Pemalite said:
HoloDust said:
Yeah, I know about that one too...or variation of it...yet I have no idea how anyone come up with GC and XBOX numbers.
GC has 4:1:4:4 @162MHz, while XBOX has 4:2:8:4 core config @233MHz (pixel:vertex:TMU:ROP)...how one comes to actual FLOPS is beyond me without knowing particular architectures.
For example, I still can't figure out how to convert PS3's RSX GPUs specs into FLOPS (24:8:24:8 part), since, to me at least, something seems to be off with quoted numbers, as if they are conflicting each other. For example, current GFLOPS at wiki are 192 for pixel shaders (I remember this being changed numerous time), and this is quoted from K1 whitepaper, which states 192GFLOPS for whole PS3's GPU.
- 24 parallel pixel-shader ALU pipelines clocked at 550 MHz
- 5 ALU operations per pipeline, per cycle (2 vector4, 2 scalar/dual/co-issue and fog ALU, 1 texture ALU)
- 27 floating-point operations per pipeline, per cycle
- Floating Point Operations per a second : 192 GFLOPs
- 8 parallel vertex pipelines
- 2 ALU operations per pipeline, per cycle (1 vector4 and 1 scalar, dual issue)
- 10 FLOPS per pipeline, per cycle
|
They could be including Vertex performance in that calculation.
Citing nVidia is a pretty dubious affair, because nVidia will want to fluff up numbers as much as possible.
|
Yeah - I remember around that time people were using some wild numbers, 400 or so GFLOPS. Which when you get down to math, it really comes to something as silly as that:
(24x27FLOPS + 8x10FLOPS)*550MHz = 400GFLOPS
And those 27 per cycle and 10 per cycle numbers are indeed from official nVidia document - or it seems so... https://www.pcper.com/reviews/Graphics-Cards/NVIDIA-GeForce-7800-GTX-GPU-Review/Hardware-Details
Yet again, 192GFLOPS for whole GPU is from official nVidia document as well... https://www.nvidia.com/content/PDF/tegra_white_papers/Tegra_K1_whitepaper_v1.0.pdf#page=18
This is the very reason why i said I would love to see math behind those numbes for GC and XBOX - because, without knowing underlying architectures it's just guesswork, and given base config (4:1:4:4@162MHz vs 4:2:8:4@233MHz), some of those numbers look...well, quite silly.
