*Official Retail G0 Q6600 Overclocking Thread*

[justin_c]

looks like we have a problem; while i was tweaking the gtls, i noticed this:

small ffts=min ffts size of 8k, max fft size of 64k

your screenshot says "1280k passed". if you dont understand what im talking about, read again.

so you were running blend, not small ffts as most of us do. try running small ffts. i just helped out a guy @ forum.ncix.com saying that his comp failed blend after 22hrs but failed 10 mins of small fft. dont you agree GripS?

[crossg]

By the way when it ask for a user name put in TheWolf as user & team number 13285. Team Abit & I will appreciate it.

LOL.Worth a try.

[GripS]

you are quite borish with your fantasy of using prime for a minimum of 8 hours as a justification that a system is stable...you need to grow up and join the adult world...get beyond this fantasy world of "show me your prime 95" to justify a high overclock.

when you get that evga 680i board, come on back and splash your sucess!...i'll be waiting for you!

Is the insult really necessary? It is very much possible to disagree with someone without insulting them. Prime95 is well known and has been for quite some time as one of the "true" tests for stability. That's why a lot of us use it and expect to see someone post a screenshot when they are claiming stability. You need to learn to 'agree to disagree' once in a while.

looks like we have a problem; while i was tweaking the gtls, i noticed this:

small ffts=min ffts size of 8k, max fft size of 64k

your screenshot says "1280k passed". if you dont understand what im talking about, read again.

so you were running blend, not small ffts as most of us do. try running small ffts. i just helped out a guy @ forum.ncix.com saying that his comp failed blend after 22hrs but failed 10 mins of small fft. dont you agree GripS?

Yes i agree with you. After you mentioned small fft to REALLY stress the CPU i did myself. Higher temps and failed at the current voltage(1.4v). I had to increase a touch more to achieve stability(1.42v). Without droop i'm sure it could have been fine at 1.4.

I did not specify running small fft's and i think running blend at that speed for that length of time is remarkable. I think that truly is a gem of a chip he has there.

Let's just say we test stability one way and he tests another way and be done with it. It's really not worth arguing about anymore.

[fornowagain]

Blend doesn't push the cpu that hard at all, good for testing the system memory after small fft's has passed with higher temps.

Oh btw.

[crossg]

Just trying to find my CPU's max and get a low 12's 1 m SPI. It may go higher but there is really no sense. Req'd 1.6125 Vcore. No stability tests whatsoever.

[pookguy88]

my vid in coretemp fluctuates between 1.1625 and 1.3x ... why is that?

[ChaosMinionX]

my vid in coretemp fluctuates between 1.1625 and 1.3x ... why is that?

Odd, I have never seen that. Mine shows up as a solid 1.250.

Some 4ghz screens and benches coming soon as I get some new hardware. Have only benched Futuremark Trio at 3.9ghz DDR2 974. Will be going higher shortly Its been to damn hot out to bench unfortunately

[Boble76]

1,51V from motherboard
3751MHz 8M
3901MHz 1M ok

1.54V
3901MHz 8M ok
1.57V
4002MHz 8M ok

1.60V
4002MHz wprime 1024m ok.
4052MHz wprime 32m ok.

P5K Deluxe WiFi AP
Q6600 G0 L726A951 (water + 28C)
PC8000

1.7V @ 4280MHz with SS Rotary 3DMark06 ok

[ChaosMinionX]

1,51V from motherboard
3751MHz 8M
3901MHz 1M ok

1.54V
3901MHz 8M ok
1.57V
4002MHz 8M ok

1.60V
4002MHz wprime 1024m ok.
4052MHz wprime 32m ok.

P5K Deluxe WiFi AP
Q6600 G0 L726A951
PC8000

1.7V @ 4280MHz with SS Rotary 3DMark06 ok

What batch you get? 720B like all the other monsters?

[situman]

Better NB cooler = better clocks

[crossg]

my vid in coretemp fluctuates between 1.1625 and 1.3x ... why is that?

You probably have speed step or CI enabled in Bios and it shows lower vid with speed step.

[eagle_]

With my G0 I can only get 3200 (@1.25v) stable.

Higher than that and it starts throttling (lowers multi to 6x automatically), even with low temps (using WC, temps are near 48º C full load on all cores at 3200). For example, at 3400 it runs for a minute or so at full speed but when it gets to mid 50's (coretemp & RT) it's starts throttling between multi 6 and 9.

HW used:
- Asus P5K vanilla (vdroop mod)
- 2x 1GB Team Elite DDR2-667
- LC Power 550w GP
- Ninja with TT2 fan (air)
- Eheim 1048 + Thermalright XWB-1 + 240 radiator & noctua NF-S12-1200 + 120 radiator + Scythe 1200

Any ideas? I really would like to have it stable at least at 3500...

[situman]

You must be on stock cooler, cause those temps look ridiculously high at 2ghz.

[justin_c]

read his post again situman :P

Don't use TAT, it was designed for pentium M processors, so please do not use extinct software. instead use prime v25.4 and prime all 4 cores with small ffts, reading temps from coretemp (or speedfan, if you use speedfan, you must set a 15 degree offset for each of the cores in the settings for INTEL CORE).

as for the 6x issue, i have had that before with extremely high temperatures. look for a "disable temperature throttling" something like that in your bios.

use rmclock and go to Advanced CPU settings, then uncheck TM1, TM2, and extended throttling, then apply.

[Nater]

read his post again situman :P

Don't use TAT, it was designed for pentium M processors, so please do not use extinct software. instead use prime v25.4 and prime all 4 cores with small ffts, reading temps from coretemp (or speedfan, if you use speedfan, you must set a 15 degree offset for each of the cores in the settings for INTEL CORE).

as for the 6x issue, i have had that before with extremely high temperatures. look for a "disable temperature throttling" something like that in your bios.

use rmclock and go to Advanced CPU settings, then uncheck TM1, TM2, and extended throttling, then apply.

What's the advantage of small FFTs nowadays? I remember, that some time ago small FFTs were used because they would fit into the processors cache, alleviating any dependence on ram. However, considering the Core 2 Quads have 2x4MB of cache, the largest FFTs are 1024KB, you'd think that Large FFTs would fit completely in the cache assuming you were doing a pure test session and not web browsing or doing anything else at the time.

Is this not the case? I've found that my machine usually craps out faster on Large FFTs than the smaller ones, heat is slightly higher on the small units. If the larger units are using ram, I doubt it's what's causing errors. I am running my two sticks of GeIL Ultra PC2-8500 way below it's spec.

[justin_c]

well my machine craps out right away if i do small ffts at 3.95, otherwise its done large for 1/2hr without crapping out (so far).

[pokerezzet]

Some benchmarks at 4 ghz. Isn´t the superpi result to low or it seems about right?

[[XC] Synthetickiller]

I'm getting 13.771s @ 3.7ghz. I'd imagine 300 mhz would shave off .8 seconds or so....

[crossg]

Did a few 3D Mark 06 runs at 3924 Mhz and couldn'd figure why I was getting lower scores than my previous Q 6600 B3 runs at 3654. I thought the GPU OC was not holding but it turns out the newer 162 Nvidia drivers give me a score average of 13600 and the 158 gave me 14200 but the cpu score was good. Seems 3dmark stable as I ran the test about 6 times trying to figure out what was wrong. Sorry about the screenies but the last time I posted a 3dmark link it didn't work.

[Ace-a-Rue]

run into windows...programs operated normally

[GripS]

Nice chip. Looks like it's true that the B's are killers.

[r4t]

hey everyone,

finally got my new b0x a few days ago!
as yall can see its runnin' @ default..
blastin' @ full load for a while the temps dont seem to go above that, courtesy of a Zalman CNPS9500..

now i've been followin' this thread, checkin' out what this baby can do, n i've decided maybe ima keep mine @ like 3.4GHz, since mine's not WC'ed nor do i wanna risk killin' it

now the thing is i've never done any OCing up until this point
so i got no clue as to how to get started

i'll be needing some of u wizhats to gimme a few pointers

AI Booster? AI NOS? BIOS straight up?

comparing with the examples here the default FSB seems very low..
if possible i'd like to keep the vcore intact unless recommended otherwise..

TIA

PS: mobo=P5K

[Boble76]

run into windows...programs operated normally

How much vCore you have?

Now I'm testing 8x475FSB and seems to be ok.

[Ace-a-Rue]

How much vCore you have?

Now I'm testing 8x475FSB and seems to be ok.

i put it at 1.6 actual! ...might have been a possibility that i could have tweaked it down a notch or two...i just went into windows made a post sent some email without a hitch...ran one instances of superpi 1M and got about 12.79 secs....i was doing 1:1, loose timings and the memory was at 456mhz.

[Ace-a-Rue]

Nice chip. Looks like it's true that the B's are killers.

AND...the amazing part is the heat output was hardly even noticeable for idle,; at least for my dual fan "TRUE".