2.4.6-ac5 gives wrong cache info for Duron in /proc/cpuinfo

kernel 2.4.6-ac5
CPU AMD Duron 700

/proc/cpuinfo gives :
cache size: 64 KB

This is wrong :
- the Duron has 192 kilobytes of cache ( 64 L1 I, 64 L1 D , 64 L2 unified )
- what is KB ?
- "kilo" is abbreviated to 'k' , not 'K'
- "B" means "Bell" :-)

David Balazic
--------------
"Be excellent to each other." - Bill & Ted
- - - - - - - - - - - - - - - - - - - - - -

As far as I know older Durons have a bug. They report a wrong size for the cache.

Hmm, I think kilo is 1000 and K is 1024.

With an Athlon I get 256KB.
So I guess, that cache size shows only the 2nd level cache size.

I believe it is normal to write 'K' for 1024, 'k' for 1000 and 'B' for bytes and 'b' for bits.

Have a look at acronymfinder.com, they distinguish the capital and lowercase b for bits and bytes.

Ketil

It never ceases to amaze me how ANAL some people on this list are. :-)

----- Original Message -----
From: "Ketil Froyn" <[email protected]>
To: "David Balazic" <[email protected]>
Cc: <[email protected]>
Sent: Tuesday, July 17, 2001 8:35 AM
Subject: Re: 2.4.6-ac5 gives wrong cache info for Duron in /proc/cpuinfo

| On Tue, 17 Jul 2001, David Balazic wrote:
| 
| > /proc/cpuinfo gives :
| > cache size: 64 KB
| >
| > This is wrong :
| >  - the Duron has 192 kilobytes of cache ( 64 L1 I, 64 L1 D , 64 L2 unified )
| >  - what is KB ?
| >    - "kilo" is abbreviated to 'k' , not 'K'
| >    - "B" means "Bell" :-)
| 
| I believe it is normal to write 'K' for 1024, 'k' for 1000 and 'B' for
| bytes and 'b' for bits.
| 
| Have a look at acronymfinder.com, they distinguish the capital and
| lowercase b for bits and bytes.
| 
| Ketil

ý:.žË›±Êâmçë¢kaŠÉb²ßìzwm…éb�ïîžË›±Êâmébžìÿ‘êçz_âžØ^n‡r¡ö¦zË�ëh™¨è­Ú&£ûàz¿äz¹Þ—ú+€Ê+zf£¢·hšˆ§~†­†Ûiÿÿï�êÿ‘êçz_è®æj:+v‰¨þ)ߣømšSåy«­æ�¶…
­†ÛiÿÿðÃí»è®å’i

Ouch, you got me there!! ;-)

Ketil

-----Original Message-----
Date: Tue, 17 Jul 2001 06:51:02 -0500
From: William Scott Lockwood III <[email protected]>
Reply-To: William Scott Lockwood III <[email protected]>
To: [email protected]
Subject: Re: 2.4.6-ac5 gives wrong cache info for Duron in /proc/cpuinfo

It never ceases to amaze me how ANAL some people on this list are. :-)

----- Original Message -----
From: "Ketil Froyn" <[email protected]>
To: "David Balazic" <[email protected]>
Cc: <[email protected]>
Sent: Tuesday, July 17, 2001 8:35 AM
Subject: Re: 2.4.6-ac5 gives wrong cache info for Duron in /proc/cpuinfo

| On Tue, 17 Jul 2001, David Balazic wrote:
|
| > /proc/cpuinfo gives :
| > cache size: 64 KB
| >
| > This is wrong :
| >  - the Duron has 192 kilobytes of cache ( 64 L1 I, 64 L1 D , 64 L2 unified )
| >  - what is KB ?
| >    - "kilo" is abbreviated to 'k' , not 'K'
| >    - "B" means "Bell" :-)
|
| I believe it is normal to write 'K' for 1024, 'k' for 1000 and 'B' for
| bytes and 'b' for bits.
|
| Have a look at acronymfinder.com, they distinguish the capital and
| lowercase b for bits and bytes.
|
| Ketil

N‹§²æìr¸›yúèšØb²X¬·ë�ÛazX§zéì¹»®&Þ–)îÆG«�ébžØ^n‡r¡ö¦zË�ëh™¨è­Ú&¢ø®G«�éh®(­éšŽŠÝ¢j"�ú¶m§ïÿ¾«‘êçzZ+ƒùšŽŠÝ¢j"�ú!¶iO•æ¬z·švØ^¶m§ïÿÃ-»+ƒùdš_

It's called attention to detail, and it's the _reason_ why a lot of people are here.

The correct prefix to signify a multiple of 1024 is 'Ki'.

Index: arch/i386/kernel/setup.c
===================================================================
RCS file: /inst/cvs/linux/arch/i386/kernel/setup.c,v retrieving revision 1.4.2.57

diff -u -r1.4.2.57 setup.c
@@ -2406,7 +2406,7 @@
 
 		/* Cache size */
 		if (c->x86_cache_size >= 0)
-			p += sprintf(p, "cache size\t: %d KB\n", c->x86_cache_size);
+			p += sprintf(p, "cache size\t: %d KiB\n", c->x86_cache_size);
 		
 		/* We use exception 16 if we have hardware math and we've either seen it or the CPU claims it is internal */
 		fpu_exception = c->hard_math && (ignore_irq13 || cpu_has_fpu);

--
dwmw2

Oh, I understand attention to detail. After spending eight years in the United States Navy as a Hospital Corpsman, you either understand attention to detail, or you are not a Corpsman any longer. :-) It just struck me as funny that, while an understanding of attention to detail is important, so is the way in which information is communicated. In fact, it's almost as important as the information itself. I've seen more fights start over how something was said, or not said - I'm sure you get the picture. :-)

Me, I mainly lurk here for that very reason - I've received too much hatemail from posting to the list, so I try to follow it instead and see what I can learn. Today, I learned that multiples of 1024 is actually Ki. Thank you! I did not know that.

Scott
----- Original Message -----
From: "David Woodhouse" <[email protected]>
To: "William Scott Lockwood III" <[email protected]>
Cc: <[email protected]>
Sent: Tuesday, July 17, 2001 10:18 AM
Subject: Re: 2.4.6-ac5 gives wrong cache info for Duron in /proc/cpuinfo

| 
| [email protected] said:
| > It never ceases to amaze me how ANAL some people on this list are.
| > :-)
| 
| It's called attention to detail, and it's the _reason_ why a lot of people
| are here.
| 
| The correct prefix to signify a multiple of 1024 is 'Ki'.
| 
| Index: arch/i386/kernel/setup.c
| ===================================================================
| RCS file: /inst/cvs/linux/arch/i386/kernel/setup.c,v
| retrieving revision 1.4.2.57
| diff -u -r1.4.2.57 setup.c
| --- arch/i386/kernel/setup.c 2001/05/14 10:32:23 1.4.2.57
| +++ arch/i386/kernel/setup.c 2001/07/17 15:13:54
| @@ -2406,7 +2406,7 @@
| 
|   /* Cache size */
|   if (c->x86_cache_size >= 0)
| - p += sprintf(p, "cache size\t: %d KB\n", c->x86_cache_size);
| + p += sprintf(p, "cache size\t: %d KiB\n", c->x86_cache_size);
| 
|   /* We use exception 16 if we have hardware math and we've either seen it or the CPU claims it is internal */
|   fpu_exception = c->hard_math && (ignore_irq13 || cpu_has_fpu);
| 
| 
| --
| dwmw2
| 
| 
| -
| To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
| the body of a message to [email protected]
| More majordomo info at  http://vger.kernel.org/majordomo-info.html
| Please read the FAQ at  http://www.tux.org/lkml/
| 

ý:.žË›±Êâmçë¢kaŠÉb²ßìzwm…éb�ïîžË›±Êâmébžìÿ‘êçz_âžØ^n‡r¡ö¦zË�ëh™¨è­Ú&£ûàz¿äz¹Þ—ú+€Ê+zf£¢·hšˆ§~†­†Ûiÿÿï�êÿ‘êçz_è®æj:+v‰¨þ)ߣømšSåy«­æ�¶…
­†ÛiÿÿðÃí»è®å’i

It's a reasonably new standard that hasn't caught on, because many people think that "kibibyte" is stupid.

Alan Shutko <[email protected]> - In a variety of flavors!
Tempt me with a spoon!

Kibibbles N Bits! N Bits!! N Bits!!!

Sorry, I had to say it ;)

http://www.koin.com/athome/wisebuys/athome-wisebuys-980223-184529.html

Ignacio Vazquez-Abrams  <[email protected]>

Well, it does sound stupid but hey - they have to call it something right? I have always been a bit peeved that I get a 30 gig drive, take it home and plug it in only to find that it is actually much less than 30 gig. :-)

----- Original Message -----
From: "Alan Shutko" <[email protected]>
To: "William Scott Lockwood III" <[email protected]>
Cc: "David Woodhouse" <[email protected]>; <[email protected]>
Sent: Tuesday, July 17, 2001 10:50 AM
Subject: Re: [VERY OT] Re: 2.4.6-ac5 gives wrong cache info for Duron in /proc/cpuinfo

| "William Scott Lockwood III" <[email protected]> writes:
| 
| > Today, I learned that multiples of 1024 is actually Ki.
| 
| It's a reasonably new standard that hasn't caught on, because many
| people think that "kibibyte" is stupid.
| 
| --
| Alan Shutko <[email protected]> - In a variety of flavors!
| Tempt me with a spoon!
| 

ý:.žË›±Êâmçë¢kaŠÉb²ßìzwm…éb�ïîžË›±Êâmébžìÿ‘êçz_âžØ^n‡r¡ö¦zË�ëh™¨è­Ú&£ûàz¿äz¹Þ—ú+€Ê+zf£¢·hšˆ§~†­†Ûiÿÿï�êÿ‘êçz_è®æj:+v‰¨þ)ߣømšSåy«­æ�¶…
­†ÛiÿÿðÃí»è®å’i

Well then it should say "L2 cache size:"
It is a bug in any case.

David Balazic
--------------
"Be excellent to each other." - Bill & Ted
- - - - - - - - - - - - - - - - - - - - - -

The kernel messages at boot have no trouble finding out the correct cache info.

David Balazic
--------------
"Be excellent to each other." - Bill & Ted
- - - - - - - - - - - - - - - - - - - - - -