Did you change the vcore in BIOS or is it set to auto? I bet it is at auto, and if so it upped the voltage by about 6 steps, hence the large jump in temps.
I didnt have to up the vcore until I hit 4Ghz on my stock 3.3ghz chip to get it to stay stable… until I bumped the voltage to get past 4ghz, the temps only increased about 4-5C. once I bumped the volts it jumped up about 10c over stock.
It’s set to auto. I didn’t change the memory timings either. The computer theoretically should be able to maintain a 1:1 ratio itself.
I have my 9550 at 3.4. Didn’t have to change the voltage, and under load it only went up a few degrees from stock. I’d say it’s safe to take it up to there without anything besides a better cooler (which you have).
I wanna try and break 4.0 myself.
What are your temps? Did you change the VCore like KrazyKid suggests? Or did you set it to auto in the Bios?
u can do it, i have a weaker setup (in some ways) and can push 18k without overclocks
nice water cooling job… looks clean !
nvm i see u already did, nice man !
Thanks man!
Yeah last night I broke 20K! Thats a mile stone for sure, with any single GPU machine.
I know this all started when I water cooled my pc so just to one up ya I am building another one. Here is the pic of my case
Kenmore White 5.0 cu. ft. Manual Defrost Chest Freezer
Liquid cooled my ass I’m freezing my pc this time.
Better get out the puty and insolate that bisch.
If you have the vcore set to auto, I hear it usually over volts it a bit.
Ilya, I would set yours back to default clocks on auto Vcore. then see what it runs under load for a vcore.
Then manually set the vcore to that in bios and bump your clocks up from there. once you see it get unstable, not post, hang what ever, bump the vcore on click at a time.
This thread srsly hurts my head, and seems wayy over the top! But pretty cool nonetheless :thumbup. Porn has to be pretty epic on that comp :lol
Yeah look at Vcore like fuel pressure, and the CPU speed X.XXGhz as PSI of boost.
More boost, you need more fuel or your lean
Same shit, more ghz, needs more Vcore volts to run stable.
But yeah higher FPS = more FAPS :rofl
Modified.
Two things.
Since your processor paths have a fixed resistance, increasing the voltage to the core will increase the power and therefore the heat of your processor. If you are having thermal issues with your processor you should reduce the vcore to the minimum level to which it still performs. Note that an undervoltage condition my also increase temps if the amperage is not controlled.
Why is the radiator inside the case? this is really hurting your cooling performance. Having the radiator inside will be contributing heat through convection to the system. You did all this work to remove this heat from the components, but the radiator will be putting it right back into the air temperature! I’d say your next project would be to move the internal radiator to the outside like you did with the other one. You will see much lower temperatures.
Nice work! :thumbup
What? I am so confused…it looks to me like the rad is mounted to the external rear of the case via four green anodized standoffs??? :lol
Right you see that cute little thing at the bottom of the case with the green coolant lines going to it???
:ponder
Good point, but he should be fine for a while. If all the critical components have a water block on them than the air temp shouldnt effect them to much. But, I’d watch the memory modules, maybe include those in your water loop.
:lol Yeah you’re right, I only picked up on the one stuck to the outside of the case. Fail on my part.
Case in point, I now concur with your statement.
yeah , u can have all the water blocks u want, i still think that is unnecessary as well, i didnt look at the pics too much as i was in a hurry and just checked his 3d mark score…
radiator in the case is just kinda silly
what are the temps? obv more voltage means more heat… lol
ya im way to lazy to read the whole thread for temps and clocks and timings
OK heres the deal.
I will start with the flow of the system. The pump pushes from the res. to the inlet on the big external radiator on the back of the case. then that goes out of the rad into the CPU block. Cpu block heats the water a little then dumps it into the first mosfet block. that block jumps to the second mosfet block, then into the North&South bridge chipset block. At this point the water was heated up by 4 water blocks.
Then that water hits the snigle 120 rad inside the case, to get cooled back down (even if it is a little bit) PRIOR to passing the water to the GPU block, then out of the GPU block back to the res.
I put that in the middle to cool the water back down so the gpu would be as cool as it could. If I didnt, the water would just keep increasing after it exited each block in order and by the time it hit the GPU (final block) it would be much higher in temp. If I just went with a bigger single radiator on the back, I would have to run CPU>GPU>NB/SB>MOSFET1>MOSFET2… and there would be alot of criscrossing water lines and it wouldnt be as clean lookin.
The ambient temps in the case used to be about 6C+ higher than ambient temps in the room when I was fully air cooled (with the ultra kaze full bore for the intake and exhaust on the case Now that the only heat producing air items are the HD (30C full load on board, and the case of the hd probes out to be 28C). The RAM, full load probe stuck inbetween the chip gaps on the stick and the heat sink 30C) and the PSU but that pulls case air and pushes it right out the case so the case of the psu doesnt even get warm to the touch. Ambient temps in the case as you see it now is only 1-3C higher than ambient. so the air hitting the small rad in the case isnt any higher than the one hanging off the back… even the one hanging off the back gets some of the case exhaust air thrown at it anyway.
Tonight I will post up a full load temp readout from everest, and post all the probe temps too from the case.
