watercooling :)

[pilotw]

So, after my rig crashed during running 3DMark05 due to GPU temps I decided to buy a watercooling setup.

 

I'm currently running the Evercool WC-202 on only my GPU. I had to go ahead and create a custom bracket to hold the second radiator onto my 120mm fan(since I cutout the 80mm mount bracket)

 

So far I've run 5x all 3DMark05 test with no temps over 40 degrees

 

Once it's stable for a 20x run I'll be happy. But so far, so good. Once this is done, I'm going to run 3DMark 05 on different ram divider and timing settings at the same FSB to see how the different memory settings affect gaming...

 

more to come...

[dracos]

Yeah the Wc-202 is not bad if you rip it apart and just use the parts seperately..

 

I have a Danger Den setup, but I poked a hole in my Black Ice Rad (stupid me!!)

 

So I had the Evercool sitting here from a review I did, and ripped it apart, I am using it to cool my Cpu, getting 28c-30c on idle and about 38c-40c on load, My Zalman cooler was 38c-39c on idle.. I had originally intended to use it for my vid card, witht he DD setup on the cpu, but need to get new rad first..

[pilotw]

so 3dmark05 kept crashing on me. Not sure exactly why so I picked up 3dmark03 and ran those tests:

 

Here are the numbers:

Stock settings:

6300

CPU1 113.2

CPU2 18.7

 

OCed to 2.5 with 180 divider 2.5-3-2-5 timing 2.7 volts

6506

CPU1 134.8

CPU2 22.0

 

OCed to 2.5 with 166 divider 2-2-2-5 timing 2.7 volts

6522

CPU1 132.8

CPU2 21.5

 

 

My process was 3x all tests after fresh reboot for each setting.

The interesting result is the 166 divider provided more overall performance due to the improved memory timing eventhough it had a minor impact on CPU test performance.

 

So, my conclusion is keep the memory timing as tight as possible with a lower memory divider to yield better results(in my system). Though going belwo the 166 divider will only hurt since I can already run the 2-2-2-5 timing