public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* perf: hsw model numbers
@ 2013-12-11 10:36 Peter Zijlstra
  2013-12-11 11:24 ` Borislav Petkov
  2013-12-11 18:53 ` Andi Kleen
  0 siblings, 2 replies; 4+ messages in thread
From: Peter Zijlstra @ 2013-12-11 10:36 UTC (permalink / raw)
  To: andi, Ingo Molnar, linux-kernel, H. Peter Anvin

Hi Andi,

arch/x86/kernel/cpu/perf_event_intel.c

3a632cb229bfb (Andi Kleen       2013-06-17 17:36:48 -0700 2500)         case 60: /* Haswell Client */
3a632cb229bfb (Andi Kleen       2013-06-17 17:36:48 -0700 2501)         case 70:
3a632cb229bfb (Andi Kleen       2013-06-17 17:36:48 -0700 2502)         case 71:
3a632cb229bfb (Andi Kleen       2013-06-17 17:36:48 -0700 2503)         case 63:
0499bd867bd17 (Andi Kleen       2013-08-08 17:26:23 -0700 2504)         case 69:

Can you add comments with each model number describing what actual part
it is?

Also, we seem to lack snb/ivb EX part numbers, are those already known?

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: perf: hsw model numbers
  2013-12-11 10:36 perf: hsw model numbers Peter Zijlstra
@ 2013-12-11 11:24 ` Borislav Petkov
  2013-12-11 18:53 ` Andi Kleen
  1 sibling, 0 replies; 4+ messages in thread
From: Borislav Petkov @ 2013-12-11 11:24 UTC (permalink / raw)
  To: Peter Zijlstra; +Cc: andi, Ingo Molnar, linux-kernel, H. Peter Anvin

On Wed, Dec 11, 2013 at 11:36:31AM +0100, Peter Zijlstra wrote:
> Hi Andi,
> 
> arch/x86/kernel/cpu/perf_event_intel.c
> 
> 3a632cb229bfb (Andi Kleen       2013-06-17 17:36:48 -0700 2500)         case 60: /* Haswell Client */
> 3a632cb229bfb (Andi Kleen       2013-06-17 17:36:48 -0700 2501)         case 70:
> 3a632cb229bfb (Andi Kleen       2013-06-17 17:36:48 -0700 2502)         case 71:
> 3a632cb229bfb (Andi Kleen       2013-06-17 17:36:48 -0700 2503)         case 63:
> 0499bd867bd17 (Andi Kleen       2013-08-08 17:26:23 -0700 2504)         case 69:
> 
> Can you add comments with each model number describing what actual part
> it is?
> 
> Also, we seem to lack snb/ivb EX part numbers, are those already known?

Yeah, it would be nice if Intel generally documented that whole model
mapping somewhere. For example, I spent some serious time grepping and
searching the net what family 6, model 47 actually is.

I asked hpa at some point about such doc and he said, CPUID doc needs to
get updated.

-- 
Regards/Gruss,
    Boris.

Sent from a fat crate under my desk. Formatting is fine.
--

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: perf: hsw model numbers
  2013-12-11 10:36 perf: hsw model numbers Peter Zijlstra
  2013-12-11 11:24 ` Borislav Petkov
@ 2013-12-11 18:53 ` Andi Kleen
  2013-12-11 20:18   ` Peter Zijlstra
  1 sibling, 1 reply; 4+ messages in thread
From: Andi Kleen @ 2013-12-11 18:53 UTC (permalink / raw)
  To: Peter Zijlstra; +Cc: andi, Ingo Molnar, linux-kernel, H. Peter Anvin

> Can you add comments with each model number describing what actual part
> it is?

It would probably be confusing, they have lots of different marketing
names :/ Just think of them as some flavour of Haswell.

> 
> Also, we seem to lack snb/ivb EX part numbers, are those already known?

They are the same as respective EP (and there is no SNB-EX)

-Andi

-- 
ak@linux.intel.com -- Speaking for myself only.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: perf: hsw model numbers
  2013-12-11 18:53 ` Andi Kleen
@ 2013-12-11 20:18   ` Peter Zijlstra
  0 siblings, 0 replies; 4+ messages in thread
From: Peter Zijlstra @ 2013-12-11 20:18 UTC (permalink / raw)
  To: Andi Kleen; +Cc: Ingo Molnar, linux-kernel, H. Peter Anvin

On Wed, Dec 11, 2013 at 07:53:10PM +0100, Andi Kleen wrote:
> > Can you add comments with each model number describing what actual part
> > it is?
> 
> It would probably be confusing, they have lots of different marketing
> names :/ Just think of them as some flavour of Haswell.

So they're all 'client' chips? Why do they have different model numbers
then?

> > 
> > Also, we seem to lack snb/ivb EX part numbers, are those already known?
> 
> They are the same as respective EP (and there is no SNB-EX)

Urgh, so we cannot differentiate between EP/EX anymore? Does that mean
that the EP chips can be used in 4-8 socket systems?


^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2013-12-11 20:18 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-12-11 10:36 perf: hsw model numbers Peter Zijlstra
2013-12-11 11:24 ` Borislav Petkov
2013-12-11 18:53 ` Andi Kleen
2013-12-11 20:18   ` Peter Zijlstra

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox