Hi Kevin,

Here's a question, one that has stumped the PAPI team for quite a long
time. Have you been able to get reasonable counts for L1 Dcache misses?
This is a replay event, but even with the IA32 docs we've never been
able to get numbers that 'make sense' out of it. L2 is no problem...

Phil


On Thu, 2006-08-24 at 08:44 -0700, Stephane Eranian wrote:
> Kevin,
> 
> On Thu, Aug 24, 2006 at 10:10:17AM -0500, Kevin Corry wrote:
> > On Mon August 21 2006 10:42 am, Stephane Eranian wrote:
> > > I have also released a new libpfm, libpfm-3.2-060821, which includes:
> > >
> > >   - preliminary support for P4 32 and 64 bit mode by Kevin Corry from IBM
> > >
> > > Also a new version of pfmon, pfmon-3.2-060821, to take advantage of the
> > > update in libpfm:
> > >
> > >   - preliminary support for P4 32 and 64 bits by Kevin Corry from IBM.
> > >     Do not get too excited and it does not count anything just yet!
> > 
> > I've definitely been able to count things on P4 with pfmon.
> 
> That's excellent. No offense for the comment, I think I was still under
> the impression it was very preliminary (from your own comments)
> and that it was not counting yet.
> 
> Obviously, I was wrong, my apologies.
> 
> 
> Now, I ran some tests just now and it seems to work for me as well.
> Here is what I could get with my stupid noploop program and by
> increasing the number of iterations:
> 
> $ pfmon --us-c -einstr_retired:NBOGUSNTAG ~/noploop 100000
> loop=100000
> 609,960 instr_retired:NBOGUSNTAG
> 
> $ pfmon --us-c -einstr_retired:NBOGUSNTAG ~/noploop 1000000
> loop=1000000
> 6,009,996 instr_retired:NBOGUSNTAG
> 
> $ pfmon --us-c -einstr_retired:NBOGUSNTAG ~/noploop 10000000
> loop=10000000
> 60,010,027 instr_retired:NBOGUSNTAG
> 
> But when I use instr_completed, it gets more flaky. But maybe I do
> not undertstand what the event actually measures:
> 
> $ pfmon -iinstr_completed
> Name     : instr_completed
> Code     : 0x7
> Counters : [ 6 7 8 15 16 17 ]
> Desc     : Instructions that have completed and retired during a clock cycle
> Umask    : 0x01 : [NBOGUS] : Non-bogus instructions.
> Umask    : 0x02 : [BOGUS] : Bogus instructions.
> 
> Does anyone get something meaningful out of this one?
> 
> 
> > [EMAIL PROTECTED] /home/corry]$ pfmon -i global_power_events
> > Name   : global_power_events
> > Code   : 0x13
> > counter: [ 0 1 9 10 ]
> > Unit-mask 0: RUNNING
> > Desc   : Counts the time during which a processor is not stopped.
> > 
> 
> In the rewritten AMD K8 event table, I have use the rule that if an event only
> has one unit mask, then we implement it without a unit mask, i.e., it is
> collapsed with the event name. It makes it much easier to use. On AMD K8,
> I have encoded the umask value starting at bit position 8. On P4, you
> probably need to use another trick, though.
> 

_______________________________________________
perfmon mailing list
[email protected]
http://www.hpl.hp.com/hosted/linux/mail-archives/perfmon/

Reply via email to