Intel Q6600 quad overclock questions

Status
Not open for further replies.

scatmanwfm

Distinguished
Dec 11, 2009
34
0
18,530
Hey all. I've inquired about overclocking my q6600 before and was grateful for the help. I've had it running stable at 3.2ghz at about 1.19v. Under prime95 it never goes above 40c. I've been trying to maybe push it a little further and have some quick questions.

I've noticed that when I overclock it to 3.4 with a multiplier of 9 it usually fails as soon as I start the prime test and I noticed that the voltage will fluctuate between 1.19 and 1.24 when I have it on auto but when I set it to run at anything past 1.27 it becomes stable but when I looked on the asus software and on cpu-z it still says it's 1.19-1.24 after I've set it to past 1.27. Is this what is called vdroop?

Also to be able to clock it higher is it better to leave it on auto? Can some overclocks that fail on auto be made stable by manually setting voltage? Thank you for any help I appreciate it.
 
Solution
First of all, overclocking memory (running memory faster than 1:1) has, except for memory i/o benchmarks, little effect on real world performance. Here is one place that we have discussed that:
http://www.tomshardware.com/forum/251715-29-ratio-myth

Second, for a given core speed, performance does not scale linearly with FSB frequency.

You will maximize performance by running the multiplier at its maximum, running FSB:memory ratio at 1:1, pushing the FSB freq as high as you can, and then tightening the memory timing as much as you can.

It also helps to get a CPU with a low VID and get a good case/cooler combination. An ACF7P is, by today's standards, a better than stock but not great cooler.

A C2Q will most likely reach temperature...

rand_79

Distinguished
Apr 10, 2009
829
0
19,010
I have a q6600 and a p35 chipset.

I do 8x420~~3.36ghz

the fsb oc is more important than another 100 or 200mhz of cpu oc.

it also works out well with my memory at a 1.2:1 multiplier to get an overclock of near 1000mhz (stock 800)

how many threads of prime95 are you using. with a good heatsink mine still gets to 65C or so.

 
i have a core 2 duo e6600 and yes the vcore set by the bios is higher than what is actually reported by hwmonitor. Im not sure of the reasoning for this. On my e6600 i have to go to 1.28v in bios to get to 3ghz stable. When I get home ill check what the actual readings are compared to what is set in bios, but im pretty sure most people find that the voltage in bios is not the actual voltage being read by monitoring programs.
 

scatmanwfm

Distinguished
Dec 11, 2009
34
0
18,530


Running 4 threads and I use a freezer 7 pro 92mm heatsink.

I thought running with a 9 would generate less heat. My motherboard seems to put out a little more heat with a higher fsb. My case has great airflow and stays real cool but I live in florida so I was trying to keep the heat as low as possible. Is it that much more of a performance increase going with 8? I was able to get 3.6 stable @ 44c with prime95 and thought I would try and see what I could get with 9 and it doesn't seem I can go as high with 8 unless I put the voltage up into the 1.3 range. So it would def be better going with 8?



 

rand_79

Distinguished
Apr 10, 2009
829
0
19,010
well you are overclocking more of the computer that way.= faster.

I benchmarked it at 8x and 9x and there was a significant difference with the fsb overclocked and 8x. I also had problems with 9x and 3600 doing overnight testing.

its been rocksolid at 3.4~~ for over 2 years now.

my voltage is also 1.3x

but reads as 1.28 from cpu-z
 
First of all, overclocking memory (running memory faster than 1:1) has, except for memory i/o benchmarks, little effect on real world performance. Here is one place that we have discussed that:
http://www.tomshardware.com/forum/251715-29-ratio-myth

Second, for a given core speed, performance does not scale linearly with FSB frequency.

You will maximize performance by running the multiplier at its maximum, running FSB:memory ratio at 1:1, pushing the FSB freq as high as you can, and then tightening the memory timing as much as you can.

It also helps to get a CPU with a low VID and get a good case/cooler combination. An ACF7P is, by today's standards, a better than stock but not great cooler.

A C2Q will most likely reach temperature limits before it reaches voltage limits. A C2D is just the opposite.

44 C load temps give you a lot of headroom. You have about 25 C to spare.
----------
Overclocking since 1978 - Z80 (TRS-80) from 1.77 MHz to 2.01 MHz
:)

 
Solution
Status
Not open for further replies.