Uncomfortable with temps

OK, so I have a 7950 OC'd to GPU:1100MHz, and MEM:1450MHz, Overvolted +3%.
I have a Swiftech Komodo 7950 Full Cover block, Black Ice GT Stealth 360mm rad, with 3 Cougar Vortex fans in push.
In FurMark, I'm getting ~53C max after 45min. My room is 19-20C. I was getting upwards of 70C on air cooling, which was the reason I decided to go to WC.
I am kind of uncomfortable with these temperatures, and i expected lower numbers.
It instantly jumps up from an idle of 25C up to about 38C before the test even starts (displays min of 38C). Is this supposed to happen?
Did I put too much/little thermal compound on it (AS5)?

Directly after i first installed the waterblock, and everything was running fine, i tested my GPU, and it got so hot it shut down (over 95C) after about 10 seconds of FurMark. After about a day it came down a little, but would still get to the point of shutting down. Does the AS5 just need the "break-in" that I've heard it needs? I've ran FurMark for probably a total of 5 hours after i applied it.

My CPU gets up to 35C from idle of 27C when running P95 for about 30min (separate test from GPU), so I'm not worried about that at all.

I just need to know if this is normal, or if i did something wrong installing it, or if it's my card.

Thanks for the help!

I dont see any problems at all...does that rad have the CPU on it as well?

EarthDog said:

I dont see any problems at all...does that rad have the CPU on it as well?

Click to expand...

I just thought I could get better cooling than that. from 25C up to 53C?
My friend gets 49C with a GTX570 on an aftermarket air cooler.

Yes the CPU is in the same loop.

You can try remounting it, but, with that amount of rad and fans, I wouldnt expect much different.

Is there any way to add three additional fans in pull for P/P? I would imagine that would help greatly.

TheHutchTTU said:

Is there any way to add three additional fans in pull for P/P? I would imagine that would help greatly.

Click to expand...

I could switch to Cougar Turbine fans, but they are slightly worse, so it would probably be the same. I measured the air with my hand (I know, terrible way of measuring air flow/pressure) of the Turbine fans and the air of the Vortex fans with the rad in the way, and I don't think it would add any cooling capacity if I put the Turbine fans pulling the air from the Vortex fans (I hope I made sense).

Also, the temps jump up about 13C before the test even starts. I have no idea if that is normal, or if my card is weird, or if my block is weird, or if I installed it weirdly.

your temps are fine GPU's can handle 53c with no problems that actually a pretty cool running gpu.

for a 30mm thick radiator, you're probably getting about the performance you should expect. however, that radiator has 30 fpi (high) and might actually cool a little better if you add secondary fans for pull. (even if they're not the same fans) either way, your temps seem fine. you sure you have enough air flow through the case?

53c is absolutely nothing for a modern GPU. Even Keplers and their twitchy de-boost system don't do anything till 75c.
AMD stuff is happy well into the 80s, those temps are GREAT.

You really can't compare between manufacturers or generations, and especially not both.
Temp sensor location plays a huge part in the displayed temps, that varies by generation and manufacturer (AMD vs Nvidia), just like with CPUs.

Can you give us more specs on your whole water cooling loop as I only see the specs of your PC in your sig. I would though move all the settings back to stock and run some Prime95 and Furmark and see what temps you get. That CPU reminds me of my 130w i7-950 while yours is 125w. Our CPUs run hotter stock than the ones that have been out the last couple years (SB/IB) so you OCing that and the GPU too with only a 120.3 rad will give you temps like that. Your temps are fine.

I don't know what happened with the whole 95c on the GPU once the WB was installed. Maybe the pump was off is my only guess and did you reinstall the block after that? Also, AS5 is an old paste that was great than but theres new and better performing pastes. I myself use the MX-4 which is great.

Do you have this radiator?

http://www.frozencpu.com/products/4283/ex-rad-88/Black_Ice_GT_Stealth_360_Radiator_-_Black.html

If so that rad really needs good fans. Easily 2000+ to perform. Back in the day when this rad was new (6-7 years ago, it's really outdated now) folks really tossed some fans on it to make it work. It's 30 FPI, it takes a LOT of pressure to get the air through to dissipate the heat.

Your underraded to start with, but with the right fans that rad can pull more heat than the more modern low FPI rads.

With the right fans. Think about something from this list I'd say, that what everyone was using back in the day when that rad was close to the best out there. It will make that rad sing, it will be cooler and you won't miss the sound of it cooling like it needs to be cooled.

http://www.frozencpu.com/cat/l3/g36/c15/s562/list/p1/Fans-12_Volt_Fans-120mm_x_38mm_Fans-Page1.html

Not sure I see how he's under-radded if his GPU load temps are 53c.

The mechanics of raddage show with his overclocks he could be short on raddage. Maybe I'm missing something, it's kinda hard to read. Has the OP run P95 small ffts and Heaven or Fur at the same time for 30 minutes? That will tell us his capability with his current setup.

Lastly, not sure what he's using for temp measurement or his room temps.

There's theory, and there's reality.
Reality is showing 53c on the GPU.

It's only a 200w tdp, CPU is 125w. An OC to 3.8GHz is going to add squat for heat if it's a C3.


As a sidenote to the OP: Temps will always jump at the start of the test, that's totally normal. You're seeing the thermal conductivity (or lack thereof) between where the sensor is on the silicon die and the actual working fluid (be it air or water).

Yea, your right Bob. Still don't know if he's fun concurrent tests though. Ahh well, his temps really are okay, but could be better..

Conumdrum said:

The mechanics of raddage show with his overclocks he could be short on raddage. Maybe I'm missing something, it's kinda hard to read. Has the OP run P95 small ffts and Heaven or Fur at the same time for 30 minutes? That will tell us his capability with his current setup.

Lastly, not sure what he's using for temp measurement or his room temps.

Click to expand...

Tested last night, P95 small fft's and Fur same time for 28 minutes, and i get 32C CPU and 56C GPU, 71F(21.66C) room.
I am using a cheapy thermometer next to my monitor for my room temps, and i'm using SpeedFan for my hardware temps.

I do have this: http://www.frozencpu.com/products/1...xoS_UT60_Full_Copper_Dual_120mm_Radiator.html laying around. Would it help much if I put this in the loop? The fin density is roughly 7FPI (estimate) but I didn't know that when I bought it.

I'm going to test the Black Ice in push/pull tonight and see my temps. I'll post all of my configurations when i finish that. I already tested just push.

Speedfan? Obsolete... try Coretemp and see if those temps for the CPU are still 32C.

EarthDog said:

Speedfan? Obsolete... try Coretemp and see if those temps for the CPU are still 32C.

Click to expand...

alright, i'll look for that when i get home.

Alright, so i did some more tests.

Push - Cougar Vortex
P95 small FFT's:
Idle: 24C
Max: 30C
FurMark 1440x900 fullscreen 8x aa:
Idle: 26C
Max: 52C
P95 and FurMark:
Idle: 24C/26C
Max: 32C/56C

Push/pull - Cougar Vortex/Cougar Turbine
P95 small FFT's:
Idle: 23C
Max: 29C
FurMark 1440x900 fullscreen 8x aa:
Idle: 26C
Max: 49C
P95 and FurMark:
Idle: 23C/26C
Max: 31C/52C

Overall: not much temperature gains from this setup. 1C on CPU, 3C on GPU, and combined 5C in the combined test. I know my temperatures are 'safe', but i just don't feel comfortable with my GPU getting that warm. (I am OCD about heat, which is why i went to watercooling).
I will try adding my 120x2 60mm thick radiator with 9FPI (just measured), and test that. I just need to order some more fittings.

Side note: I looked into CoreTemp, and it only reads the core temperature of the CPU (obviously). Speedfan gets the same readings on the core, as well as displaying the GPU temp, North Bridge temp, and since i opted to not get a fan controller, I can have it automatically turn my fans up at a certain temperature, or i can manually control them. -I just see no advantage with CoreTemp-

Edit: next, I'm gonna test it like this.

Turbine/Turbine/Vortex
======radiator======
Turbine/Turbine/Vortex

So that the same fans are push/pull with the same fans. I think that should help.