550 MHz G71 based GPU on 90 nm process
* 300+ million transistors (600 million with Cell CPU)
* Multi-way programmable parallel floating-point shader pipelines
o Independent pixel/vertex shader architecture
o 24 parallel pixel-shader ALU pipes
+ 5 ALU operations per pipeline, per cycle (2 vector4 or 2 scalar/dual/co-issue and fog ALU)
+ 27 FLOPS per pipeline, per cycle
o 8 parallel vertex pipelines
+ 2 ALU operations per pipeline, per cycle (1 vector4 and 1 scalar, dual issue)
+ 10 FLOPS per pipeline, per cycle
o Maximum vertex count:1.1 billion vertices per second
+ polygon count: 366.6 million polygons per second
o Maximum shader operations:100 billion shader operations per second
o Announced: 1.8 TFLOPS (trillion floating point operations per second) (2 TFLOPS overall performance)[5] [6]
* 24 texture filtering units (TF) and 8 vertex texture addressing units (TA)
o 24 filtered samples per clock
+ Maximum texel fillrate: 13.2 GigaTexels per second (24 textures * 550 MHz)
o 32 unfiltered texture samples per clock, ( 8 TA x 4 texture samples )
* 8 Render Output units / pixel rendering pipelines
o Peak pixel fillrate (theoretical): 4.4 Gigapixel per second
o Maximum Z sample rate: 8.8 GigaSamples per second (2 Z-samples * 8 ROPs * 550 MHz)
o Maximum anti-aliasing sample rate: 8.8 GigaSamples per second (2 AA samples * 8 ROPs * 550 MHz)
* Maximum Dot product operations: 51 billion per second
* 128-bit pixel precision offers rendering of scenes with high dynamic range rendering (HDR)
* 256 MB GDDR3 RAM at 700 MHz
o 128-bit memory bus width
o 22.4 GB/s read and write bandwidth
* Cell FlexIO bus interface
o 20 GB/s read to the Cell and XDR memory
o 15 GB/s write to the Cell and XDR memory
* Support for OpenGL ES 2.0
* Support for S3TC texture compression
[ source ]
