**- Official Asus P5K Thread -**

What max temps under load of having 4 instances of Orthos running for 20+ mins should I really be aiming at ?

I'm a little surprised and disappointed that at 8x425@3400 with only 1.4v that I'm hitting between 71 and 74 temps, taking into account I've got the Antec900 case with both front 120mm intakes on full, rear 120mm full, top 200mm full and an extra 80mm on the side window. Used AC5 between cpu and Tuniq, the Tuniq fan is also on max, room temp isn't out of the ordinary either

Should I really be aiming at <70

Think I'm going to try some other stress tests next
 
I'm sure some have hit about 540, I hoped to get 500 myself but initial tests I can't get over 485 whatever I try, however maybe my DDR-800 memory doesn't like 970 at 1:1 even with more volts and relaxed timings
 
Just curious what the Max FSB anyone here has hit? If they could post their BIOS setttings that would be great.

I have an e6300 coming in the post and am hoping its a good clocker.

Your P5K will trash an E6300 within an inch of its life. why go from a Q6600 to an old C2D though???

What max temps under load of having 4 instances of Orthos running for 20+ mins should I really be aiming at ?

I'm a little surprised and disappointed that at 8x425@3400 with only 1.4v that I'm hitting between 71 and 74 temps, taking into account I've got the Antec900 case with both front 120mm intakes on full, rear 120mm full, top 200mm full and an extra 80mm on the side window. Used AC5 between cpu and Tuniq, the Tuniq fan is also on max, room temp isn't out of the ordinary either

Should I really be aiming at <70

Think I'm going to try some other stress tests next

I wouldnt go over your current temps tbh, of course thats going to be the highest you will ever see as its stressing your cpu 100%

95C you chip will start to throttle and 100C it will shut down, thats the Tj max of our quads
 
Why is it i can run 400 x 8 fine in windows, i ran orthos for 8 hours with that, but when i try to go past it, ie 8 x 425 or 7 x 377 to get 3400mhz windows becomes extreamly slow and sluggish and nothing works properly, eventully resulting in orthos throwing up an error but it hasnt restarted as of yet.

Ive enabled the cpu dampener and thats helped with the voltages, is it worth upping the sounthbridge voltage ? and setting a relax level? and updating the bios ?
 
Last edited:
What max temps under load of having 4 instances of Orthos running for 20+ mins should I really be aiming at ?

I'm a little surprised and disappointed that at 8x425@3400 with only 1.4v that I'm hitting between 71 and 74 temps, taking into account I've got the Antec900 case with both front 120mm intakes on full, rear 120mm full, top 200mm full and an extra 80mm on the side window. Used AC5 between cpu and Tuniq, the Tuniq fan is also on max, room temp isn't out of the ordinary either

Should I really be aiming at <70

Think I'm going to try some other stress tests next


That does seem high, my high VID Q6600 at 3.5Ghz in an antec 900 never goes over 59C when priming and I have with front and back fans on low, top fan on medium and a lowly Freezer 7 cooler on it. Maybe your needing to do a little lapping?
 
Settled for [email protected] 1.425v
Memory at 1:1 900MHz 4-4-4-12 2.2v
FSB 1800MHz 1.55v
Had to drop GTX clocks slightly

cpu pll 1.6v
fsb termination 1.3v
sb 1.2v

Prime95, Orthos & ITAT stable - no errors and max temps a shade over 70

3DMark01 up from 47k to 64,317
3DMark05 up from 15k to 21,223
3DMark06 up from 12k to 15,206

Temps are fine for synth tests

Crysis demo benchmark 48.79fps
F.E.A.R in game test ave 82fps, min 41fps
(on 1920x1200 with everything maxed out except soft shadows off - 16AF, 8AA)

This is as far as I think I'm going, got to resist temptation of meddling any more I think
But then again how about memory divider to 1000+ with relaxed timings or 8x460 for 3.7GHz..........lol
 
Last edited:
Your P5K will trash an E6300 within an inch of its life. why go from a Q6600 to an old C2D though???

Good to hear. Anyway I'm going back to an e6300 because the power of my quad is wasted tbh. It's only good for 3dmarks and raring files :p ...but the main reason is because I've just purchased a 22'' Samsung monitor and need to pay off my credit card...and the e6300 will do everything I need it to (having it clock to 3.4-3.6ghz would be awesome..I've been wanting a 100% oc for awhile now :D )
 
Hi

Just got the basic version of the P5K, and trawled through this very useful thread (thanks). Using 4x1gig sticks seems to have a recommendation with it of increasing the Northbridge voltage for stability - just wondering if this also applies to a system running everything at stock speeds?.

joe

*EDIT* one other thing, when installing the xp drivers from the supplied cd, are there any parts of it to avoid? - just a bit paranoid from experience with my old asus A8N Premium, where installing the nforce IDE drivers was a no no at the time.
 
Last edited:
anyone?. I have set the NB to 1.4v for now, rather than auto. I do plan on taking the cpu to 3.2ghz tonight (Q6600), so not sure if that NB will need another step?.

(disregard the *EDIT* question - i used the drivers anyway - seem ok)

thanks
joe
 
sorry as I am sure this is covered but cant find it...

New build with q6600 and AF7 is showing an idle temp of 18 degrees in Bios??? Did the first BIOS have a problem with under reporting?
 
Slower 3dmark06 after overclocking, why????

Hi,
Have got my Q6600 running at 460 FSb @ x7 to give 3.24MHz and it is rock solid on Orthos for 12 hours. I then started to run some graphics benchmarks to start on the GPU overclocking but have come across a very strange effect where the results are as below:

CPU FSB Mult Memory GPUBus GPUmem 3dmar06 3dmark06CPU
3.0 333 9 1066 650 972 13907 4784
3.0 333 9 1066 794 1074 15093 4867
3.23 460 7 920 650 972 4908!!! 5090

Basically i seem to be getting about a quarter of what I would expect. I cannot understand what can be happening to the graphics card that is keeping its frame rate right down when the FSB and CPU are rock solid. The CPU parts of the test seem consistent with the increase in FSB and CPU speed but the rest do not. Has anyone any ideas? The graphics card is running at stock speeds in theory and can only benefit I would have thought from the high FSB allowing 1:1 access to memeory at 920MHz. I guess there must be some other aspect of this speed FSB that affects the graphics card but I cannot understand what. I have not set any particular voltages other than Vcore as it was not necessary to do so for a stable system. Is it possible that an increase in voltage on some other parts of the hardware is required in order to drive the graphics correctly?

To get this overclock the settings are as below.

Vcore 1.3875
DRAM voltage 2.2v
CPU voltage damper Enabled
FSB 460
Multiplier x7
DRAM timings 5-5-5-16
The rest are on Auto generally.

Spec of machine is
P5KC Bios 0903
Q6600
Nvidia 8800GTS 512
Zalman CNPS 9700NT
Crucial Balistic 8500 5-5-5-15
Coolermaster extreme power duo 650W

Help would be greatly appreciated.

Thanks
 
Hi,
Have got my Q6600 running at 460 FSb @ x7 to give 3.24MHz and it is rock solid on Orthos for 12 hours. I then started to run some graphics benchmarks to start on the GPU overclocking but have come across a very strange effect where the results are as below:

CPU FSB Mult Memory GPUBus GPUmem 3dmar06 3dmark06CPU
3.0 333 9 1066 650 972 13907 4784
3.0 333 9 1066 794 1074 15093 4867
3.23 460 7 920 650 972 4908!!! 5090

Basically i seem to be getting about a quarter of what I would expect. I cannot understand what can be happening to the graphics card that is keeping its frame rate right down when the FSB and CPU are rock solid. The CPU parts of the test seem consistent with the increase in FSB and CPU speed but the rest do not. Has anyone any ideas? The graphics card is running at stock speeds in theory and can only benefit I would have thought from the high FSB allowing 1:1 access to memeory at 920MHz. I guess there must be some other aspect of this speed FSB that affects the graphics card but I cannot understand what. I have not set any particular voltages other than Vcore as it was not necessary to do so for a stable system. Is it possible that an increase in voltage on some other parts of the hardware is required in order to drive the graphics correctly?

To get this overclock the settings are as below.

Vcore 1.3875
DRAM voltage 2.2v
CPU voltage damper Enabled
FSB 460
Multiplier x7
DRAM timings 5-5-5-16
The rest are on Auto generally.

Spec of machine is
P5KC Bios 0903
Q6600
Nvidia 8800GTS 512
Zalman CNPS 9700NT
Crucial Balistic 8500 5-5-5-15
Coolermaster extreme power duo 650W

Help would be greatly appreciated.

Thanks

Try locking your PCI-E bus speed to 100Mhz from auto. Maybe a slight increase in NB voltage?
 
Thanks for the response.
I have tried it at 100 but with no difference. I haven't played with the NB voltage as it is stable on Auto but is it possible that it can be stable for CPU/memeory tests using orthos but not for other parts of the system such as the PCI_E?
 
Quite possibly, the PCI-E demands another 8GB/s bandwidth via its bus that connects directly to the P35 northbridge. It could well be dropping packets/or missing frames to keep the system stable, trying to keep the ram and CPU happy at that rather high FSB.

I’m certainly no tech head but I’ve seen people struggle to get the best 3d scores when they’re running high FSB.
 
Back
Top Bottom