Printable View
Thanks for the info sp3n, but it didn't help.
Anyway I did a bit of of lapping on my Scythe Infinity...
...my temperatues have dropped by 2 degrees!!
If you included the trip to halfords for the fine sandpaper, thats over 2 hours work.
Oh well.
At 1.45V, TAT at 100% load is topping out at around 70c (coretemp reads 1-2c higher). Should I lower the voltage, or is it fine to leave it at that voltage with those temp? with orthos temps usually stays around the 59c mark.
After reading up on temps, anything under 65C on TAT when under load is acceptable. While TAT was specifically developed for mobile CPU's, it also works correctly with desktop versions of C2D's, although the E4300 may be the exception (although these should still be pretty reliable)
As Clunk mentions, it's worth checking the application of thermal compound - I'm using an unlapped Tuniq Tower 120 and applied the AS5 thermal compound according to the AS guide. Compared to the stock Intel HSF, my temps lowered by at least 10C. However, due to my recent system problems I had to remove the Tuniq and found that there was too much TC for my liking. I'm sure with the correct amount I may have improved by up to 15C over the stock Intel HSF (so it's definitely worth checking).
For anyone that's interested, I'll be posting on my experiences of a noob bulding a new PC from bare components, including any wisdom gained from my selections and toils :)
sswats - C2D's are supposedly resilient to higher voltages, although you do run the risk of reducing the lifespan of the CPU. As mentioned in my previous post, it would be ideal to have TAT temps of 65C or lower, although 70C is not unacceptable.
The best advice I could give would be to weigh up the advantages of a few extra fractions of a MHz (which can offer reasonable speed improvements in CPU intensive apps) against a slightly lower speed with potentially safer operating parameters (ie. a slightly longer lifespan).
As I've said though, CPU's are generally hardy beasts that don't give up the ghost easily ;)
Ill give the TC stuff a shot. Dont have any around at the moment though so that will be on my next order no doubt.
I tried to run Orthos again. It has got passed the 12min mark and is still running.
Temps are as follows:
TAT - Mainly around 62c-63c but it has jumped to 65c for a small amount of time
Core Temp & Speedfan - Lowest has been 49c highest has been 52c
Do you think these temps sound accecptable for the overclock it has done so far? Or should I just throw everthing back to stock and wait for the TC?
EDIT - also the Vcore is at 1.32
EDIT again - The machine restarted itself after running for 28mins, the error log doesnt say why or anything though. So something isnt right somewhere. Any ideas guys?
Temps seem ok, the restarting itself isnt though.
Did you enter the ram timings and subtimings manually?
Did you enter the correct ram voltage for your sticks (Vdimm).
Did you leave the cpu voltage on auto, or did you set it manually to 1.32?
Cheers for the reply ClunkQuote:
Originally Posted by Clunk
I cant see your ram needing 2.5v, I think 2.0-2.1 should be ok for those, but you can always double check that in the corsair forum.
Auto is right for the cpu for now.
Ok ill put the Vdimm to 2.0 volts (as it doesnt do 2.01) and see how it goes. I looked at the forums and it seemed to be mix, some did 2.3v and others did 1.9v. Is there a threshold for the modules?
Edit - Orthos ran for 32 mins this time. Then halted with the following error. Im starting to get the feeling this overclock isnt going to go well :(
Test Output Results
===============
2:Test 1, 560000 Lucas-Lehmer iterations of M212991 using 10K FFT length.
2:FATAL ERROR: Resulting sum was 2263170623160884, expected: 2263115432831123
2:Hardware failure detected, consult stress.txt file.
2:Torture Test ran 32 minutes 30 seconds - 1 errors, 0 warnings.
2:Execution halted.
1:Torture Test ran 32 minutes 30 seconds - 0 errors, 0 warnings.
1:Execution halted.
What voltage does it actually say on the sticks?
I havent actually had a look on the sticks. Ill open up the case and have a look.
Edit - on the sticks it doest actually have a voltage. it just says the timings which are 5-5-5-12, 800mhz. Should i set that in the bios rather than 5-5-5-15?
Thats odd, it usually has the rated voltage on a sticker. But yes, enter the timings as it says on the sticks. I would also set the voltage to 2.1v for now (not 2.01v).
Edit: May be some usefulness in this thread.
Can you just confirm that you havent set the ram to 800mhz?
You started off with 266/533 right?
Hey Clunk, timings were 233/533 as it said in your guide. Right now I have it at 300/600. Ram isnt at 800.
Are you sure about the 2.1v?
http://www.houseofhelp.com/forums/sh...ad.php?t=52351
(search for "Testing parameters for Twinx/XMS2-6400")
It says the RAM should be 1.9v, just dont want to be pushing too much through them.
If they are rated at 1.9v @ 800mhz then 1.9v it is :)
Have you got all the other settings correct?
In the ram subtimings, change write to precharge delay from 10 to 15.
Instead of 300, try 320 or somewhere round there (up to 333 say) and see if its stable.
Ok so no the problems start. I changed the timings on the ram to what it said on the sticker. I ran orthos it was at 8mins when I last checked it. Came back to the machine and the screen was blank. Had to do a hard reset.
Now everything spins up fine, except there is still nothing on the screen and no single post beep. Any suggestions?
Edit - Nevermind, read your handy dandy troubleshooting guide and got it up again. Just had to take out the power for a bit.
Ill have to come back to this as need to go and take my brother out and then take care of some other stuff. I set the memory timing back to 5-5-5-15 even though the sticker says 5-5-5-12 I had somewhat better results with that.
Cheers for your help so far though Clunk, dont worry Ill be back ;)