Prime95 error after 20 mins. Now what?

Primus462

Distinguished
Dec 19, 2005
244
0
18,680
I upped my 4200+ X2 to 2.507GHZ (228x11) and started two instances of Prime95 and went to bed. When I got up this morning I found out that it had crashed. I increased the voltage by .025 to 1.3125 and started it again, and headed off to work.

If it errors again, where do I start on figuring out what caused it to fail? I'm new to this.

Thanks!
 

mkaibear

Distinguished
Sep 5, 2006
678
0
18,990
It's just a guess, but I suspect what caused it to fail is THE OVERCLOCK!

Try running Prime95 with no overclock at all, see if it runs with no errors. If it does, you've clocked your processor too high and it's erroring.
 

Primus462

Distinguished
Dec 19, 2005
244
0
18,680
I believe that comment was unnecessary. Obviously, it's unstable because I overclocked it, I am aware of that. I'm trying to figure out what I should try adjusting in order to make it stable. How do I know what caused the error? Where do I start? I had it running at 2.4GHz for a couple of weeks with no problems. Now I'm trying to get it to 2.5GHz.
 

samir_nayanajaad

Distinguished
Feb 22, 2006
331
0
18,780
Okay few questions.

Did you oc the cpu and ram separately to find their max oc individually or did you just start to up the HTT bus?

Have you changed your ram timings, voltages, and dividers?

Have you taken as much load off the chipset as possible by disabling unused things such as a floppy drive, firewire, extra usb ports ect.?

So not to leave you with no help and only more questions I'm gonna assume some answers to those questions and maybe give you some help.

I bet you just started by upping the voltages and htt bus speed and that is most of how you overclock but you need to tweak your bios so that other parts can handle that. First I would look at your memory dividers and adjust them so your ram is running below their speed rating.

That will hopefully eliminate the ram as the cause and you should also make the timings kinda loose too just to make sure.

next is the chipset load. the chipset handles all those bells and whistles on your mobo, so to lessen the load on the chipset disable every bell and whistle you dont use. then bump up the voltage to it a small amount to give it some more stability.

I know this article is getting dated but you should find some useful information to give you a good starting point. It’s not for x2's but many basic principals should still hold true.
short media guide to oc 939 and 754
 

htoonthura

Distinguished
May 21, 2006
651
0
18,980
It's just a guess, but I suspect what caused it to fail is THE OVERCLOCK!

Try running Prime95 with no overclock at all, see if it runs with no errors. If it does, you've clocked your processor too high and it's erroring.

You can not call it overclocking too high when he overclocks to 2.4 from 2.2
 

mkaibear

Distinguished
Sep 5, 2006
678
0
18,990
>can not call it overclocking too high

No, you really can. Some processors are speed binned at a particular point but will get a lot faster when o/c'd - some will not o/c at all (though you'd have to be quite unlucky to get that!).

Processors o/c as far as they will go, and no further. You can't ever say "this is not too high" - if it's erroring, it's too high!
 

mkaibear

Distinguished
Sep 5, 2006
678
0
18,990
>obviously, it's unstable because I overclocked it,

Fair enough. Your question implied you hadn't considered the possibility that it might be the overclock...

Follow samir's advice!
 

htoonthura

Distinguished
May 21, 2006
651
0
18,980
>can not call it overclocking too high

No, you really can. Some processors are speed binned at a particular point but will get a lot faster when o/c'd - some will not o/c at all (though you'd have to be quite unlucky to get that!).

Processors o/c as far as they will go, and no further. You can't ever say "this is not too high" - if it's erroring, it's too high!

I see your points but it can not be called too high if it is erroring. You can fail to get a good overclocking result due to many facts like processor, ram, cooling issue , for example. When you overclock to 2.5 from 2.2 , and you fail it, i would not say the processor has reached its limit. It could be the processor , but it is not very likely.

Bye.
 

Primus462

Distinguished
Dec 19, 2005
244
0
18,680
I have not messed with the ram except to drop it from 800 to 667. Windows wouldn't boot when I did this. Also, the timing was automatically changed to 5-5-5-12. I changed it back to 4-4-4-12 and 800MHz. It'd boot fine then.

How do I figure out how fast my RAM is running based on how much I OC'd my FSB?

I increased the voltage by .025 this morning and then started Prime95 before I left. Who knows if it's still running or not. I'll find out when I get home.
Have you taken as much load off the chipset as possible by disabling unused things such as a floppy drive, firewire, extra usb ports ect.?
No. I haven't tried that yet.
 

Primus462

Distinguished
Dec 19, 2005
244
0
18,680
It's just a guess, but I suspect what caused it to fail is THE OVERCLOCK!

Try running Prime95 with no overclock at all, see if it runs with no errors. If it does, you've clocked your processor too high and it's erroring.

OK, last night I lowerd it back to 2.2GHz and ran Prime95 on both cores. One instance ran for 24 minutes and the other ran for 3+ hours before I stopped it. What's the deal!!!!!!!!!!
 

coroder

Distinguished
Oct 16, 2006
8
0
18,510
First of all a personal experience,

I have managed to get my 4400+ running at 2.7Ghz and i find it unlikely, though possible, yours is at its limit.



Second, about prime95

More common than not in dual core processors one of the cores will fail before the other, there is no way around this and a system cant really be considered stable until both cores pass.
 

Primus462

Distinguished
Dec 19, 2005
244
0
18,680
How can I determine why mine is failing when it's at stock speed. Should I try resetting my BIOS and starting from scratch?
 

shata

Distinguished
Dec 10, 2005
727
0
18,980
Put the HT multiplyer at 1x then you should be safe up to 1.5-1.55 Volts (not responsible when somthing happends) But i have my 4000+ at 1.5 and its just peachy so far.

Sometimes the HT causes instablity thats why 1x is a good choice, also ive read the HT does nothing for you really so it doesnt matter what its at.
(what ive heard not sure if thats true, it hasnt had effect on super pi bench)
 

shata

Distinguished
Dec 10, 2005
727
0
18,980
I dunno what the x2's can handle but my 4000+ is stable at 2.8ghz with vcore of 1.5v

1.33 isnt that much voltage as far as im concerend.