Pixel filling and shader power
As usual, let's kick things off with a look at how our cards compare when it comes to theoretical fill rate and memory bandwidth peaks. Theoretical peaks don't always determine real world performance, but they should give us a rough idea of what to expect from the cards we've assembled. The chart below is arranged in order of ascending multi-texturing fill rate.
| Core clock (MHz) | Pixel pipelines | Peak fill rate (Mpixels/s) | Texture units per pixel pipeline | Peak fill rate (Mtexels/s) | Memory clock (MHz) | Memory bus width (bits) | Peak memory bandwidth (GB/s) | |
| Albatron Trinity PC6600 | 300 | 8* | 1200 | 1 | 2400 | 500 | 128 | 8.0 |
| Radeon X700 | 400 | 8 | 3200 | 1 | 3200 | 600 | 128 | 9.6 |
| Abit RX700 Pro-128PCIE | 420 | 8 | 3360 | 1 | 3360 | 864 | 128 | 13.8 |
| Gigabyte GV-RX70P256V | 420 | 8 | 3360 | 1 | 3360 | 864 | 128 | 13.8 |
| Chaintech SE6600G | 500 | 8* | 2000 | 1 | 4000 | 1000 | 128 | 16.0 |
| XFX PVT43GNDD7 | 500 | 8* | 2000 | 1 | 4000 | 1200 | 128 | 19.2 |
Although their unconventional eight-pipe, four-ROP configuration gives them a comparatively weak single-texturing fill rate, our GeForce 6600 GT cards are well ahead of the field when it comes to multi-texturing. The 6600 GT cards also have a healthy advantage over the X700 Pros when it comes to peak memory bandwidth.
In the vanilla department, the Albatron GeForce 6600 may have a hard time keeping up with the standard Radeon X700. The X700 not only has superior single- and multi-texturing fill rate, it's also packing 1.6GB/sec more memory bandwidth.
Let's see how these theoretical peaks pan out in 3DMark05's synthetic fill rate tests.
With today's latest games loaded with shader-driven effects, shader power has become increasingly important. 3DMark05 has a handful of synthetic pixel and vertex shader tests that can help us isolate each card's shader potential.
