Archive 1997–2024

Gigabyte goes "Green" with DES software

Test Setup

Gigabyte GA-X48T-DQ6 Testbed
Processors Intel Core 2 Extreme QX9650
Quad-core, 3GHz, 2x6MB Cache, 1333FSB

Intel Core 2 Extreme QX6800
Quad-core, 2.93GHz, 2x4MB Cache, 1066FSB
CPU Voltage Various according to DES Level
Cooling Water cooling: Swiftech Apogee GTX, PA120.3 Rad, 3 X Laing DDC Ultra Petra Top pumps in series, 3x Panaflo 120mm Fans 7-10V
Power Supply OCZ 1000W
Memory OCZ DDR3 PC3-14400 (DDR-1800) Platinum Edition
2GB XP, 4GB Vista 64
Memory Settings Default via SPD
Video Cards Asus 8800 GTS 640MB
Video Drivers Nvidia 169.09 (XP)
Hard Drive Western Digital 7200RPM 250GB SATA 3/Gbps 16MB Buffer
Optical Drives Plextor PX-B900A
Case Lian Li -75
BIOS F4B
Operating System Windows XP 32 Bit SP2
Watt Meter Paget 9149
.

We decided to use Windows XP 32-bit as our test operating system with 2GB of memory installed. The Stanford Folding@Home client was used to simulate 100% loading tests and to test for stability at a variety of processor core voltage levels. Prime95 Torture tests were also run in small FFT mode to place a maximum load on the CPU in order to determine the worst case scenario for voltage droop at each of the available DES levels. As we are monitoring for outright system power consumption and savings, we did not account for PSU losses in any of the idle and full load wattage figures. However, PSU load efficiency will obviously affect the total level of power saving that can be realized. The most important factor to us today is how much overall power we can save at the wall. We did find that the watt-meter cycled between a variety of voltages at both idle and full load of the processor. For this reason, the figures shown in our tables represent averages that are accurate to around +/- ~.5W.

Originally published at https://www.anandtech.com/show/2425 (Wayback Machine).