[commit: ghc] master: Add a comment about oddity with yieldThread() and timing results on Linux (40dd50a)

[Simon Marlow]

Repository : ssh://darcs.haskell.org//srv/darcs/ghc

On branch  : master

http://hackage.haskell.org/trac/ghc/changeset/40dd50a9d1d92a7bd8526b54ca5fcdd1eb95d42f

>---------------------------------------------------------------

commit 40dd50a9d1d92a7bd8526b54ca5fcdd1eb95d42f
Author: Simon Marlow <marlowsd at gmail.com>
Date:   Mon Dec 12 16:28:29 2011 +0000

    Add a comment about oddity with yieldThread() and timing results on Linux

>---------------------------------------------------------------

 rts/sm/GC.c |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)

diff --git a/rts/sm/GC.c b/rts/sm/GC.c
index 928f444..993de51 100644
--- a/rts/sm/GC.c
+++ b/rts/sm/GC.c
@@ -1060,6 +1060,11 @@ gcWorkerThread (Capability *cap)
 
     // Wait until we're told to wake up
     RELEASE_SPIN_LOCK(&gct->mut_spin);
+    // yieldThread();
+    //    Strangely, adding a yieldThread() here makes the CPU time
+    //    measurements more accurate on Linux, perhaps because it syncs
+    //    the CPU time across the multiple cores.  Without this, CPU time
+    //    is heavily skewed towards GC rather than MUT.
     gct->wakeup = GC_THREAD_STANDING_BY;
     debugTrace(DEBUG_gc, "GC thread %d standing by...", gct->thread_index);
     ACQUIRE_SPIN_LOCK(&gct->gc_spin);