GA-EX58 EXTREME OVERLOCKING

[techno2k8]

well here is a few pics of mine

1st -  


2nd -


3rd -


i will be messing with the settings and see what happens but i can run the intel burn test and that is fine it doesnt crash but will try it for longer now and see what happens.

i have the loop as follows

res - pump
pump - dual rad
dual rad - cpu
cpu- single rad
single rad - gpu
gpu - res

[Chris_Waddle]

Ok, well I did some more investigating yesterday and have decided that after reading this article to turn off Hyper Threading. I may well turn it back on in the future if it improves, but for now, I don't see any reason to leave it on. Infact, after doing this, my games (GTA4 especailly) ran far smoother.

I am still running 4.2ghz as a 24/7 overclock, but with HT off and LLC enabled, it's steady as a rock at 1.262v.

If I upped the Vcore to 1.30v it overclocked to 4.4ghz and ran Prime95 for an hour at least. I may well go back to this and tweak it to get better stability. With HT disabled the temps have really dropped. At the moment it idles around 36deg and maxes out at around 57deg.

I didn't like the idea of disabling HT, but for day to day running, it does appear to make my PC more responsive.

[techno2k8]

i tried those settings provided by all and they didn't work pc wouldn't even post, let alone boot.

it still wont do more than 5 mins of prime95, or occt but i can play games do video converting and host game servers no problem but as soon as i try to stress test it, it falls over i can also benchmark it aswell.

i am however going to try prime95 at stock speed with no overclock and if it fails then, then there is something wrong with either the board or cpu or ram.

[Chris_Waddle]

Sorry I can't be of more help. If stock speeds do fail, try running with 1 stick of ram (swapping it, if it still fails).

Have you tried running Memtest? If it is your ram, this might help identify it.

Going back to your temperature problems, I have noticed that even though you are watercooled you don't appear to have any fans on the radiators. If this is the case, you really should fit some. The rads do need air moving through them to cool your system. Without fans, your water temps will be much higher.

[techno2k8]

Sorry I can't be of more help. If stock speeds do fail, try running with 1 stick of ram (swapping it, if it still fails).

Have you tried running Memtest? If it is your ram, this might help identify it.

Going back to your temperature problems, I have noticed that even though you are watercooled you don't appear to have any fans on the radiators. If this is the case, you really should fit some. The rads do need air moving through them to cool your system. Without fans, your water temps will be much higher.

i have got fans on the rad's it's just that you can't see them as they are pulling air through the rad's.
i will post some pic's later of the fans etc.

i looked at getting the temps down but i noticed that with a vcore of 1.25v my temps were only 5c lower per core than with 1.35v, and with lower voltage it would not even run any games etc.

even if i put the overclock down to 3.0ghz or even 2.7ghz it won't remain prime stable, and the temps are still around the same.
if i put more voltage through it then it remains stable for longer but the temps shoot up to nearly 90c per core which i will not run it at, so at the moment it runs at 4.0ghz fine for gaming and everything else i do but it is not prime stable so i might just leave it if it passes the 10min mark at stock speed.

also i have noticed that my patriot ram kit which is 1600mhz 8-8-8-24 1 or 2t cant quite remember, when i set the bios to standard it drops my ram down to 1066mhz with timings of 30-30-30-60, then when i put it to 1600mhz it gives timing's of 10-11-11-34 so i have to put the timings in myself for the ram as 8-8-8-24-2 could this mean the board is faulty or just that it doesn't like to pick up the speed of the ram itself?

[Chris_Waddle]

If temps are doing that then I'd have to say that it sounds like a problem with the cpu block.

Are you are running Real Temp? If so, go into settings and activate the gpu reading. If this is saying that the GPU is normal (your's looks like a GTX280 H20C; mine runs around 40 deg) then I would really take a look at the CPU block.

I'm not sure how long you've been running green liquid through it, I used to use this stuff, but found on one re-build that the CPU block had totally blocked up with green gunk. I took me quite a while of soaking it in white vinegar to get it all out.

Your temps really shouldn't be that high. If the gpu is running hot then maybe you have a blockage somewhere or an air lock.

Can't think of much more than cleaning up both the cpu and water block. I would recommend lapping the waterblock. My block had a slight convex surface to it originally. This meant I got great contact in the middle, but poor contact at the edges. After lapping it the contact was much improved between the CPU and waterblock. My temps dropped around 8 degrees after doing this.

If I can think of any other things to look for, I'll let you know.

[techno2k8]

what i have is a swiftech apex kit which contains:

dual rad and pump and res.

i have no other waterblocks other than the xspc delta v3 cpu bloc, which has had green fluid running through it ever since i got it last year, minus a few times where i have had to drain it etc due to adding and taking cards out.

to monitor the temps i use coretemp and also everest ultimate, the temps are lower by around 10c on this loop that when i was using the corsair h50 unit.

i am using artic silver mx2 paste.

i have a dual rad and a single rad both with fans on them pulling air through the rad rather than pushing air through as i was told that pulling was better, rather than pushing.

the graphics card is a bfg gtx280 that was originally air cooled but i got the waterblock for it and watercooled it myself, the idle temp is 37c.

as i am typing this my temp's read like this:

Motherboard: 38c
Cpu: 28c
Cpu Core 1: 40c
Cpu Core 2: 36c
Cpu Core 3: 40c
Cpu Core 4: 36c
North Bridge: 47c
Gpu: 37c
Gpu Ambient: 34c
Gpu Memory: 30c
Gpu VRM: 34c
Samsung hdd: 21c
WD Velociraptor: 26c

here is a screenshot when idle
 http://img237.imageshack.us/img237/1484/tempsetc.png

heres when it's running intel burn test
http://img237.imageshack.us/img237/4903/tempsetc2.png

let me know what you think cheers

[techno2k8]

Just a update as here in the uk it is moving in to winter and i thought my problems with my overclock were heat related i started prime 95 64bit running and it didn't crash after 40 mins of running without hyperthreading turned on, so i turned the hyperthreading back on and it has been running for 40 mins again without crashing.

i think my issue was that the motherboard was overheating, as all i have done is put a desktop fan pointing in to the computer and blowing air over the mobo, and the temps of my cpu each core is hitting a max of 70c with ht turned on and a max of 66c with ht turned off.

so i'm now looking at watercooling my mobo to see if that drops the temps even more.

temps were without fan:
core's 70c ht on 66c ht off
northbridge ht on 55c ht off 50c

temps with fan:
core's 70c ht on 66c ht off
northbridge 44c ht on 38c ht off

so i'm fairly confident it is heat related, when i get it watercooled i may look at seeing if the chip will go any higher than 4.0ghz but not sure yet.

[techno2k8]

well i went to the pub at just after 6pm and left prime 95 running and i just got back and it was fine no crash or reboot, so it was running for just over 5 hours so i would say it's stable as it wouldn't even run for more than 5 mins before i added the fan blowing on to the mobo.