of the mark stack, and prefetching addresses as they enter the queue. However, I wasn't able to see any performance improvement there despite testing a number of queue sizes from 4 to 64. Its possible I implemented this wrong, or misjudged the access patterns (if e.g. the memory of a VALUE is accessed before push_mark_stack is called, it would invalidate this approach). The code for that alternative is here: https://github.com/bpowers/ruby/commit/d790d0c15047c36c23850a112093fe0e32fd3262 ---Files-------------------------------- 0001-gc-prefech-objects-seems-to-improve-full-GC-performa.patch (2.29 KB) -- https://bugs.ruby-lang.org/ Unsubscribe: