elle urilise directx 10 si.
1) The Xenos will have 382 Million Transistors, The most in Any GPU, and that includes G70
2) The main GPU has 232, the 150 alone are in EDRam , So you can call the Edram a mini GPU
3) ATI's Xenos current ALU utilization in tests ran at 95% which is higher than X850 (55%) the highest ever in a card tested
4) The Way Tesselation works is it will literally mean the end of Blocky models in best of games
5) The way multiple threads are interleaved so that instruction dependency becomes a non-issue
6) We will have Dynamic Branching at No literally no cost, first time ATI has ever managed that or Nvidia
7) Every game will have 4xFSAA, even if the developer wants it or not.
8) The Hierarchial Stencil buffer was a smart move. That should make stencil shadowing scream. And all the improvements to the system make shadow mapping a very high performance option too.
9) MSAA works on all render targets, all render targets sit in eDRAM, etc. The more common way to say this is that it "just works." There's a welcome lack of features that cannot play with each other.
10) ATI will design All Future GPU based on the Xenos. You will get Xenos like cards in 2007/2008 in the form of R600
11) HDR rendering is truly reasonable for the first time.
12) The tesselation ability might make some developers create very smooth models at high performance.
13) Better performance at shadowing than any ATI Card ever.
14) Xenos work began in Q4 2002 and ended Q1 2005 (3 years of R&D), now you know why X850 sucked
15) Some design features are solely design for 512 Ram and Solely designed for 3 cores of 3.2 Ghz PPC, Those design features will never work in any other sort of machine.
- *16) The Xenon has WG2.0 , Xenos implements it. Directx 10 is implemented in Xbox 360 design. You will see this design in PCs only after final Longhorn ships
17) SIMD engines can be individually assigned to different workloads, Just like the Cell processor
18) Article says the next step will be to see not how efficient the GPU is in Xbox 360, but how inefficient current generation GPUs are compared to Xenos
19) Alpha to Mask - means there's no need for supersampling to get anti-aliased alpha-blended textures
20) Ultra-fast Z-only pass - should massively reduce overdraw compared to current generation including R500
21) MEMEXPORT - arbitrary tesselation algorithms
22) The 192 processors on the daughter die can be equivalent to what you can call a mini PPU
23) This is probably the first time that general purpose physics calculation would be achievable, with a reasonable degree of success, on a graphics processor and is a big step towards the graphics processor becoming much more like a vector co-processor to the CPU."
24) The Xenos shader contains a large number of independent groups of pixels and vertices (threads) which are 16 wide. In order to hide the latency of an instruction for a given thread, a number of other threads are used to "fill in the gaps". By doing this, the ALU's are fully utilized all the time, and the shader can have direct data dependency on every instruction and still run full rate."
Bon ba ducoup j'ai trouvé une réponse:
La tesselation n'est pour le moment utilisé que par rare dans "viva pinata" (jeux bizard^^) pour le rendu du terrain.
Mais du coup... j'ai baucoup de mal situé la nature exact de la chose si je compare avec la definition de la fonction que j'ai pu lire.!
