Montecito processor family
Could someone, please, tell me what the family of the Montecito processor will be in /proc/cpuinfo? (text representation, numerical ID)
Thanks,
Zoltan
The family number is 32. Currently there is no text representation yet. Probably a piece of text should be there later.
Thanks.
-Fenghua
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Menyhart, Zoltan
Sent: Monday, June 13, 2005 12:04 PM
To: [email protected]
Subject: Montecito processor family
Could someone, please, tell me what the family of the Montecito processor will be in /proc/cpuinfo? (text representation, numerical ID)
Thanks,
Zoltan
-
To unsubscribe from this list: send the line "unsubscribe linux-ia64" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Hi,
I see that nothing has happend to Montecito's family name in both Linus' and Tony's tree since then.
Is adding 'Itanium 2' for case 0x20 enough or do we need something different?
Intel people may want to decide what it should be, according to their marketing/branding plan ;)
From: "Yu, Fenghua" <[email protected]>
Subject: RE: Montecito processor family
Date: Mon, 13 Jun 2005 12:17:34 -0700
Quoted 14 lines
The family number is 32. Currently there is no text representation yet. Probably a piece of text should be there later.
Thanks.
-Fenghua
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Menyhart, Zoltan
Sent: Monday, June 13, 2005 12:04 PM
To: [email protected]
Subject: Montecito processor family
Quoted 7 lines
Could someone, please, tell me what the family of the Montecito processor will be in
/proc/cpuinfo? (text representation, numerical ID)Thanks,
Zoltan
---
Takayoshi Kochi
Quoted 8 lines · Takayoshi Kochi
I see that nothing has happend to Montecito's family name in both Linus' and Tony's tree since then.
Is adding 'Itanium 2' for case
0x20enough or do we need something different?Intel people may want to decide what it should be, according to their marketing/branding plan ;)
I'm sure they're having trouble getting that patch through Intel legal ;-)
Maybe we could put in a patch temporarily that calls it "Montecito", then Intel can patch it to whatever the official marketing name is upon release?
I'm not quite sure why we're so allergic to using the codenames in the ia64 port. I mean, look at arch/i386/kernel/cpu/intel.c
Hi,
Quoted 6 lines
Quoted 6 lines
I see that nothing has happend to Montecito's family name in both Linus' and Tony's tree since then.
Is adding 'Itanium 2' for case
0x20enough or do we need something different?
Quoted 8 lines
Quoted 2 lines
Intel people may want to decide what it should be, according to their marketing/branding plan ;)
I'm sure they're having trouble getting that patch through Intel legal ;-)
Maybe we could put in a patch temporarily that calls it "Montecito", then Intel can patch it to whatever the official marketing name is upon release?
IIRC the field once was "McKinley" for Itanium2 but later replaced with "Itanium 2". So this sounds ok. But probably as no real application depends on or relies on the field, I think we can display '32' as it is until the 'official' name is given.
Quoted 2 lines
I'm not quite sure why we're so allergic to using the codenames in the ia64 port. I mean, look at
arch/i386/kernel/cpu/intel.c
Maybe we can add 'codename' field in cpuinfo so that people can recognize codename:marketing name correspondense, and we can distinguish McKinley, Madison, Deerfield, ... (I don't remember all) among Itanium2s ;)
---
Takayoshi Kochi
Digging up this from the distant past (November, 2005):
Quoted 5 lines
Quoted 5 lines
I see that nothing has happend to Montecito's family name in both Linus' and Tony's tree since then.
Is adding 'Itanium 2' for case
0x20enough or do we need something different?
Printing "Itanium 2" hides information since that is the same string as we use for family 0x1f. So this would make it hard/impossible for a script reading /proc/cpuinfo to tell whether the cpu was McKinley/Madison or Montecito.
Quoted 2 lines
Quoted 2 lines
Intel people may want to decide what it should be, according to their marketing/branding plan ;)
It seems that the branding people focus most of their time and energy on processor implementations, rather than on families of processors. So there isn't a clear answer from them on just what they'd like 0x20 to be tranlated to. One did suggest:
family : Dual-core Intel(r) Itanium(r) 2 processor 9000 series
But that rather seriously overflows the "char family[32]" buffer into which it would need to be copied, doesn't fit into the style of earlier family names as used by Linux and might be outdated if we are still using family 0x20 for processors with more than two cores.
Quoted 1 line
I'm sure they're having trouble getting that patch through Intel legal ;-)
Marketing is a much tougher sell than legal :-)
Quoted 3 lines
Maybe we could put in a patch temporarily that calls it "Montecito", then Intel can patch it to whatever the official marketing name is upon release?
"Montecito" would be a poor choice. Next processor in the series is "Montvale", and it will also have family 0x20.
Quoted 4 lines
IIRC the field once was "McKinley" for Itanium2 but later replaced with "Itanium 2". So this sounds ok. But probably as no real application depends on or relies on the field, I think we can display '32' as it is until the 'official' name is given.
32 is good ... it even makes us more like arch/i386/kernel/cpu/proc.c which makes no attempt to interpret the number, just prints it in decimal. As a bonus, no patch is needed, nor will any future changes be needed as new family numbers are assigned!
The only remaining question is whether we should drop the translation of 0x7 to "Itanium" and 0x1f to "Itanium 2", and just unconditionally print the family as a decimal number. This would make it all symetrical (but sadly requires strong evidence that there are no applications that depend on the old string values for the "family" field in /proc/cpuinfo).
-Tony
Quoted 7 lines · Tony
Digging up this from the distant past (November, 2005):
It seems that the branding people focus most of their time and energy on processor implementations, rather than on families of processors. So there isn't a clear answer from them on just what they'd like
0x20to be tranlated to. One did suggest:
Quoted 7 lines
family : Dual-core Intel(r) Itanium(r) 2 processor 9000 series
But that rather seriously overflows the "char family[32]" buffer into which it would need to be copied, doesn't fit into the style of earlier family names as used by Linux and might be outdated if we are still using family
0x20for processors with more than two cores.
And is a bit crass ... I mean we already plug it as a GenuineIntel in the vendor field. How about more plainly:
family : Multi-core Itanium 2
I suppose this risks confusion with HP's Hondo package, though. Maybe
family : 9000 series Itanium 2
would do better?
Quoted 6 lines
Quoted 3 lines
Maybe we could put in a patch temporarily that calls it "Montecito", then Intel can patch it to whatever the official marketing name is upon release?
"Montecito" would be a poor choice. Next processor in the series is "Montvale", and it will also have family
0x20.
I would hope that the family name would be decided before Montvale was released ;-)
Quoted 4 lines
32 is good ... it even makes us more like
arch/i386/kernel/cpu/proc.cwhich makes no attempt to interpret the number, just prints it in decimal. As a bonus, no patch is needed, nor will any future changes be needed as new family numbers are assigned!
Yes, I like this idea. Being More Like i386 is always a good idea.
Quoted 5 lines
The only remaining question is whether we should drop the translation of
0x7to "Itanium" and0x1fto "Itanium 2", and just unconditionally print the family as a decimal number. This would make it all symetrical (but sadly requires strong evidence that there are no applications that depend on the old string values for the "family" field in/proc/cpuinfo).
We've always felt free to change these strings before. Anyone relying on them for anything more than display-to-the-user purposes is on a losing streak.
Quoted 6 lines · Tony Luck
Digging up this from the distant past (November, 2005):
Quoted 3 lines
Quoted 3 lines
I see that nothing has happend to Montecito's family name in both Linus' and Tony's tree since then.
Quoted 7 lines
Quoted 2 lines
Quoted 2 lines
Is adding 'Itanium 2' for case
0x20enough or do we need something different?Printing "Itanium 2" hides information since that is the same string as we use for family
0x1f. So this would make it hard/impossible for a script reading/proc/cpuinfoto tell whether the cpu was McKinley/Madison or Montecito.
Why not have "model name" with the text name, like on ia32?
vendor_id : GenuineIntel
cpu family : 15
model : 4
model name : Intel(R) Pentium(R) 4 CPU 2.80GHz
stepping : 1
Quoted 9 lines
Quoted 2 lines
Quoted 2 lines
Intel people may want to decide what it should be, according to their marketing/branding plan ;)
It seems that the branding people focus most of their time and energy on processor implementations, rather than on families of processors. So there isn't a clear answer from them on just what they'd like
0x20to be tranlated to. One did suggest:family : Dual-core Intel(r) Itanium(r) 2 processor 9000 series
That's almost as bad as "Intel(R) Pentium(R) 4 CPU 2.80GHz".
I like the "Pentium III (Coppermine)" style, based on family and model. "Itanium 2 (Madison)" and "Itanium 2 (Montecito)" would be consistent with that style.
Quoted 6 lines
Quoted 3 lines
Maybe we could put in a patch temporarily that calls it "Montecito", then Intel can patch it to whatever the official marketing name is upon release?
"Montecito" would be a poor choice. Next processor in the series is "Montvale", and it will also have family
0x20.
That's why a combination of family and model is needed.
Quoted 9 lines
Quoted 4 lines
IIRC the field once was "McKinley" for Itanium2 but later replaced with "Itanium 2". So this sounds ok. But probably as no real application depends on or relies on the field, I think we can display '32' as it is until the 'official' name is given.
32 is good ... it even makes us more like
arch/i386/kernel/cpu/proc.cwhich makes no attempt to interpret the number, just prints it in decimal. As a bonus, no patch is needed, nor will any future changes be needed as new family numbers are assigned!
It would be nice to have a "model name" text description, for people that don't know the meaning of 32. There are people that need to use the right flavor of CPU that know the English name. It's easier to have the code print the text name than teach them the cryptic numbers.
Signature · 2 lines
Russ Anderson, OS RAS/Partitioning Project Lead SGI - Silicon Graphics Inc [email protected]
Quoted 5 lines · Russ Anderson
It would be nice to have a "model name" text description, for people that don't know the meaning of 32. There are people that need to use the right flavor of CPU that know the English name. It's easier to have the code print the text name than teach them the cryptic numbers.
I second this. It shouldn't be as hard as it currently is to
determine what processor is installed in the system. The common names are well enough advertised, even if unofficial. The family/model numbers don't mean much to most people. Thanks,
Alex
Signature · 1 line
Alex Williamson HP Open Source & Linux Org.
Quoted 7 lines
Why not have "model name" with the text name, like on ia32?
vendor_id: GenuineIntel
cpu family : 15
model : 4
model name : Intel(R) Pentium(R) 4 CPU 2.80GHz
stepping : 1
This sounds like an excellent plan. We can even use the new (in SDM2.2 p. 2:345) PAL_BRAND_INFO call to get the right (Intel marketting/legal approved) string. For older systems that don't implement PAL_BRAND_INFO we could have some fallback code that decodes family/model to Merced/McKinley/Madison.
-Tony
Quoted 3 lines · Matthew Wilcox
We've always felt free to change these strings before. Anyone relying on them for anything more than display-to-the-user purposes is on a losing streak.
Debian Installer uses it for determining the best kernel to install - is there a better way? Of course, old installers aren't expected to work with new kernels, so the change won't hurt us as long as someone notices the change.
Signature · 1 line
dann frazier | HP Open Source and Linux Organization
Quoted 10 lines · dann frazier
Quoted 3 lines · Matthew Wilcox
We've always felt free to change these strings before. Anyone relying on them for anything more than display-to-the-user purposes is on a losing streak.
Debian Installer uses it for determining the best kernel to install - is there a better way? Of course, old installers aren't expected to work with new kernels, so the change won't hurt us as long as someone notices the change.
That seems like a good reason for family and model to be numbers straight from the hardware. Applications can key off the numbers. Then the text "model name" string could change without breaking applications.
vendor_id : GenuineIntel
cpu family : 15
model : 4
model name : Intel(R) Pentium(R) 4 CPU 2.80GHz
stepping : 1
Signature · 2 lines
Russ Anderson, OS RAS/Partitioning Project Lead SGI - Silicon Graphics Inc [email protected]
Quoted 10 lines
That seems like a good reason for family and model to be numbers straight from the hardware. Applications can key off the numbers. Then the text "model name" string could change without breaking applications.
vendor_id: GenuineIntel
cpu family : 15
model : 4
model name : Intel(R) Pentium(R) 4 CPU 2.80GHz
stepping : 1
Using PAL_BRAND_INFO on Montecito I end up with this:
vendor : GenuineIntel
arch : IA-64
family : 32
model : 0
model name : Intel(r) Itanium(r) 2 Processor 1.6GHz with 24M L3 Cache for 533MHz Platforms
Which is stylistically pretty similar (though somewhat more verbose ... PAL_BRAND_INFO allows for up to 127 characters + NUL so it could get even chattier, but at least there is an upper bound).
-Tony
Quoted 1 line · Tony
model name : Intel(r) Itanium(r) 2 Processor 1.6GHz with 24M L3 Cache for 533MHz Platforms
Presumably this would be the same string that's used in the processor manuals? If so, that would be really helpful to have.
--david
Signature · 1 line
Mosberger Consulting LLC, http://www.mosberger-consulting.com/
Quoted 4 lines
Quoted 1 line
model name : Intel(r) Itanium(r) 2 Processor 1.6GHz with 24M L3 Cache for 533MHz Platforms
Presumably this would be the same string that's used in the processor manuals? If so, that would be really helpful to have.
The same person at Intel provides the string to the PAL team for the PAL_BRAND_INFO call as gets used in the manuals. So the goal is that they should be the same. But the reaction time of print-media and PAL releases is different, so when branding changes happen, the strings may get out of sync. That's apparently the case here: print manuals will include the 9xxx processor model number, but current PAL doesn't have that.
-Tony
Luck, Tony wrote on Thursday, June 01, 2006 9:33 AM
Quoted 8 lines
Quoted 7 lines
Why not have "model name" with the text name, like on ia32?
vendor_id: GenuineIntel
cpu family : 15
model : 4
model name : Intel(R) Pentium(R) 4 CPU 2.80GHz
stepping : 1
Quoted 5 lines
This sounds like an excellent plan. We can even use the new (in SDM2.2 p. 2:345)
PAL_BRAND_INFOcall to get the right (Intel marketting/legal approved) string. For older systems that don't implementPAL_BRAND_INFOwe could have some fallback code that decodes family/model to Merced/McKinley/Madison.
I would also like to take this opportunity to revamp feature string. It is way too long, and should be either two or three letter acronym like x86:
X86 feature string: "flags : fpu vme de pse tsc msr pae mce ... "
- features : branchlong, 16-byte atomic ops
+ features : lb ao
Quoted 9 lines · Kenneth W
I would also like to take this opportunity to revamp feature string. It is way too long, and should be either two or three letter acronym like x86:
X86 feature string: "flags : fpu vme de pse tsc msr pae mce ... "
- features : branchlong, 16-byte atomic ops
+ features : lb ao
On the down side, I find that I sometimes have to go look in the code
to figure out which flags I care about on x86. Are "lb" and "ao" defined somewhere?
Alex
Signature · 1 line
Alex Williamson HP Open Source & Linux Org.
Alex Williamson wrote on Tuesday, June 06, 2006 11:17 AM
Quoted 6 lines · Kenneth W
Quoted 5 lines
I would also like to take this opportunity to revamp feature string. It is way too long, and should be either two or three letter acronym like x86:
Quoted 8 lines
Quoted 4 lines
X86 feature string: "flags : fpu vme de pse tsc msr pae mce ... "
- features : branchlong, 16-byte atomic ops
+ features : lb aoOn the down side, I find that I sometimes have to go look in the
code to figure out which flags I care about on x86. Are "lb" and "ao" defined somewhere?
I thought the x86 acronym was taken straight out of SDM, refer to CPUID instruction detail there. We should do the same on ia64.
Section 3.1.11 out of SDM volume 1:
field bit description
lb 0 long branch (brl) instructions. sd 1 spontaneous deferral ao 2 16-byte atomic operations rv 63:3 Reserved.
It's only my wishful thinking. The acronym should be improved a bit to be more creative and meaningful in its current form. Oh well.
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Ken" = Chen, Kenneth W <[email protected]> writes:
Ken> I thought the x86 acronym was taken straight out of SDM, refer to Ken> CPUID instruction detail there. We should do the same on ia64.
Ken> Section 3.1.11 out of SDM volume 1:
Ken> field bit description lb 0 long branch (brl) instructions. sd 1 Ken> spontaneous deferral ao 2 16-byte atomic operations rv 63:3 Ken> Reserved.
Ken> It's only my wishful thinking. The acronym should be improved a Ken> bit to be more creative and meaningful in its current form. Oh Ken> well.
IMHO single words are much nicer to read than strange letter combos. If we take this path, could we at least make sure it gets documented in Documentation/ somewhere?
Thanks,
Jes