Asetek Low Cost Liquid Cooling (LCLC) System
Test Setup & Methodology
This page contains an explanation of the test methodology and the specifications of the test system used to perform the tests which produced the results on the next page.
 |
|
- Intel Core 2 Duo E6850 (3.0 GHz) Dual Core (1333 MHz FSB)
- ASUS P5K3 Deluxe Motherboard
- NVIDIA GeForce 8800 GTX
- 2 x 1GB PC3-11000 DDR3 Memory
- Western Digital Raptor 74GB 10,000 RPM SATA Hard Disk
- DVD+/-RW Serial ATA Optical Drive
- PC Power & Cooling Silencer 610W EPS12V Power Supply
- Arctic Silver 5 Thermal Compound
- Microsoft Windows XP Professional (32-bit)
 |
|
For our thermal tests, we used two air coolers for reference in our CPU temperature tests and the stock GeForce 8800 GTX reference cooler for GPU tests. We chose the Intel stock cooler and a large, heat-pipe based air cooler; the Silverstone NT-06. For all tests, a Scythe S-FLEX SFF21D (800 RPM) was installed on the LCLC's heat exchanger. All other coolers used the fans they are packaged with. The Silverstone was tested twice, once at 750 RPM and a second time at 2640 RPM (max speed for stock fan). All thermal tests were performed on an open-air test bench. All testing was performed with the CPU and GPU at stock frequencies.
Two separate tests were performed. The CPU was stressed with Everest's built-in stress test function. This stresses the CPU and system memory, but the graphics card is idle. This simulates a common scenario during CPU intensive tasks and we believe that Everest's stress test does a good job of stressing the CPU. This test should result in peak CPU temperatures and it was used to measure CPU cooling performance.
The second type of test performed was a real-world gaming test using Call of Duty 4. This test was conducted with all graphical settings set to their highest available levels at a resolution of 1920x1200. Anti-aliasing was set to 4x while anisotropic filtering was set to 16x. Call of Duty 4 is not the most stressful game for hardware currently available (that distinction obviously belongs to Crysis), but we believe it provides a good representation of a fully-stressed system. This test was used to measure GPU cooling performance. The CPU temperatures from this test were not measured, but they were generally below those from the first test.
Temperatures were measured with Everest Ultimate Edition. CPU temperature accuracy was double-checked with Core Temp, which matched perfectly. For all CPU results, the recorded temperature is the hottest of the two cores in the Core 2 Duo E6850 used for testing.
For all tests, the test was allowed to run for 15 minutes at which point Everest begin monitoring temperatures. After another 30 minutes, the average temperature that Everest had recorded over the proceeding 30 minutes was recorded. There was a 30 minute cool-down period between tests where the system sat idle.
All testing was performed 4 consecutive times, over 3 different days at different times of the day. The results of all 4 test iterations were averaged to produce the numbers provided in the graphs. This was done to reduce the effect that slight variances in ambient temperature (which can change throughout the day) had on our test results. Ambient temperature throughout testing was maintained between 21 and 22 degrees Celsius.
Two separate tests were performed. The CPU was stressed with Everest's built-in stress test function. This stresses the CPU and system memory, but the graphics card is idle. This simulates a common scenario during CPU intensive tasks and we believe that Everest's stress test does a good job of stressing the CPU. This test should result in peak CPU temperatures and it was used to measure CPU cooling performance.
The second type of test performed was a real-world gaming test using Call of Duty 4. This test was conducted with all graphical settings set to their highest available levels at a resolution of 1920x1200. Anti-aliasing was set to 4x while anisotropic filtering was set to 16x. Call of Duty 4 is not the most stressful game for hardware currently available (that distinction obviously belongs to Crysis), but we believe it provides a good representation of a fully-stressed system. This test was used to measure GPU cooling performance. The CPU temperatures from this test were not measured, but they were generally below those from the first test.
Temperatures were measured with Everest Ultimate Edition. CPU temperature accuracy was double-checked with Core Temp, which matched perfectly. For all CPU results, the recorded temperature is the hottest of the two cores in the Core 2 Duo E6850 used for testing.
For all tests, the test was allowed to run for 15 minutes at which point Everest begin monitoring temperatures. After another 30 minutes, the average temperature that Everest had recorded over the proceeding 30 minutes was recorded. There was a 30 minute cool-down period between tests where the system sat idle.
All testing was performed 4 consecutive times, over 3 different days at different times of the day. The results of all 4 test iterations were averaged to produce the numbers provided in the graphs. This was done to reduce the effect that slight variances in ambient temperature (which can change throughout the day) had on our test results. Ambient temperature throughout testing was maintained between 21 and 22 degrees Celsius.
