E8400+XFX 780i

[brendano257]

When I first got this mobo/cpu all I did was OC it to 3.6 and leave it there on pretty much stock voltages. I've finally gotten it stable in Prime for w 1.275 core and the other voltages raised too. The problem is this is what it took to make it stable, and now the temps go from 40 idle to 75 on load, so far, And I've only been running Prime for 10 minutes

Any suggestions are much appreciated

EDIT: Prime failed a minute after this post, I don't think I'll be continuing unless there is a way to avoid the heat because unstable at 75 C is not ok with me.

BTW: 8800GT, Corsair 2x1gb XMS2, and CM Hyper TX2

 

[ibp_john]

What kind of thermal paste are you using?

From the couple of reviews I've read it looks like that TX2 isn't that great of a cooler.
http://www.bit-tech.net/hardware/2007/11/19/cooler_master_hyper_tx2_and_212/6

 

[aevm]

Looking at newegg reviews of that cooler, I'd say it should do much better. There are people claiming 45 degrees at 3.6GHz with that CPU and cooler.

Maybe your fan speed is too low, or you need to reapply the paste.

 

[brendano257]

I reseated (with AS5 just like before) it and remove part of the plastic on it....Right now it's at 3.6 and 1.225 Vcore and idling at 38 C, haven't seen load yet. I'm going to try it at 3.825GHz again in a bit. Anyway this is what I removed:

 

[alfrido]

omg omg omg hahahhaha
Isn't it upside down? Shouldn't it be the other side up? Cause the air is transported downwards. Wait wait that isn't heat that's being transferred downwards. It's normal air and maybe the fan on back of the case sucking the rest out.

 

[brendano257]

omg omg omg hahahhaha
Isn't it upside down? Shouldn't it be the other side up? Cause the air is transported downwards. Wait wait that isn't heat that's being transferred downwards. It's normal air and maybe the fan on back of the case sucking the rest out.

Come back when your not high, but anyway I primed it for 30 min without errors..... When I tried to open up another in place fft window, the new one had an error instantly, when I opened one blend ontop of the one in place fft, it ran, then several minutes later the in-place one failed. So I'm back at 3.6, I'll try 3.8 and put the voltage up a bit more than last time and prime it over night.

 

[iluvgillgill]

some one is on pot here!lol

that thingy at the back is for guiding air to get blow downward to the mobo's component so it gets cooled at the same time. you see that shoud is pointing at the MOSFET heatsink. maybe you are upto a point where you need to adjust the VTT and GTL ref if your BIOS have it.

but for now prime it and see what you get.

 

[brendano257]

The odd thing was that I had raised the voltages (still within safe limits) on everything else to get it stable. But after I reset the heatsink and retried the 3.8GHZ OC I didn't add any voltage to anything except the core, and got a longer time in P95 w/o errors.... I think my bios has those, any suggestions or cautions when changing them? Just so everyone knows I cut my thumb on the fan while checking the airflow

BTW does it look to anyone else like the mobo in the ^^^picture^^^ look slightly bent?

 

[iluvgillgill]

yes the mobo is bent. but thats a good mount because the pressure is big enough so contact should be good.

i never manage to cut my fingers on those cpu fans. as they are not that sharp.

leave those value for now. you dont need them yet i think.

 

[brendano257]

Alright right now I'm at this for voltages:
CPU Core:1.25v
CPU FSB:1.3v
Mem:1.85v
nForce SPP:1.35v
nForce MCP: 1.5v
HT SPP<->MCP 1.2V
GTLVREF lanes 0-3 (all at auto, no additions)

It's not my mobo, it's just the only side pic I found on google.....I'm going to Prime what it is now, and see if it fails before I feel like going to bed, if it does I'll up the core, monitor temps for a few mins, then leave it for the night...... Which P95 test should I use? And should I run two at a time or just one? Two "blend" tests do not put my CPU usage at 100%, and do not crash within a few minutes....Two "in place ftt's" one fails, and the other keeps on going, retrying the one that fails fails very quickly again...... One small, or in-place FFT does 50% CPU usage.

 

[iluvgillgill]

just run prime 25.6 and small ffts. 24hours is my recommended time.

 

[brendano257]

Just started running 25.6 small fft's... One fails almost immediately...upping core voltage a bit.

 

[iluvgillgill]

lol so what you been using before?omg!you done all that and its not "tested" to stable!

 

[brendano257]

CPU Core:1.25v---> 1.275
CPU FSB:1.3v
Mem:1.85v
nForce SPP:1.35v--->1.4v
nForce MCP: 1.5v--->1.525v
HT SPP<->MCP 1.2V
GTLVREF lanes 0-3 (all at auto, no additions)

One of the work threads failed at two min after the 1.275v and so I upped the SPP and MCP and it failed at 2 min again.......

I've been using P95 24.14..... it needed a comand line switch -A1 to run multiple windows......

 

[brendano257]

Same 4 dumb pronged fasteners
as the defective Intel HSF:

http://www.supremelaw.org/systems/heatsinks/warning.htm


Sincerely yours,
/s/ Paul Andrew Mitchell, Inventor and
Webmaster, Supreme Law Library
http://www.supremelaw.org/

All Rights Reserved without Prejudice

Erm... fixed that atleast 5 hrs ago today........I reseated it with new AS5 again.

 

[brendano257]

It's not my mobo, it's just the only side pic I found on google.....

Please read the whole thread first....

GILL THE LOGIC DEFEATS ME!!@#$#@#

CPU Core:1.25v---> 1.275
CPU FSB:1.3v
Mem:1.85v
nForce SPP:1.35v--->1.4v
nForce MCP: 1.5v--->1.525v
HT SPP<->MCP 1.2V--->1.25v
GTLVREF lanes 0-3 (all at auto, no additions)

Upped the HT by .05v and one instance P95 failed after one minute instead of 2..... Maybe I should just start cutting myself with fans.... Can it just not take the full 100% load or something, or could it be a CPU option that needs to be turned on/off I head most things should be off, the only CPU option in BIOS I have on is Virtualization.....

With those voltages it was stable in CS:S for 30 min+ ............ If it can stable Crysis 1 hr+ and no odd system freezes/crashes I might be able to live with myself/go back to 3.6 and be content.....

 

[iluvgillgill]

what FSB you runnong on now?

i dont think those voltage increase apart from vcore is needed after 3.6Ghz. can you get me a screenshot of CPU-Z?

 

[brendano257]

When I woke up I went back to 3.6, the stability problems are with 3.8GHz...

Edit: Really odd, booted up again with 3.825GHz and it Primed past for 6 min, then I stopped it so I could do other things heres the screenshot CPU-Z+Realtemp+Prime.

During Prime the temp only hit 58C

One work thread failed after 11 min......Grrrrrrr

 

[iluvgillgill]

the ss above what bios setting are you on?

 

[brendano257]

I'm using the newer one P05 (flashed it yesterday). BTW I've been playing Garry's Mod for probably over three hours now, and stable :S

I'm also running the ram at latencies of 6 and 20 instead of 5 and 18. And of course syncronously FSB of 1700 and ram at 850.

 

[iluvgillgill]

i think you can go higher then 4Ghz by the looks of it.

 

[brendano257]

i think you can go higher then 4Ghz by the looks of it.

I tried 1800 FSB for the heck of it and froze during windows boot-up sound.....Is there anything I can do to make it stable now?

 

[iluvgillgill]

have you up the NB voltage yet?

 

[brendano257]

Yep, up from 1.5 to 1.525, I've upped it further bc I'm attemping to run Prime on 4.05GHz, instant Prime failure so far.

CPU Core:1.35v
CPU FSB:1.3v
Mem:2.0v
nForce SPP:1.45v
nForce MCP: 1.6v
HT SPP<->MCP: 1.3v
GTLVREF lanes 0-3 (all at auto, no additions)

That's what I'm at now with 450x9(4.05Ghz), One thread of P95 fails after less than a minute. Yet the other keeps on going.

 

[iluvgillgill]

you will need more then 1.3V on the FSB.
CPU Core:1.35v
CPU FSB:1.5v?
Mem:2.1v
nForce SPP:1.55v
nForce MCP: 1.6v
HT SPP<->MCP: 1.3v+1more notch?

dont think loose your memory timing to 6 can allow you speed up without voltage increase. you will still need some ram voltage bump to get it up.