#4322: High CPU usage during idle time due to GC
--------------------------------------+-------------------------------------
Reporter: hesselink | Owner:
Type: bug | Status: closed
Priority: normal | Milestone:
Component: Runtime System | Version: 6.12.3
Resolution: invalid | Keywords: idle GC
Testcase: | Blockedby:
Difficulty: | Os: MacOS X
Blocking: | Architecture: x86
Failure: Runtime performance bug |
--------------------------------------+-------------------------------------
Comment(by hesselink):
That sounds like it could be a solution.
Note that we don't see any difference between an idle GC time of 2s and
30s. Once we get above the 1s of the threadDelay, the CPU usage drops
sharply from ~10% to ~0.25% and stays stable.
It seems I still don't understand something. If the threadDelay causes a
idle GC, I should see a GC run every 1s, right? (the threadDelay is 1e6)
But I just created a log with -vt, and, assuming times are in microseconds
as well (which they seem to be), I see a GC run every 1000-2000
microseconds, which is 1000x more than I'd expect.
I've attached the log again.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/4322#comment:8>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
