Testing & Results

Testing Methodology

The following programs were used during testing;

• FurMark build 1.6.0 - For monitoring temperatures under load.
• ATI Tool v0.26 - For attaining max stable O/C data.
• RivaTuner - For driver level overclocking.
• CoD 4 modern warfare - For real world stability testing.

Ambient temperatures were measured using a relatively cheap digital case temperature monitor and were consistently 23°C / 24°C. 30 minutes were allowed for the temperatures to stabilize before starting each test, Idle temperatures were then recorded after that FurMark was allowed to run for 10 minutes to load the GPU and find maximum temperatures. Each test was conducted 3 times and an average was taken from the 3.

The tests were conducted under the following conditions;

• 9800GT with stock cooler attached
• 9800GT with Xigmatek Bifrost attached
• Max O/C on 9800GT with Xigmatek Bifrost attached - ATI Tool (Core 720MHz / Memory 1140MHz / Shader 1800MHz)
• Max stable O/C without instability during gaming on 9800GT with Xigmatek Bifrost attached - CoD4 (Core 675MHz / Memory 1080MHz / Shader 1687MHz)

Fan speeds were always at 100% which wasn't a problem as the two 100mm fans on the Bifrost are nearly inaudible at around 20db. The Bifrost uses one standard 3 pin fan connector for both fans so I was able to connect it to my fan controller, the speed of the stock fan was controlled using Riva Tuner. I'm relatively new to overclocking graphics cards, I'm sure that with a little more time and experience then I could probably achieve a better stable in game O/C. With this said, The point that I would really like to get across is that there is no real temperature difference between each of the overclocks. No matter how much I pushed the card, the Bifrost kept it nice and cool.

Test System

• Video: PNY GeForce 9800 GT 512mb (G92 65nm)
• Motherboard: ASUS P5B Deluxe WIFI/AP
• System Memory: 2Gb PC2-6400 OCZ platinum rev2
• Processor: Intel Core 2 Duo E6600 2.4 GHz @ 3.60GHz
• Audio: Integrated digital HD audio
• Disk Drive 1: Western Digital SE Caviar 200GB SATA 7200rpm
• Optical Drive: LITE -ON DVDRW LDW-411S
• Enclosure: Xigmatek Midgard-W
• PSU: OCZ StealthXStream 500w
• Operating System: Windows XP Pro 32 bit

Results

The stock cooler was tested purely for reference, after just 2 minutes I was worried about the temperatures, as you can see in the graph below, the card ran super hot with the stock cooler attached. I decided to end the test early so that I didn't toast my GPU before I had the chance to test the Xigmatek Bifrost.

With the Xigmatek Bifrost attached it was a different story altogether, temperatures stabilized at their maximum after 3 minutes and stayed put for the duration of the test. I am still trying to get over the shock from recording such low temperatures. A whole 18°C difference between idle temps and 38°C difference between load temps. I even decided to run FurMark in extreme burning mode to see if I had the settings wrong and there was no difference. The message here is that my O/C was not restricted by temperature, the amazing Bifrost was consistently able to keep the load temperature below the idle temperature recorded using the stock cooler.

The graph says it all, this is a really extreme cooler. If only there was a way I could attach it to my CPU and replace my noisy zalman to see what it could do there, I'm sure it could work wonders alot like the HDT-S1283 also by Xigmatek.

<< Previous Page - Next Page >>