Hi,
I realize that this is an old thread, but it seems quite relevant to
me now... My apologies if I shouldn't be replying on old threads...
Citeren Stephane Eranian <[EMAIL PROTECTED]>:
Kevin,
On Thu, Aug 24, 2006 at 01:23:47PM -0500, Kevin Corry wrote:
On Thu August 24 2006 10:44 am, Stephane Eranian wrote:
> On Thu, Aug 24, 2006 at 10:10:17AM -0500, Kevin Corry wrote:
> > 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.
Not a problem. You're right that I did mention that the support is not
complete - there are features of the P4 PMU that I still want to add support
for: event filtering, event tagging, and event cascading. But the
code that's
there now is useable for basic event counting.
Ok, so you are saying that basic counting (and sampling) should work
fine then.
That is already quite an accomlishment given the complexity of the P4.
As some of you may have noticed in a thread on the PAPI mailinglist,
counts on P4 don't seem right.
I've counted instructions for SPEC CPU2000 using PIN (a dynamic
instrumentator) and compared them with papiex (latest version,
together with latest perfctr patch) and an Fedora Core 4 Linux / Intel
Pentium 4 machine (see attachment).?
For some benchmarks the counts are quite similar, but for others, they
are way of. Mesa is a nice example of this: papiex counts almost
7*10^9 instructions more, or 2.5% more than PIN does ! The PIN counts
have been validated using another instrumentator (DIOTA), and counts
using perfex on an AMD Athlon machine. Even for mesa, counts didn't
differ more than 300,000 instructions.
So my conclusion is that something must be going horribly wrong with
the HPCs on the Pentium4. Maybe the instr_completed event (see below)
is actually what we want, but trying that also yields a zero count
with me, as it did below. Has anyone figured out yet why
instr_completed isn't working? The documentation mentions it will
only work on model 3 or model 4 Pentium 4s, but looking at my
/proc/cpuinfo tells me my machine is a model 3, so it _should_ work...
Any ideas or comments on this are highly appreciated.
greetings,
> $ 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?
I've noticed this as well, actually.
The IA32 Developers Manual, Appendix A lists all the events and
related info.
It has this to say about instr_retired and instr_completed:
instr_retired: This event counts instructions that are retired
during a clock
cycle. Mask bits specify bogus or non-bogus (and whether they are tagged
using the front-end tagging mechanism).
instr_completed: This event counts instructions that have completed and
retired during a clock cycle. Mask bits specify whether the instruction is
bogus or non-bogus. This metric differs from instr_retired, since it counts
instructions completed, rather than the number of times that instructions
started.
I am wondering if instr_completed does not count instruction that
execute and retired
in 1 cycle as opposed to instructions that take more. Yet, I doubt
we don't have any
of those in a program such as date for instance. We need to verify
that instr_comp;leted
does not use a counter (CTR,ESCR,CCR) that is not used by the events
that do seem to
return valid data. There maybe some bugs in the kernel in this case.
<snipped>
--
-Stephane
_______________________________________________
perfmon mailing list
[email protected]
http://www.hpl.hp.com/hosted/linux/mail-archives/perfmon/
----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.
_______________________________________________
perfmon mailing list
[email protected]
http://www.hpl.hp.com/hosted/linux/mail-archives/perfmon/
