Unreal Tournament 2003

    We all know UT2003 for its beautiful graphics, and the game itself isn't too bad either.  The quality of the graphics is what brings many modern CPU's and video cards to their knees.  So therefore this is a good test to see how well both video cards and motherboards perform.  We used the Hard OCP CPU test, which is run at 640*480 and uses the dm-inferno map, which is one of the most power hungry maps included.   The Video tests were run with a slightly modification to the program, instead of all the maps, we only tested dm-antalus, which apart from dm-inferno, is a very power hungry level.  This test was run at both 1024*768 and 1600*1200 with the high settings.  So lets see how the Chaintech board and the IGP did.

   We can see that in the CPU test the nForce 2 doesn't show any advantage at all compared to the SiS 745 chipset.  Apart from the single channel mode IGP the systems all have extremely similar graphs, with only a few small differences between the different cards/platforms.  The differences between the different systems is small to non-existent differences between the SiS 745 and the nForce 2, in either single or dual channel mode.  In UT2003 the difference between with IGP enabled and it disabled is pretty high, about 25% in dual channel mode, though this may be more a video card difference than that of the lowered bandwidth to the CPU.  However the differences between single channel mode and dual channel mode are very high, about 40.4%, which is nice to see.  How about when we increase the resolution to 1024*768?

    Here we see the limitations of the GeForce 4MX series start to creep in.  Even the higher clocked MX 440, which has a 75MHz core speed difference and a 90MHz (180DDR) memory speed difference, is only 19% faster than the dual channel IGP.  One may be able to play at 1024*768 with the integrated video card, as long as you lower the quality settings to normal or slightly lower levels.  When we look at the graph of the IGP modes we see only three upward spikes in either mode and no real downward spikes.  With the Parhelia there was very little difference between both chipsets, though in single channel mode the Parhelia tended to have more in the way of spikes.  Does the performance situation change when we up the resolution to 1600*1200? 

   Here we see all three modes that we tested the Parhelia in are the same, which is good as this turns into a video test.  The only noticeable difference in graphs between the GF4 cards is that the MX card has slightly higher peaks than that of the IGP.  However it would be reasonable to say that anything from the GeForce 4MX card down is definitely not 'playable' at this resolution.  The difference between the MX and the dual channel IGP isn't all that much, about 20%, and the difference between dual and single channel mode is a rather large 55%.  But Unreal Tournament 2003 is only one of the newer games, can the nForce 2 and IGP perform better in a slightly older game?

Jedi Knight II

    Jedi Knight II has been around for a fair while, and while based on the 'dated' Quake III engine it still proves itself an adequate test of the motherboard/memory and at higher resolutions it can still stress the video card fairly well.  So lets see how the nForce 2 does, with and without IGP enabled.

   Here we see something interesting, the integrated graphics of the nForce 2 in dual channel mode performs better than that of the Parhelia in any configuration.  Though it still loses to the GeForce 4MX by a small amount it does very well, and is very playable at this resolution, even that of the single channel mode is playable.  We see the improvements that adding more memory bandwidth and the other enhancements of the nForce 2 chipsets, as the chipset provides a 9-10% improvement over that of the SiS 745 chipset.  This benchmark takes a lot of bandwidth and as such we see a 47% increase in frame rate when we feed the chipset twice the bandwidth and 3.2GB/s more than the CPU needs at maximum to the IGP.  How though does the chipset perform at 1600, is it still playable or is it as with UT2003?

   When we increase the resolution we see the performance of the IGP decrease compared to that of the GeForce 4 MX.  Here the MX performs 53% better than the dual channel IGP, and it is also close to and can be considered playable.  The differences between the two platforms is small to negligible, and the dual channel result can be considered a minor blip in the results.  Overall from the Jedi Knight results bandwidth plays a big part in the performance of the system at 1024 for all video cards and at 1600 for the integrated video of the chipset.  But people don't always play games how does the nForce 2 perform compared to the SiS 745 chipset in other applications?