#5282: Bizarre results from -P profiler on OS X
---------------------------------+------------------------------------------
Reporter: bos | Owner:
Type: bug | Status: new
Priority: high | Milestone: 7.2.1
Component: Runtime System | Version: 7.0.3
Keywords: | Testcase:
Blockedby: | Difficulty:
Os: Unknown/Multiple | Blocking:
Architecture: x86_64 (amd64) | Failure: Incorrect result at runtime
---------------------------------+------------------------------------------
Comment(by simonmar):
So the single-threaded RTS uses a CPU timer (that is, the signals are
generated every 20ms of CPU time), whereas the multi-threaded RTS uses a
wall-clock timer. This is explained in a comment at the top of
source:rts/posix/Itimer.c (the comment says "Major bogosity", but I now
think that's overstating it a bit, it's been this way for ever).
So perhaps on Mac OS X the CPU time signals are occurring at predictable
places, whereas wall-clock signals are more accurate. It would be
unfortunate if we had to switch to using the wall-clock timer, because
it's less accurate for profiling (periods when the process is descheduled
introduce noise into the results).
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/5282#comment:15>
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
