Originally Posted by 10acjed
You should always configure your GTL's from a stable OC, so if you can get 100% stable at a lower OC then start adjusting the GTL's one at a time and test (usually 20 - 25 passes IBT will do) you will find that there will only be 1 or 2 settings per GTL will be stable.
Good advice, thank you.
Here's a few question about GTLs i hope you can answer.
From my understanding the GTLs are not dependant on the CPU's clock or multi, but rather FSB term voltage.
So if i have a overclock 100% stable at say 3.6ghz, and all i do is raise my motherboard's FSB and cpu vcore to compensate, should the GTLs need to be adjused if vtt stays the same?
To put it another way. If my VTT remains the same, should the GTLs also remain the same?
Also is it possible for me to have my GTLs wrong, even though i can pass hours worth of stress testing?
is it possible i can pass these tests, even for prolonged periods of time, but the GTLs still be off?
Stress tested the machine last night.
1.36250v vcore in bios
0.645 GTL (0/2)
0.655 GTL (1/3)
NB GTL set to Auto
Speed step disabled
These settings have provided me with 8 hours prime blend stable. 4 hours OCCT stable. 50 passes of Intel Burn Test on high stable. 2 hours P59 small FFT stable, and 2 hours large 2000k - 4000k fft stable with 2gbs of ram.
I guess based on this it's fair to assume my CPU can actually do 3.8ghz? I just haven't nailed some of the settings thus causing the 124?
I haven't yet tested this overclock to see if i can cause the 124, but i'm going to say it will.
I found if i lowered the gtls any more, the machine would blue screen. Setting the GTLs to auto also makes the PC unstable. I certainly need these to be manually configurable.
If i have the gtls at say 0.640 and 0.655 - BSOD, same as 0.645, 0.660, etc, etc. So i feel i've hit the sweet spot again.
That said previously, as you can see on the my first post. I had my GTLs at 0.720 and 0.725 with the same VTT (1.28) and that was 100% stress test stable too..... ***
Originally Posted by Denim-187
From my experience, an idle random BSOD generally comes back to your ram.
I managed to fix my 124 Bsod.
Try this for me.
Set your bios up to a state where you were happy and stable in prime95 etc.
Now once you have done this and re stressed again, go back to your DRAM voltage and drop it by .05v. So if your @ 1.65, lower it to 1.6 or if your @ 1.5, lower to 1.45.
Re stress. If it doesnt crash, your 124 random bsod wont happen anymore, if it does crash while stressing, raise it by .01v increments till its good.
Sent from iPhone4s using Tapatalk
Thanks, i will try. All my research and reading suggested this was a VTT/Vcore issue, glad to see someone has had instances where it is not.... might explain why i can't identify the cause of this.
In regards to memory voltages.
At stock speeds, 1066mhz, 5,5,5,15 CAS, this memory is rated between 2.0 and 2.1v by Gskill.
I have tried voltages rainging from 2.0 - 2.10 and that yielded no improvement. I have heard that these Asus motherboards tend to overvolt the memory by up to .05v, as such i've currently been running the memory at 2.04. So currently i'm running the memory right in the middle of its rated overated voltages.
I'll try undervolting and overvolting it a bit. I assume 1.9v - 2.2v this would be a wide enough range?
I have also run 8 hours of memtest86 on this PC to test the FSB and memory were stable before i started raising the CPU's multi. The ram did pass memtest86 without a problem @ 1077mhz with stock cas timings.
That said, i have heard that memtest86 is not that great at identifying ram faults related to overclocking. It's really only good for identifying hardware faults within the ram modules. It doesn't replicate 'real world' testing, as it doesn't stress the NB, CPU, and bandwidth like a PC would be in normal and intence operation such as Prime95 or benchmarking. As such faulty memory configurations that only show when the CPU and north bridge are hammering the ram dont show in memtest.
Also just for note i have also tried dropping the ram back as low as 895mhz, still same 124 bsod.Edited by matthew87 - 11/8/11 at 2:31pm