LisPi · @lispi314
381 followers · 6380 posts · Server mastodon.top

@violet Regarding low-collection overhead, there are also strategies like 's storage pools and 's 's concurrent garbage collection (wiki.openjdk.org/display/zgc/M), where the variance is very predictable (strict guarantee: won't be more than 1ms iirc; I think it qualifies as a firm real-time garbage collector) .

Storage pools have been implemented in before, and there's no reason implementations cannot have realtime garbage collectors (iirc NASA has done it before).

#ada #java #zgc #commonlisp

Last updated 2 years ago

LisPi · @lispi314
344 followers · 5078 posts · Server mastodon.top

@programming_discussions

It also sounds like they didn't even try (wiki.openjdk.org/display/zgc/M) since they're mentioning GC pauses as a problem and apparently their best answer to that was restarting nodes rather than configuring them to use a better GC.

Heap settings are really just tuning the same GC, not changing it.

@natty

#zgc #java #jvm #garbageCollection

Last updated 2 years ago

Christian · @schrumpelnase
24 followers · 775 posts · Server sueden.social