diff options
| author | Simon Marlow <marlowsd@gmail.com> | 2011-12-09 10:35:46 +0000 |
|---|---|---|
| committer | Simon Marlow <marlowsd@gmail.com> | 2011-12-13 08:58:57 +0000 |
| commit | a02eb298d3f6089e51a43307ffb37e3a8076c8fd (patch) | |
| tree | 6174e46a959a2449b54e470c8b627e3ba1a277bb /rts/Capability.h | |
| parent | 55d7931d7fb55865a592f91dfa65435ee2859158 (diff) | |
| download | haskell-a02eb298d3f6089e51a43307ffb37e3a8076c8fd.tar.gz | |
New flag +RTS -qi<n>, avoid waking up idle Capabilities to do parallel GC
This is an experimental tweak to the parallel GC that avoids waking up
a Capability to do parallel GC if we know that the capability has been
idle for a (tunable) number of GC cycles. The idea is that if you're
only using a few Capabilities, there's no point waking up the ones
that aren't busy.
e.g. +RTS -qi3
says "A Capability will participate in parallel GC if it was running
at all since the last 3 GC cycles."
Results are a bit hit and miss, and I don't completely understand why
yet. Hence, for now it is turned off by default, and also not
documented except in the +RTS -? output.
Diffstat (limited to 'rts/Capability.h')
| -rw-r--r-- | rts/Capability.h | 3 |
1 files changed, 3 insertions, 0 deletions
diff --git a/rts/Capability.h b/rts/Capability.h index a4655dd36d..f60adf9de4 100644 --- a/rts/Capability.h +++ b/rts/Capability.h @@ -46,6 +46,9 @@ struct Capability_ { // catching unsafe call-ins. rtsBool in_haskell; + // Has there been any activity on this Capability since the last GC? + nat idle; + // The run queue. The Task owning this Capability has exclusive // access to its run queue, so can wake up threads without // taking a lock, and the common path through the scheduler is |