My socket 754 overclock

Pixle

Senior member
Apr 9, 2004
435
0
0
running a 2.41ghz socket 754 3400 cpu at @ 2.58 ghz.

Fsb is at 215... it seems to run aquark and i get a 64k score.

My question is...when I run Priime 95 it gives me an error saying the the accuracy of the rounding is greater than .5 or something like that. Is this a sign of unstable?

I uped the voltage a level 1.57.

The ram is corsair value select pc 3200. Anything I should do to make for a better and more stable overclock.

What am I currently running at in terms of speed against a 3800 or 4000 64?
 

CheesePoofs

Diamond Member
Dec 5, 2004
3,163
0
0
If prime gives you an error, it means that the cpu is unstable.

Lower your mem divider to overclock better more with the value ram.

A 4000 runs at 2.4ghz, so you are running faster than one.
 

Pixle

Senior member
Apr 9, 2004
435
0
0
I dont get why my cpu came clocked at 2.24 ghz...is this normal for a 64? socket 754?
 

Pixle

Senior member
Apr 9, 2004
435
0
0
Originally posted by: CheesePoofs

A 4000 runs at 2.4ghz, so you are running faster than one.

I dont understand this...my 3400 cpu came at a stock cpu speed of 2.41 ghz...did I get a special one or what?

 

Duvie

Elite Member
Feb 5, 2001
16,215
0
71
AMD

According to AMD that is the right speed for that chip....

It is also common for some board manufacturers for whatever reason, and maybe a rounding of frequencies, that a chip can be like 1fsb or some percentage of 1fsb off. At that multiplier being used that can easily throw out an oddball 2.41ghz reading...2.4ghz is stock...


Also realise that when you fail prime there is an issue. If you fail within minutes it will likely take up to .075v more actual vcore to gain stability to levels of 12-24+ hours....If it fails around 1hour or slightly less then it may only take .05v more....If it can go 2-4 hours and then throws out an error it likely only requires .025v to gain stability...However that may be moot if the board you give offers percentages alone so often you can only jump in bigger increments and not nice .025v increments....

Also remember when trying to test stability of the cpu one needs to ISOLATE (as Zebo has coined) the CPU...this means run more relaxed memory divider, speed, and timings as to rule out the ram as being a likely cause of any errors generated in prime95....ONly then can you really find the cpus limit....Failing to do this can make the adventure to find stability a long and tedious one....