I was pondering...

I was pondering a question last night (could have been the flavored brandy...it's brandy season).

The question was: "Is the cooling I am using optimal for my setup".

My rig is in my sig, and my "current state" was to have:
1. Intake fans on front of the case
2. Exhaust fan on back of case
3. Radiator mounted on top of case, with fans taking air from inside the case, blowing through CPU radiator, and exhausting outside the case
4. 2 GTX 970s in SLI, dumping their heat into the case

I read a lot of the forums on the best way to setup a water cooling system. So, I wanted to see the effect of the heat-load of the graphics cards on CPU cooling performance.

I devised a set of experiments to keep the CPU power draw constant, and varied the GPU power. I did a "run" for 30 minutes, taking data at 5 minute intervals. At the start of each run, I brought the coolant back to "nominal temperature" by running the AIO cooling fans at maximum.

I used Seti@Home and Heaven 4.0 to load the system.

Experiment Run 1:
- Seti@Home at 90% CPU, turn off GPU
- CPU power at 153 W, GPU power at 95 W (idle)

Experiment Run 2:
- Seti@Home at 90%, both GPUs running Seti@Home
- CPU power at 155 W, GPU power at 171 W

Experiment Run 3:
- Seti@Home at 60%, Heaven 4.0 running continuously
- CPU Power at 152 W, GPU power at 290 W

---------------------

Setup #1:
- Intake fans on front of case
- Exhaust fan on back of case
- AIO radiator on top of case, fans pulling air from inside the case through radiator as exhaust

Here is a graph of CPU temp (notice it goes up as the GPU power increases):



Here is a graph of the AIO coolant temp (goes WAY up as GPU power increases):



---------------------

Setup #2:
- AIO radiator in front of case, fans pulling outside air into case
- Exhaust fans on top of case
- Exhaust fan on back of case

The results of this shocked me. Here is a graph of CPU temp:



A MASSIVE drop in CPU temp.

Here is the AIO coolant temp:



---------------------

Conclusions:
1. CPU temperature showed an 8 to 10 °C temperature drop with the radiator fans drawing in outside air into the case
2. AIO coolant temperature was 3 to 6 °C cooler with the radiator fans drawing in outside air into the case
3. Drawing air into the case through the radiator is superior

I was concerned about taking this approach due to the AIO blowing heated air into the case. The net effect was that the motherboard temperature increased by about 5 °C, GPU temperatures increased by about 3 °C, and GPU fan speed increased by about 10%.

Interesting read JrClocker, I wouldn't have thought the temps would drop that much. Before getting to the bottom of the post I was thinking, "but what did the Gpu temps go up" thinking it would be more then the 3 c it did. Keep drinking the Brandy maybe it will make it go sub ambient.

Mandrake4565 said:

Keep drinking the Brandy maybe it will make it go sub ambient.

Click to expand...

The brandy didn't influence physics...not that kind of drug - hehe.

There were a couple of uncontrolled items in the experiment:

1. Ambient temperature - my house is air conditioned, but my office does not have it's own temperature zone. However, it was pretty close between the two days.
2. In the first part, the radiator was mounted horizontal and above the AIO CPU pump. In the second part, the radiator was mounted vertical, with the center of the radiator "in line" with the AIO CPU pump.

Item (1) would have a couple of degree influence either way.

Not sure about item (2). To rule out this, I would have to run two more test cases:
- Reverse fans in vertical mount to take air from inside the case
- Reverse fans in horizontal mount to take air from outside the case

As I have much better temps with the same gear, I'm too lazy to do that - hehe.

So, while your results may very, I think the conclusion is still valid: cooling the radiator with "outside air" venting into the case is superior.

As an ancillary observation, after running for 12 hours last night looking for ET (i.e. experiment run #2 with CPU power at 155 W and GPU power at 171 W) my office was noticeably cooler this morning!