On 16/07/2013 4:06 AM, David Jeske wrote:
> So to re-iterate, all x86 GCs I'm aware of except Azul-Zing's "no 
> pause" have a stop-the-world pause during initial-mark of tenured 
> generation which is proportional to heapsize.

This isn't true of on-the-fly collectors. The pause is proportional to 
the root set size, which will incur a latency of hundreds of 
microseconds. If you compile your program into CPS form, then it's tens 
of microseconds. This has other downsides of course, but the technology 
for near-zero latency exists.

The CHICKEN, CLOVER and STOPLESS collectors I previously linked also 
have pauses on the order of hundreds of microseconds, which is an 
acceptable latency for just about any application I can think of.

Sandro


_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to