ATI R580 Specification
R580 vs R520
| ATI R580 GPU Properties | ||
|---|---|---|
| GPU | ATI R580 | ATI R520 |
| Process and Fabricator | 90nm @ TSMC, 90GT w/ low-k | |
| Die Size | 18.5mm² | 13x12mm |
| Transistor Count | 384 million | 321 million |
| DirectX Shader Model | Shader Model 3.0 | |
| Basic Configuration (VP/FP/ROP) | 8/48/16 | 8/16/16 |
| Vertex Shader Info | VS3.0 (no texld) 5D FP32, co-issue MADD, branch single cycle trig functions |
|
| Fragment Processor Info | PS3.0 4D FP32, dual-issue ADD+MADD, branch single cycle trig functions |
|
| ROP Info | 6x Int8/FP16/FX16/Int10 MSAA (2 subsamples/cycle) 2x Z-only rate FP/FX16 blender |
|
| Texture processing | 16 FP32 address units, 4 samplers Bilinear filter for integer samples |
|
| Memory Interface | 256-bit, 512-bit ring bus GDDR->GDDR4 |
|
| GPU |  |
 |
| Display output | 2x dual-link DVI TMDS transmitters ATI Avivo |
|
As you can see, ATI spent around 60 million transistors adding thrice the fragment ALUs and all the associated extras, including thread control logic and register file space.
In the top configuration of X1900 XTX - 650MHz GPU, 774MHz memory - theoretical rates work out as follows, versus X1800 XT.
| Theoretical Rates for ATI Radeon X1900 XTX, versus Radeon X1800 XT | ||
|---|---|---|
| ATI Radeon X1900 XTX | ATI Radeon X1800 XT | |
| Core Clock | 650MHz | 625MHz |
| Memory Clock | 775MHz | 750MHz |
| Pixel fillrate | 10.4G pixels/sec | 10.0G pixels/sec |
| Texture sampling rate | 41.6G samples/sec | 40.0G samples/sec |
| Z-only fillrate | 20.8G samples/sec | 20.0G samples/sec |
| Vertex transform rate | 1.30G tris/sec | 1.25G tris/sec |
| VP MADD issue rate | 5.20G instr/sec | 5.00G instr/sec |
| FP MADD issue rate | 31.2G instr/sec | 10.0G instr/sec |
| Memory bandwidth | 49.60 GB/sec | 48.00 GB/sec |
Fragment processing monster? Looks that way. Want to see a board? Thought you'd never ask.