Re:Java 12
by mehrdadn on Tuesday March 19, 2019 @07:24PM
(#19437230)
Well that's exactly the point too. GC's aren't magic wands, and people just need to be honest about the inherent limitations of whatever they build or use. You can't claim X is independent of Y when it very clearly cannot be, no matter how much you want that to be the case. That's all I'm saying.
[ Parent
| on HN ]
The length of individual pauses is independent of your heap size. It still depends on your root set size (which includes thread stacks). Other factors (CPU cycles consumed, throughput) depend on your heap size and allocation rates.
Those performance goals are only valid when given enough spare resources (heap > live set size, cpu cores). If there are insufficient resources then the VM may eventually enter those "failure modes" to keep the application running, albeit at reduced performance, instead of crashing. I think there are flags to make it just exit if that's preferable over latency spikes.
Ultimately you cannot ask it to do the impossible, this much should be obvious.