3.2 Prescott with 14x multiplier

[flaamer]

This weekend I picked up a couple of mobo/cpu combos from fry's, retail box 3.2P4's and cheap ECS mobo for $249. One turned out to be a prescott that I installed in an IC7G. Booted it up and found the multiplier fixed at 14x instead of 16x. I thought it was a mobo glitch so I took it out and installed it in a P4C800E-dlx... same 14x multiplier. I'm not complaining as this clocks pretty well 3.808 @1.4875vcore(bios setting) but I need to use 5:4 divider at 272 fsb. Just thinking if it were a 16x multiplier I could run 238fsb with my bh5 at 1:1. Anyone else seen a processor/multiplier mismatch?

P4 3.20E
SL7B8
7409A888

[DrJay]

Maybe whoever made the retail system overclocked it and tried to pull a fast one?

[Soulburner]

Originally posted by DrJay
Maybe whoever made the retail system overclocked it and tried to pull a fast one?
It is not a retail system.

I have seen this happen before, but it was an ES...

Latest BIOS?

[flaamer]

I'm running 1017 beta and don't have multiplier options :(

[hallowen]

Use the RELEASE 1017 Bios for the P4C800E, Not the Beta.

http://www.asus.com/support/download/item.aspx?ModelName=P4C800-E%20Deluxe&Type=BIOS

[skate2snow]

It is only a BIOS issue, Update both to the LATEST released version of BIOS....

[flaamer]

I'll try the latest bios when I get back to the computer. Do think it'll give more available multipliers or lock it at 16x?

[lutjens]

Post a CPU-Z of the chip...it'll determine what kind of chip it is without a doubt.

Its behavior sounds like a mobile chip...rin Sisoft Sandra to see if it supports SpeedStep.

Anyhow, a lucky chip...to get a high yield part and lower the multiplier is the reason why people buy ES chips.;)

[flaamer]

I've updated to the official release of bios 1017 on my P4C800Edlx. Multiplier is still locked at 14x and no sign in the bios to change.

Where in SiSoft Sandra would it show if it were speedstep capable?

[lutjens]

Should be under CPU Information. Check there for min and max multiplier as well.

CPU-Z should have showed it as a mobile CPU if it were in fact one, so it's looking more and more like Intel screwed up.

A good find, as Intel doesn't screw up like this very often, especially on things like multiplier...this is the first time, I myself have heard of anything like this happening.

Hopefully, this is a recurring trend.;)

[flaamer]

Unless they lock it at 18x :)

[freecableguy]

Wow. Just, wow. Intel dorked this chip up. Which leads me to believe that they probably LABELED it wrong, not set the wrong multiplier. Based on your OC I would bet that you are running a remarked 2.8E @ 3.8E. You MAY have gotten a chip that was supposed to be a 3.2E but I doubt that.

[NoAffinity]

This is a normal occurrence w/ pre-Prescott BIOS'es on both the IC7 (BIOS 18 or older) and the P4C800 (BIOS 1015 and older). The older BIOS'es can't ID the processor correctly, and it apparently defaults to a 14x multiplier. It doesn't seem to occur with any other Prescott, or even newer ones....and the fact that the newest BIOS doesn't correct it is good for you, flaamer! Time to invest in some faster RAM maybe? :D

[skate2snow]

Did you try 1016 BIOS?

And he bought a 3.2 to have a 3.2. Yes, if he whants to play w/ multiplier it is good, but it aint for him since his RAMs will perform better w/ 16 multiplier. And what you meen by faster RAM?

[NoAffinity]

Well, if he's gotta go 5:4 at 272, then some PC4400 or PC4500 would allow for higher 1:1. It was a hypothetical question, not necessarily a suggestion. ;)

[lutjens]

The multiplier lock doesn't care about BIOS revision. It is a hard lock and cannot be changed by software on the board.

The fact that a 14x multiplier has been burned for a a 3.2 processor is incredibly rare. The only way a 14x multiplier on a production 3.2E would be for a mobile chip defaulting to it's lowest multiplier (14x). Which is why I thought it may have been a mobile chip incognito.

Run Intel's Frequency ID utility and see what it says. Have some fun too...call their tech support and see what they say...they'll probably try to get it back from you.

Laugh at them...tell them thank you for giving you the ability to have a 266 MHz FSB today on an "unauthorized" board and without their permission.

That should tick them off...:D

[NoAffinity]

Originally posted by lutjens
The multiplier lock doesn't care about BIOS revision. It is a hard lock and cannot be changed by software on the board.
http://forums.extremeoverclocking.com/showthread.php?t=113084&highlight=effectively

[NoAffinity]

.....so Intel would have us believe.

[lutjens]

I still can't see how it's possible....the multiplier lock is on the CPU die itself.

Has been since the P2. So either Intel has changed how they lock these chips (making it possible to change the multiplier under certain circumstances) or there is a serious chink in the armor of the multiplier lock.

Funny that after all this time, there would be a flaw in the lock...it has beaten many attempts by many smart people to defeat it.

It could also be an early iteration of the frequency optimizations that are rumored to be coming out with the E0 stepping chips, that allow the chips to be clocked back when peak performance isn't required.

It is nonetheless very interesting news indeed...i'm eager to see how it plays out.:)

[NoAffinity]

Well, my theory is that, in a rush to get these chips on the market, Intel either inadvertently side-stepped part of the manufacturing process, or did it purposefully. I was looking for this processor for about a month and a half (nobody had them in stock at that point in time, and the Prescott's were still new on the market), before Newegg got them in stock in early May. This 3.2E was in my hands a week after the pack date. :D

Of course, I have absolutely no substation to this theory, it's all merely heresay, but if this is the case, then it indicates the the lock is both hardware- and software-enforced. I'm not the only that has experienced this (at least 3-4 other peeps at EOCF, and 1 or 2 others on OC.net), and there is even a FAQ about it at Intel's web site.

[skate2snow]

There is more then one that as those chips. Here I saw a good 5 people when they were released....

But flaamer CANT change it to 16...

[NoAffinity]

Originally posted by skate2snow
There is more then one that as those chips. Here I saw a good 5 people when they were released....

But flaamer CANT change it to 16...
Yeah, that's true....well hey, he got a massively OC'ing 2.8 there! :p