Unofficial Opteron 939 Thread

[ridikolous]

i finally got 3ghz stable today at 1.55 v should i stay where im at right now or should i shot for more ghz?

[blazin-asian]

i finally got 3ghz stable today at 1.55 v should i stay where im at right now or should i shot for more ghz?

 

which stepping do you have and what multiplier, divider and such? more info!

im sure you could go for more MHz. just watch your temps. lets see 3.1

 

im on a quest for a prime stable 3.0. i can boot and benchmark, but prime gives me errors :confused:

 

i lowered the HTT to 2x to give the mem controller a bit of breathing room and priming now.

[Guest LithoTech]

I was wondering how much raising the chipset voltage helps in getting a good overclock. Is there a link that im not finding that would help me figure this out. I am currently at 1.7 and I think thats pretty good. Should/can I go higher if the temps are ok? Would it even help?

 

thx fellas

 

I've been wondering that myself, as well as the LDT voltage. Not so much as to how high to put it, but more as to how low I should be trying to get away with; and what sort of errors or problems should I be looking for when they need more or less volts.

 

If I ever get around to trying to look it up. I keep getting side tracked with other things to read and try.

 

It just seems logical to me to keep voltage settings as low as possible. So I often bump them a notch and see if it solves whatever wall I hit, if not I put them back and look for another solution. This way they rarely end up getting raised and kept raised unless it is really needed.

 

Less volts = less heat, and less heat = longer life of components. Probably a lot more stuff factoring in there, but obviously too much heat will drastically reduce the life of a component.

 

A good analogy: think of a car with an NOS system, but every time you use it the life of the engine is reduced by 10k miles... as well of the possibility that it will simply explode.

 

I'm just yappin here, thinking allowed. I'll shut up now. :nod:

[blazin-asian]

omg...this is so f-in annoying. my prime no longer gives me errors. it will just crash and freeze my whole system. did prime test and it did this 1.5 hours into it.

 

my hopes for a stable 3.0 are fading

[cjoe7]

omg...this is so f-in annoying. my prime no longer gives me errors. it will just crash and freeze my whole system. did prime test and it did this 1.5 hours into it.

 

my hopes for a stable 3.0 are fading

 

Have you tried burning it in with a slower speed and lower voltage? Are you topless, WC, Air?

 

cjoe

[blazin-asian]

Have you tried burning it in with a slower speed and lower voltage? Are you topless, WC, Air?

 

cjoe

 

sorry, but what do you mean by burning it in?

 

i still have my IHS on and im cooling with TTBT. i have yet to get to 50C load so i think my temps are still good. im too chicken to completely void the warranty by removing the IHS.

[cjoe7]

sorry, but what do you mean by burning it in?

 

i still have my IHS on and im cooling with TTBT. i have yet to get to 50C load so i think my temps are still good. im too chicken to completely void the warranty by removing the IHS.

 

Running a burn-in program like Prime95, SP2004, Sandra, OCCT, S&M, CPUBurn.

 

By running these programs, the CPU works at 100% utilization which stresses and pushes the CPU. Over time, the CPU should achieve higher OC at a lower voltage. Its analogous to stretching fabric.

 

There are many threads about burning CPU and memory strategies.

 

cjoe

[blazin-asian]

Running a burn-in program like Prime95, SP2004, Sandra, OCCT, S&M, CPUBurn.

 

By running these programs, the CPU works at 100% utilization which stresses and pushes the CPU. Over time, the CPU should achieve higher OC at a lower voltage. Its analogous to stretching fabric.

 

There are many threads about burning CPU and memory strategies.

 

cjoe

 

oh sorry. well, i run Prime95 and SP2004 for hours upon hours after every change to test for stability. i dont really like OCCT because it doesnt give me any information at all (not like i would know what its telling me anyways).

 

i also do Super PI 32M to test for stability before i use SP2004.

 

does it sound like im doing it right so far? my cpu is @ 1.6v now so im hoping my later OCs will require less voltage.

 

EDIT: :mad: tried 300x10 with 133 divider and still no go!!! wtf....im switching to 2T and see if it makes a difference. I HATE MY MEMORY CONTROLLER!

[cjoe7]

oh sorry. well, i run Prime95 and SP2004 for hours upon hours after every change to test for stability. i dont really like OCCT because it doesnt give me any information at all (not like i would know what its telling me anyways).

 

i also do Super PI 32M to test for stability before i use SP2004.

 

does it sound like im doing it right so far? my cpu is @ 1.6v now so im hoping my later OCs will require less voltage.

 

You're doing fine. OCCT is really good because its fast at finding problems.

 

Run Prime95 or Sp2004 @ 2.5G 1.37V or less volts. Try to run 2.5 at the lowest error free voltage for 8hrs+ (when you sleep). After awhile, you'll notice that it takes less volts to run that same speed. Slowly bump up the speed while maintaining the lower voltage.

 

Today, I'm priming @ 2.7G 1.28V. Tomorrow, I may prime high @ 2.9 1.435V. I'll alternate high and low to stretch the CPU's capability.

 

cjoe

[blazin-asian]

You're doing fine. OCCT is really good because its fast at finding problems.

 

Run Prime95 or Sp2004 @ 2.5G 1.37V or less volts. Try to run 2.5 at the lowest error free voltage for 8hrs+ (when you sleep). After awhile, you'll notice that it takes less volts to run that same speed. Slowly bump up the speed while maintaining the lower voltage.

 

Today, I'm priming @ 2.7G 1.28V. Tomorrow, I may prime high @ 2.9 1.435V. I'll alternate high and low to stretch the CPU's capability.

 

cjoe

 

wow, thats freakin sweet. ive already primed about 24 hours worth @ 2.97 (1.584v) so maybe i'll try 1.55v or something. if i cant hit 3.0, i might as well work on getting 2.97 on as low volts as possible.

 

gonna try 3.0 @ 1.6v (2T) for now. good thing i have a laptop or else i wouldnt get ANY work done, lol.

 

EDIT: getting a lot further with 2T then i thought. 2.5 hours prime stable so far....better not get my hopes up for at least another 10 hours.

[St0ffel]

ok, i have some really freaking timings now^^

i tested yesterday some tighter timings, but 3-4-3 is a must for my sticks, 3-3-3 couses ~500 faults test #5 @98% and 3-4-2 is a no boot to memtest^^

 

but 3-4-3-0 works

ill update my OCDB entry next time!

 

[Guest cloud_171]

So far I've been able to get it stable 2.8ghz @ stock vcore. I'm trying to hit 2.9ghz with 1.525v and the load temperatures are getting around 46 degrees. Should I still push it higher?