Theoretical Tests in 3DMark06
3DMark06 incorporates all the theoretical tests from 3DMark05 as well as two new tests, Shader Particles Test (SM3.0) and Perlin Noise (SM3.0). As the names suggest these two tests both require Shader Model 3.0 support to run.
- Shader Particles Test (SM3.0) resembles the particles processing test from 3DMark 2001 but uses Shader Model 3.0. The physical model of the particles behavior is calculated using pixel shaders. The particles are then visualized by means of vertex texture fetch. The trajectories of 409,600 particles in a simple gravitation field with environment resistance are calculated using Euler integration and are checked for collisions with a height field. Besides Shader Model 3.0, the test requires a graphics processor capable of vertex texture fetch, so it works only on graphics cards with GeForce 6 and 7 architectures. ATI’s Radeon X1000 doesn’t support this feature.
- Perlin Noise (SM3.0) uses the so-called 3D Perlin noise to simulate realistic changing clouds. Perlin noise is the basis for some procedural texturing and modeling techniques and will probably be used more widely in the future since the effects it can help create require high computational power, but do not put a big load on the graphics card’s memory whose performance is growing at a slower rate than the math1ematical performance of the GPU. The pixel shader from this test consists of 495 instructions (447 arithmetic and 48 texture lookups). For reference: the SM3.0 standard requires support of shaders consisting of up to 512 instructions as the minimum. All texture instructions create a single 32-bit texture with 256x256 resolution. It is only 64KB big, so the test is indifferent towards the amount and frequency of the memory the graphics card carries.
The rest of the tests, including the batch size tests, have been left the same. So, this is the end of the theoretical part of this review. We will now proceed to test contemporary graphics cards and CPUs in the new benchmark from Futuremark.