NV43 versus NV40
| NV43 | NV40 | |
| Process | 110nm @ TSMC | 130nm @ IBM |
| Transistor Count | Unknown | 222M |
| Geometry Pipeline | VS3.0 | VS3.0 |
| Fragment Processor | PS3.0 | PS3.0 |
| Fragment Processor Setup | 2 full ALU (not equal) each with 1 mini ALU, Fog ALU, per pipe | 2 full ALU (not equal) each with 1 mini ALU, Fog ALU, per pipe |
| Fragment Processor Precision | FP32, FP16 | FP32, FP16 |
| Traditional Render Setup | 8 x 1 | 16 x 1 |
| ROPs | 4 | 16 |
| Vertex Shaders | 3 | 6 |
| Basic Texture Filtering | Bilinear | Bilinear |
| Texture Filtering | Bilinear, Trilinear, 16X Anisotropic | Bilinear, Trilinear, 16X Anisotropic |
| Antialiasing | Multi-sampling and super-sampling | Multi-sampling and super-sampling |
| AA Sample Type | Rotated grid up to 8X with supersampling combined at 8X | Rotated grid up to 8X with supersampling combined at 8X |
| Native Bus Support | PEG16X | AGP8X |
| Memory support | GDDR3 | GDDR3 |
| Basic Core Frequency | 500MHz | 400MHz |
| Basic Memory Frequency | 1000MHz | 1100MHz |
| Memory Bus Width | 128-bit, memory crossbar | 256-bit, memory crossbar |
| Basic Pixel Fillrate | 4000Mpixel/sec | 6400Mpixel/sec |
| Basic Multitexture Fillrate | 4000Mtexel/sec | 6400Mtexel/sec |
| Basic Memory Bandwidth | ~16.0GB/sec | ~35.2GB/sec |
Now I mention the peak fillrates of NV43 at 500MHz being 4000M pixels or texels per second. In reality, due to the ROP count, that's an internal peak fillrate only, not a written pixel to the backbuffer fillrate. Those written pixel fillrate figures are halved, due to the ROP count. It's up to the performance figures to bear that out, in terms of a good design decision.
Produced on TSMC's 110nm process, besides being NVIDIA's first native PCI Express GPU, it's also their first GPU at 110nm at any foundry. Something to bear in mind, given the disparity in clocks between the 6600 GT and the plain 6600, is just how NVIDIA are getting 500MHz out of a core that's not using a low-k dielectric. It stands to reason that yeilds at 500MHz aren't likely to be high on that particular process. However, NVIDIA are bullish about yeilds and the number of GTs released to OEM and retail recently, ready for sale. Here's hoping they're not just being positive for the sake of it and that yeilds genuinely are good.
Before we have a look at performance, let's have a look at the reference board.
