[GHC] #13164: idle time full GCs (idle cpu usage)
GHC
ghc-devs at haskell.org
Thu Jan 26 18:04:46 UTC 2017
#13164: idle time full GCs (idle cpu usage)
-------------------------------------+-------------------------------------
Reporter: lspitzner | Owner:
Type: feature request | Status: new
Priority: normal | Milestone:
Component: Runtime System | Version: 8.0.1
Resolution: | Keywords: idle GC
Operating System: Unknown/Multiple | Architecture:
| Unknown/Multiple
Type of failure: None/Unknown | Test Case:
Blocked By: | Blocking:
Related Tickets: | Differential Rev(s):
Wiki Page: |
-------------------------------------+-------------------------------------
Comment (by rwbarton):
FWIW I've never cared about deadlock detection, and it sounds unreliable
in view of lspitzner's most recent comment. I do care that interactive
programs I'm not actively using don't consume 12% of my CPU. Doing a GC
when the user isn't interacting with the program sounds like a good idea
to hide latency.
Judging from the detailed GC statistics lspitzner's program isn't totally
idle; it must be doing a little bit of work every second, repeatedly
triggering a new idle GC. But one would expect it to be cheap to do a
little bit of work every second. The heap could be a lot bigger than 50
MB...
The user's guide suggests enabling the idle GC for interactive programs,
but when there is a little work to do every second it seems basically
unusable. And the effects of the idle GC would be lost completely by doing
work every 0.2 seconds, which is strange.
--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13164#comment:5>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
More information about the ghc-tickets
mailing list