• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Sandy Bridge stability ???

Shetlander

Shetlander

Shetlander

Associate
Joined
12 Jul 2004
Posts
377
Location
Poole Dorset
Im currently working on getting my build Prime95 stable but seem to be finding it impossible. Ive had it as high as 4.7Ghz where it will pass hours of LinX on max settings and Intel burn. Only problem is when I set Prime95 going it runs for a while and then either freezes the computer, BSOD,s or restarts the computer. Its done this after as much as 5 hours and usually in the middle of the night.

Has anyone managed a Prime95 test for 24 hours? Im curious to see if the program has issues with the hardware. I dont see how it could but Im sure I remember something about it happening with a previous new chip.
 
Unstable overclock probably needs more vcore. I'm having troubles with my MSI board and stability it undervolts randomly causing BSOD's not sure if this might be happening to yours?
 
I'm actually having very similar issues. Currently working on 4.8ghz with a goal of 5ghz stable overclock (have got both linx stable anyway.) Here is a screenshot of 4.8 at 1.38vcore.

4.8ghz_1.384v.png


However I've ran prime95 at multiple vcores all the way up to 1.42v and within 1 hour I get BSoD or freezing. Just also tried 1.45v just to try to see if I could get a longer run and withing 20 minutes the pc completely reset and booted to login screen with no error.

What BSoD error are you getting? Mine ends in 00124 which according to various places could be a memory controller error, so away to try either more voltage for VVCIO or running prime95 for as long as possible at stock.
 
I am finding that to get stable above 4.6Ghz requires core voltages in excess of what is currently recommended.
If your a bencher it wont bother you much, if you are worried about the welfare of your CPU then it will.

You can boot and run benches at fairly low vcore but long prime/IBT runs will blue screen or reboot the PC.

I set my own limit at 1.4v shown on cpuz under load, 1.385v in bios.
This allows me to run a stable 4.7Ghz.

Above this required a lot more voltage and wasn't truly stable especially at 5 Ghz.

I also found running the memory at very high clocks to be very unstable.
2133 Mhz is great for aida64 benches but not for stability.

I compromise at 1833Mhz on the RAM with 1.575v mem volt.
 
Ive just upgraded my bios to the latest 1203 this morning and when I left for work it was priming using the auto overclock setting to see if the board has it right. Looking at you settings above I may have been a little conservative with the volts. I was running 1.365v in the bios and had it stable running hours of LinX but not passing Prime95. I didnt appreciate it would need such a big increase to pass both. I will test it tonight overnight and see how it goes.
I have also just changed my memory to the faster G Skill 2133Mhz set and moved it into A1 & B1. Fingers crossed it will still be running when I get home.

Come to think of it I wonder if it could be the memory over heating? My new G Skill comes with a RAM cooler so I may give that a try. The airfolow isnt blowing a tornado through it as the CPU/GPU are watercooled so the fans are low for quiet.
 
Last edited:
I am using GEIL platinums, they have an XMP profile for 2133Mhz.
4 of them for 8GB so all banks filled.
With only 2 installed it may be possible to run 2133Mhz stable i briefly tried it but didn't spend much time on it.
The volts needed on the RAM were too high for my likieng
 
Asus P8P67 pro

1.485v in the bios, and my load volts read 1.440 volts when I look at cpuid. Temps are 68 degrees at the moment

Still think there is more in it to be honest

15 mins and it'll be 24 hours :D
 
cmsbfent said:
Asus P8P67 pro

1.485v in the bios, and my load volts read 1.440 volts when I look at cpuid. Temps are 68 degrees at the moment

Still think there is more in it to be honest

15 mins and it'll be 24 hours :D
Click to expand...


are you posting on the machine you are testing ? doesn't that invalidate the test ? as - if you use other programs at same time - it reduces the load from 100% on the stress test ....
 
Hey

No, was using a laptop

Cant see that being true though for what its worth. If anything, it would make more sense to try and do other things, like read discs etc, just to try and make the lines fluctuate

Stopped it at 26 hours
 
Thats some volts for 4.7GHz.

I ran OCCT last night and have a strange problem. On the 12v graph, it shows the voltage as 7v. Also, Everest shows the 12v rail as 7v. Is that an incompatability of the programs or is that likely. I wouldnt think it would run on 7v or it would be very flakey indeed. In the bios it shows as 12.338v. Is there any other progs I can use other than HW monitor as it doesnt show my 12v rail.

On a side note, when I got home last night it had been priming all day with the standard overclock the board set (103 x 42 = 4.326Ghz) and on core 0 and 2 it had failed with a rounding error after 7 hours but the other 2 cores where still going. That was with my G Skill 2133Mhz RAM.
 
Maybe a bit high, but thats what testing is all about. Gets unstable any lower, so up it went. Temps aren't really an issue though and thats my biggest concerns.

Now at 4.8ghz, was running at 70 degrees when I left for work and that had been going for 3 hours so we'll see if its still going when I get back

For that, I havent touched the voltage (still at 1.485 in the bios) and just changed the LLC to very high from high, which made my load volts go from 1.440 to 1.480 - both with +- fluctuations of .008
 
Well....

12 and a half hours into prime at 4.8ghz, and my temps are now at 69 degrees according to the asus monitor. This read 68 when I was at 4.7ghz

Prob gonna leave it over night to hopefully reach 24 hours, and then go for 4.9ghz :)

You had any luck getting stable yet mate?

These 24 hour stability tests dont half take some time to complete ;)
 
I had it running overnight at 4.7Ghz with 1.380v in the Bios. It was all ok when I went to bed but when I got up this morning there was a pop up which stated "Prime95" had stopped redponding". No BSOD, no restart. PC was working as normal. Funny thing is it stopped at 3.18am which Im sure is when its frozen in the past. Im starting to wonder if the heating coming on in the middle of the night is causing a sufficient power drop/surge to unsettle th PC enough to cause a wobble. I set it going this morning so It will be interesting to see if its still going when I get home.
 
cmsbfent said:
Asus P8P67 pro

1.485v in the bios, and my load volts read 1.440 volts when I look at cpuid. Temps are 68 degrees at the moment

Still think there is more in it to be honest

15 mins and it'll be 24 hours :D
Click to expand...

Good to hear, got the same board but savin for my cpu cooler (still using stock which I don't fancy overclocking) just gotta decide which to buy
 
You must log in or register to reply here.
Back
Top Bottom