Performance

For this test we are using one of the higher end cards which both products support, the ATI Radeon X1950 Pro. The model being used is MSI and comes with a cooler which is slightly redesigned when compared to the reference model providing quieter cooling at similar temperatures to the standard cooler.

The core and memory speeds on this card are:
• Engine Clock Speed: 575MHz
• Memory Clock Speed: 686MHz/1372MHz

Testing was performed in a standard PC case (Thermaltake Swing) with two 12cm fans installed, one front, one rear (2000rpm, 35 CFM). The other system fans were the Arctic Cooling Freezer 7 Pro on the CPU and the test systems Enermax Noisetaker 600w PSU with its stock fans. The test system also featured an Intel XBX2 motherboard, 2x SATA Drives and an IDE DVD Writer.

Load testing was performed by looping 3Dmark06 graphic tests for 1 hour. Idle temperatures were taken after the system had been running for 2 hours with no activity on Windows XP.

Don't have flash installed? then click here for a jpeg

Fanless= No case fans on
Fan= Case fans on
Extra 12cm= 12cm case fan blowing directly on to the VNF100
Turbo= Turbo module attached to the S1
MX-2= replacing the MX-1 compound with MX-2

The graph above shows the various results we obtained when testing. What we can take from these is that the two coolers being tested today vary significantly in their ability to cool our test card. When in fanless/silent mode the Zalman cooler is 4 degrees hotter than the Arctic Cooling at idle and 12 degrees at load.

NOTE: We do not recommend running either cooler in this mode, case cooling should be used at all times.

When we turn on our case cooling we see temperatures dropping significantly and the VNF100 becomes much more competitive with the S1. Both also manage to outperform the X1950’s stock cooler at load temperatures which is a good achievement.

We next applied some more powerful cooling to the system. In the case of the S1 this meant attaching the turbo module. On the VNF100 we mounted a 12CM fan above the rear heatsink which pushed a large volume of air across the cooler. In this scenario temperatures of both coolers dropped a great deal, the Arctic Cooling S1 remained the better performer, however the VNF100 was only two degrees behind at load temperatures.

Our final test was to satisfy an interest - we recently took a look at Arctic Coolings MX-2 thermal compound when applied to CPU’s so we decided to see how the compound would work within a graphics card environment. The result shows that the MX-1 based results using the turbo module are 2c higher than the MX-2 results, a nice little increase in performance/decrease in temperatures merely by changing compounds.