LOSTCIRCUITS

We used the FIC KA6110 mainboard for the setup since of all VIA 693 chipset powered boards it is currently the fastest, equal or even faster than the average BX chipset board. Benchmarks were run at 3.5x133 MHz (467MHz) except where otherwise noted.

Incoming

One of the words that always pops up in the discussions about the GeForce is scalability. Claims have been made that CPU speed doesn’t matter anymore and that the GPU will fully be able to compensate for lower CPU power. Once again, no rule without exception, actually there are more than one but Incoming is a prime example for an almost linear relation between CPU speed and frame rates

Incoming Gameindex frame rates at 467 (133x3.5) MHz, 525(150x3.5)MHz, 533(133x4)MHz, 560(140x4)MHz

Treemark

The opposite can be claimed for nVidia’s treemark. Here it really does not matter one bit whether the CPU was run at either 467 (133x3.5) MHz, 525(150x3.5)MHz, 533(133x4)MHz, 560(140x4)MHz or 600 (150x4) MHz. The result in all cases was 50.15 fps for the simple and 14.32 fps for the complex configuration.

Again, there is a certain amount of scalability but more importantly, the numbers clearly show that the V6600 profits from a higher bus speed just as well as from the overall increase in clock speed.

3Dmark and CPUmark scores at 467 (133x3.5) MHz, 525(150x3.5)MHz, 533(133x4)MHz, 560(140x4)MHz

Unreal

Unreal timedemo frame rates at 467 (133x3.5) MHz and different resolutions. Once again, there is very little difference between the settings. On a side note, Unreal in OpenGL mode did not load but terminated in an error message

Unreal Tournament

For simplicity reason we used the demoace benchmark from Ace's Hardware. It is quite impressive how little impact switching from 16 bit to 32 bit has.

Unreal Tournament Demoace frame rates at 525 (115x4.5) and 540 (4.5x120) MHz at 16 and 32 bit color depth, 1024x768 pixel resolution, skin and world levels high.

Next Page:    => P!!! OpenGL =>