nagus
2003-11-01, 18:11:51
Quelle: http://www.beyond3d.com/forum/viewtopic.php?topic=8782&forum=9
As we mentioned previously, much of the focus of ATI’s Analyst Day appeared to be on the transition across to PCI-Express. This is of course important to both ATI and NVIDIA, and hence the analysts, as numerous OEM’s will transition across to this bus once its available and its key for the graphics vendors to have product available at the right time to fill those slots. However, we have been contacted by another analyst present at ATI’s presentation that looked for some different details about the upcoming architectural cycle.
It’s long since been suggested that the next generational cycle from ATI and NVIDIA, with the R420 and NV40 architectures, will follow the DirectX9 Vertex and Pixel Shader 3.0 specification. The crucial element here is that the VS/PS3.0 specification has been set down by Microsoft since the initial release of DirectX9, hence both R420 and NV40 have a stable specification target to hit and that specification was set early in their design cycle.
At the Analyst Day ATI suggested that, technologically, the transition from DX8 to DirectX9 is actually for more important than the transition from AGP to PCI-Express, because of the complexity involved. Because of the way ATI have hit the DirectX9 specification they are presently enjoying very good margins whilst still having a performance advantage in some cases – R300/R350/R360, for instance, were heralded as performance leaders by many, and yet they are still based on the relatively cheaper 150nm process, making that competitiveness cheaper to ATI.
With the transition to the next generation architectures ATI believe that they will not have both a margin and performance advantage, probably because of the fact that both ATI and NVIDIA have been aiming for this stable specification. With the specification set down in stone the main differentiator between the two companies will likely be what performance estimate both have set, and this is the guessing game each have to play when designing their parts.
ATI mentioned to Analysts that for the next cycle that in the high end, and to a lesser extent mainstream, they would be willing to sacrifice some margin in order to gain performance. The implication being that perhaps R420 will be a relatively large chip with numerous pipelines in order to gain that performance.
As we mentioned, it seems likely that given the timescales these parts are required by R420’s design would have been finalised and sent to the fab. Likewise, at NVIDIA’s recent Editors Day they showed the silicon masks for NV40, also suggesting that this has been sent to the fab. This being the case, both parts a now set in stone and both companies are left guessing whether they picked the best performance targets.
As we mentioned previously, much of the focus of ATI’s Analyst Day appeared to be on the transition across to PCI-Express. This is of course important to both ATI and NVIDIA, and hence the analysts, as numerous OEM’s will transition across to this bus once its available and its key for the graphics vendors to have product available at the right time to fill those slots. However, we have been contacted by another analyst present at ATI’s presentation that looked for some different details about the upcoming architectural cycle.
It’s long since been suggested that the next generational cycle from ATI and NVIDIA, with the R420 and NV40 architectures, will follow the DirectX9 Vertex and Pixel Shader 3.0 specification. The crucial element here is that the VS/PS3.0 specification has been set down by Microsoft since the initial release of DirectX9, hence both R420 and NV40 have a stable specification target to hit and that specification was set early in their design cycle.
At the Analyst Day ATI suggested that, technologically, the transition from DX8 to DirectX9 is actually for more important than the transition from AGP to PCI-Express, because of the complexity involved. Because of the way ATI have hit the DirectX9 specification they are presently enjoying very good margins whilst still having a performance advantage in some cases – R300/R350/R360, for instance, were heralded as performance leaders by many, and yet they are still based on the relatively cheaper 150nm process, making that competitiveness cheaper to ATI.
With the transition to the next generation architectures ATI believe that they will not have both a margin and performance advantage, probably because of the fact that both ATI and NVIDIA have been aiming for this stable specification. With the specification set down in stone the main differentiator between the two companies will likely be what performance estimate both have set, and this is the guessing game each have to play when designing their parts.
ATI mentioned to Analysts that for the next cycle that in the high end, and to a lesser extent mainstream, they would be willing to sacrifice some margin in order to gain performance. The implication being that perhaps R420 will be a relatively large chip with numerous pipelines in order to gain that performance.
As we mentioned, it seems likely that given the timescales these parts are required by R420’s design would have been finalised and sent to the fab. Likewise, at NVIDIA’s recent Editors Day they showed the silicon masks for NV40, also suggesting that this has been sent to the fab. This being the case, both parts a now set in stone and both companies are left guessing whether they picked the best performance targets.