The system configuration is as follows:
CPU: | AMD Opteron 148 CABYE (S939; 1MB; Venus) |
---|
Clock speed: | 11 x 200 MHz = 2200 MHz, Memory at DDR-400 |
---|
Motherboard: | DFI NF4 Ultra D |
---|
Memory: | 2x 1024MB OCZ PC-3200 Platinum EL |
---|
Video Card: | ATI Radeon X1900XT PCI-e |
---|
Harddisk: | 4 x 36GB WD Raptor drives in raid 0 Maxtor 200GB PATA drive |
---|
Power Supply: | OCZ 600W GameXStream |
---|
Software: | Windows XP SP2, Catalyst 6.12 |
---|
The watercooling system consists of the following:
CPU block: | Swiftech Storm Revision 1 |
---|
Pump: | Danger Den (Laing) D5 set at #5 |
---|
Radiator: | Swiftech MCR320-QP-K |
---|
Fans: | 3 x Yate Loon D12SL-12 120mm fans |
---|
Tubing: | Masterkleer 7/16” ID tubing |
---|
Fittings: | All components use ½” OD barbs |
---|
To test the system, I ran one loop of 3DMark 06, then recorded the highest temperature logged by SysTool. I then repeated the test two more times, logging the temp after each test, then averaged the three readings. The “Before” testing was completed with a Zalman VF900-Cu installed on the card, and CPU temperature was included to see what impact adding the Tyee would have. Ambient temperature was recorded with a standard mercury thermometer, and was kept at approximately 21°Celsius (+ or – 1°).
Anyone who owns an X1900XT can tell you these temperatures with a VF900-Cu installed are much better than stock temps.
With the Tyee installed, all the card temperatures dropped considerably, and the CPU temp rose only slightly. The GPU core temp dropped by 17°C, the PCB temp dropped by 42°C (more than half), and the VRM temp dropped by 22°C. Many of the load temperatures with the Tyee are better than idle temperatures with the Zalman, but that is why we use watercooling. All this, and the CPU temp was only increased by 2°C only when idling.