CPUID name corrupted - potential instability

Discussion of all forms of processors, from AMD to Intel to VIA.

Moderators: Flying Fox, morphine

CPUID name corrupted - potential instability

Postposted on Sat Sep 07, 2013 8:21 pm

So I noticed while using AIDA16 that the CPUID CPU Name "AMD Athlon(tm) 64 X2 Dual Core Processor 4200+" must be a hardcoded ASCII string stored in the processor itself - which, according to Wikipedia's article on the CPUID instruction, is true. If so, then if you know you have a Phenom II X2, and it returns something slightly different, or off, like Phenom II B45 when you overclock/unlock it, that probably means that the registers EAX, EBX, ECX and EDX are systematically corrupted, and therefore, that the CPU is unstable... makes sense?

On the other hand, it doesn't really make sense since it seems people have unlocked CPUs, the string changes, but it's still 24/7 stable.
Mothership: Thuban 1055T@3.7GHz, 12GB DDR3, M5A99X EVO, GTX470+Icy Vision Rev.2@840/3800, Vertex 2E 60GB
Supply ship: Sargas@2.8GHz, 12GB DDR3, M4A88TD-V EVO/USB3
Corsair: Macbook Air Ivy Bridge
Crayon Shin Chan
Minister of Gerbil Affairs
 
Posts: 2237
Joined: Fri Sep 06, 2002 11:14 am
Location: Malaysia

Re: CPUID name corrupted - potential instability

Postposted on Sat Sep 07, 2013 8:55 pm

This is a wild ass guess: warnings and caveats are all in place...
Perhaps the string it returns depends upon the configuration state? It wouldn't be hard to allocate space for more than one hard coded string.
mnecaise
Gerbil First Class
 
Posts: 114
Joined: Thu Oct 30, 2008 1:17 pm
Location: Hampton Roads, Virginia

Re: CPUID name corrupted - potential instability

Postposted on Sun Sep 08, 2013 8:46 am

mnecaise wrote:This is a wild ass guess: warnings and caveats are all in place...
Perhaps the string it returns depends upon the configuration state? It wouldn't be hard to allocate space for more than one hard coded string.

That would be the most sensible way to do it. Otherwise they would need to somehow hard-code different strings into each processor after binning, which would be nasty from a manufacturing standpoint. The upshot of this is, if it sees a configuration which does not match any known version of the processor, then you may see something odd.

@OP - If the registers were being corrupted that easily, the processor would not work at all.
(this space intentionally left blank)
just brew it!
Administrator
Gold subscriber
 
 
Posts: 37632
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: CPUID name corrupted - potential instability

Postposted on Sun Sep 08, 2013 1:43 pm

I would WAG that your motherboard might be doing some unlocking black magic involving switching out the microcode in use.

Alternately, your motherboard might be the one in charge of building the CPUID string, so there's a few bits looked up against a table to get the full model name.

I sincerely doubt it makes any meaningful change.
Siglessness is boring.
Image - M4800-Eight1
Image - Vargr-Z97
Forge
Lord High Gerbil
 
Posts: 8010
Joined: Wed Dec 26, 2001 7:00 pm
Location: SouthEast PA

Re: CPUID name corrupted - potential instability

Postposted on Sun Sep 08, 2013 2:15 pm

just brew it! wrote:
mnecaise wrote:This is a wild ass guess: warnings and caveats are all in place...
Perhaps the string it returns depends upon the configuration state? It wouldn't be hard to allocate space for more than one hard coded string.

That would be the most sensible way to do it. Otherwise they would need to somehow hard-code different strings into each processor after binning, which would be nasty from a manufacturing standpoint. The upshot of this is, if it sees a configuration which does not match any known version of the processor, then you may see something odd.


That's what I was thinking. My understanding, which may be dated, is that they set a couple bits (fuses) after testing and binning to configure the cores to the default, "as shipped" configuration. Early on, AMD was doing this with jumpers on the ceramic package (unlocking a processor with a pencil, I remember doing that).
mnecaise
Gerbil First Class
 
Posts: 114
Joined: Thu Oct 30, 2008 1:17 pm
Location: Hampton Roads, Virginia

Re: CPUID name corrupted - potential instability

Postposted on Sun Sep 08, 2013 2:57 pm

seems you might have some setting messed up in your bios. check this link from anandtech about unlocking cores.http://www.anandtech.com/show/2927
AmazighQ
Gerbil In Training
 
Posts: 2
Joined: Sun Aug 18, 2013 3:08 am

Re: CPUID name corrupted - potential instability

Postposted on Wed Sep 11, 2013 5:44 pm

IIRC these strings are stored in the BIOS, not on the chip itself. The BIOS only reads the capabilities and configuration of the CPU and then picks the name from the list of known CPUs. In the case that it doesn't match, you get funky things like "Phenom II B45" and such.
Z68XP-UD4 | 2700K @ 4.4 GHz | 16 GB | 770 | PCP&C Silencer 950 | XSPC RX360 | Heatkiller R3 | D5 + RP-452X2 | HAF 932 | 1 TB WD Black w/ SRT
Waco
Gerbil Elite
 
Posts: 744
Joined: Tue Jan 20, 2009 4:14 pm

Re: CPUID name corrupted - potential instability

Postposted on Wed Sep 11, 2013 9:40 pm

Waco wrote:IIRC these strings are stored in the BIOS, not on the chip itself. The BIOS only reads the capabilities and configuration of the CPU and then picks the name from the list of known CPUs. In the case that it doesn't match, you get funky things like "Phenom II B45" and such.

According to the description of the CPUID instruction (mentioned but not linked in the OP), it is reported directly by the CPU.
(this space intentionally left blank)
just brew it!
Administrator
Gold subscriber
 
 
Posts: 37632
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: CPUID name corrupted - potential instability

Postposted on Thu Sep 12, 2013 8:46 am

just brew it! wrote:According to the description of the CPUID instruction (mentioned but not linked in the OP), it is reported directly by the CPU.

The features, model number, brand, etc are all returned. There are no strings identifying the actual model though - that's what the combination of the various fields gets you.

Unless I'm totally misreading the spec (and you can fit strings into a few 8-bit fields) I'm pretty sure that the BIOS/UEFI is responsible for decoding what CPU you have and coming up with that string.
Z68XP-UD4 | 2700K @ 4.4 GHz | 16 GB | 770 | PCP&C Silencer 950 | XSPC RX360 | Heatkiller R3 | D5 + RP-452X2 | HAF 932 | 1 TB WD Black w/ SRT
Waco
Gerbil Elite
 
Posts: 744
Joined: Tue Jan 20, 2009 4:14 pm

Re: CPUID name corrupted - potential instability

Postposted on Thu Sep 12, 2013 9:21 am

my 2c.. some software also takes the speed into account. if your model and revision and speed in mhz match another cpu it may be calling it that. if it passes a stress test (prime+3dmark looped) i would not worry about it.
Cybert said: Capitlization and periods are hard for you, aren't they? I've given over $100 to techforums. I should have you banned for my money.
maxxcool
Gerbil Elite
Silver subscriber
 
 
Posts: 644
Joined: Thu Sep 12, 2002 8:40 am
Location: %^&*%$$


Return to Processors

Who is online

Users browsing this forum: No registered users and 3 guests