Hmmm someone correct me if I'm wrong please; so far I live under the impression that:
1. DX9.0/DX9.0L = WGF1.0, whereby even Intel GMA900 will make WGF1.0.
2. Longhorn will ship with WGF1.0/DX9.0L, while WGF2.0 (DX10?) is supposed to ship later.
3. ATI/NVIDIA seem to plan this far WGF2.0 GPU releases fairly close if not at the same time as Longhorn (unless something has changed in the meantime ~H2 2006).
If by API above you should mean WGF1.0, then there might be some truth to that under certain conditionals. However if you should mean that both IHVs will release their WGF2.0 GPUs at the same time WGF2.0 appears (and the latter truly arrives later than Longhorn), then it's a totally different perspective.
Joe,
Yes and no. Yes to the console R&D investment and no to rather small change to former product lines. I don't think it's that simple to change the existing pipeline into a SM3.0 architecture with a simple twist. Even more if ATI wanted an architecture that yields more effective branching. All IMHLO
1. DX9.0/DX9.0L = WGF1.0, whereby even Intel GMA900 will make WGF1.0.
2. Longhorn will ship with WGF1.0/DX9.0L, while WGF2.0 (DX10?) is supposed to ship later.
3. ATI/NVIDIA seem to plan this far WGF2.0 GPU releases fairly close if not at the same time as Longhorn (unless something has changed in the meantime ~H2 2006).
It definitely makes the most marketting sense for both companies to release their WGF 2.0 parts at the same time that Microsoft releases the API. Yes, this may mean that ATI will have to bite a pretty big bullet on R&D with the R520 and its bretheren.
If by API above you should mean WGF1.0, then there might be some truth to that under certain conditionals. However if you should mean that both IHVs will release their WGF2.0 GPUs at the same time WGF2.0 appears (and the latter truly arrives later than Longhorn), then it's a totally different perspective.
Joe,
It remains to be seen exactly how "different" R520 is from R4xx in terms of architecture, of if its basically "R300, with FP32 support and SM 3.0 bolted on."
Clearly, ATI has spent a LOT of R&D time on the defunct R400 transformed into the R500/Xenon (and assumed to be basis for R600 PC chips). I'm thinking there really isn't all that significant R&D for R520.
Yes and no. Yes to the console R&D investment and no to rather small change to former product lines. I don't think it's that simple to change the existing pipeline into a SM3.0 architecture with a simple twist. Even more if ATI wanted an architecture that yields more effective branching. All IMHLO