Re: The sixty second pmc howto
- From: Andrew Gallatin <gallatin@xxxxxxxxxxx>
- Date: Thu, 23 Feb 2006 10:19:51 -0500 (EST)
Robert Watson writes:
(2) Run "pmcstat -S instructions -O /tmp/sample.out" to start sampling of
instruction retirement events, saving the results to /tmp/sample.out.
Dumb question, but what does "instructions" really mean? The number
of instructions, the time spent executing them, ?
<.....>
Since there is no call graph information in the sample, the first few pages of
gprof output will be of limited utility, but the summary table by function is
the bit I found most useful:
You can use gprof -l to suppress the printing of the call-graph profile
My only problem with hwpmc is that it will not work for kernel
modules. I wonder if somebody with enough toolchain fu could take
kldstat output and produce a pre-linked elf executable image
containing kernel+modules which could be used for hwpmc, and also for
crash dump analysis. Or is there another, better, way to get a
complete symbol table of the kernel & all kernel modules?
So if you're doing kernel performance work, and not already using pmc, you
probably should be.
Very much agreed.
Drew
_______________________________________________
freebsd-current@xxxxxxxxxxx mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscribe@xxxxxxxxxxx"
- Follow-Ups:
- Re: The sixty second pmc howto
- From: Robert Watson
- Re: The sixty second pmc howto
- References:
- The sixty second pmc howto
- From: Robert Watson
- The sixty second pmc howto
- Prev by Date: The sixty second pmc howto
- Next by Date: Re: The sixty second pmc howto
- Previous by thread: The sixty second pmc howto
- Next by thread: Re: The sixty second pmc howto
- Index(es):
Relevant Pages
|
|