Real World Tests
We see that the XMS memory overclocks very well and does extremely well in synthetic tests, but how does it fair given in two real world applications, games and DivX rendering. First let us look at Jedi Knight II. We know that JKII is based on the Quake III engine, which seems to thrive on added bandwidth. Will we see any improvement between the XMS memory and its competitors? Lets find out.
What can we see with Jedi Knight? The XMS memory does very well, beating or tying the previous best frame rate. A quick note that the XMS memory had a (4*)85MHz AGP bus and 42MHz PCI bus behind it, as well as a ~40MHz CPU speed improvement over any of its competitors. None the less, the fact that changing to turbo settings at 214MHz provides a 7.5% increase in frame rate over the fast settings is something that is very nice to see. The other RAM can only catch the slower settings on the XMS memory, and have no chance at catching the XMS at turbo settings. Does this trend continue with DivX rendering?
The results are very similar to the Jedi Knight II scores, as the two XMS scores are as fast or faster than any of the other RAM. There is a 6% increase in encoding speed between turbo and fast settings, which equates to a 9:10 quicker encode of a two hour movie. This isn't a lot but can be helpful in rendering smaller projects that need things done quickly, and it only needs an increase in FSB. The competition catches the XMS at fast settings, but only when the competition is running at its fastest settings and speed.
Previous Page - Overclocking/Synthetic Tests
Next Page - Conclusion
|
