Organization: University of California San Francisco--Biostatistics
On Wed, 2006-04-12 at 14:22 -0700, Ian Ollmann wrote:
> This wont fix gprof, but have you tried shark?
>
> Ian
I've had bad luck with both gprof and shark, as I've reported on this
list previously without eliciting much of a response.
My program was spending a lot of time in _szone_malloc and related
calls. In a nutshell, gprof did not indicate the calling sequence (it
marked these as <spontaneous>) and shark reported the wrong calling
sequence (see my post of March 14).
I've managed to eliminate that problem (among other things,
demonstrating what the calling sequence really was), but the performance
tools have not been as helpful as they could be, and their behavior does
not inspire much confidence.
Shark 4.3.2(1). OS X Server 10.3.9. dual G5's, gcc 3.3.
The code is C++ and compiled from make, not XCode.
Compiler options include -pg (which I know is not recommended for shark)
and -O2.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
PerfOptimization-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/perfoptimization-dev/email@hidden
This email sent to email@hidden
