P
prasanth
Guest
ALTHOUGH MICROSOFT STILL still hasn't launched DirectX 9, which perhaps it's reserving for November and Comdex, news reaches us of an upcoming change to the application programming interface.
We wrote about this API at Cebit time in March 2002 but without as many details as we now have, from a very reliable source close to one of the top two graphics companies. So it's either ATI or Nvidia. Guess away.
DirectX 9.1 will not bring the new shaders that should be implemented in hardware that will be launched at the end of next year.
Just to remind you, Microsoft decided at the last minute to accept Pixel Shader 1.1 as a standard for DirectX 8. This was a win for Nvidia since ATI's Radeon, later called Radeon 7200, had support for pixel shader 1.0 but not for 1.1.
ATI did a similar thing as Nvidia with DirectX 8.1, when it introduced Pixel shader 1.4 with more instructions and improvements meaning the Nvidia Geforce 3 card supported just DirectX 8.0 but not DirectX 8.1.
This time DirectX 9 hardware will have full support for all DirectX 9.1 feature. To clarify things for you, Radeon 9700 Pro and NV30 when it comes out will be DirectX 9.1 hardware. You won’t see any new pixel shader supported in hardware than version 2.0, which is used by Radeon 9700 PRO.
Apart from bug fixes and improvement of existing things in DirectX 9, Microsoft will introduce brand new Pixel shader 3.0.
So the new Shader 3.0 will not be supported by any hardware available in first half of 2003. We have good reason to assume that R350 and NV30 or the Spring refresh codenamed NV35 will not have support in hardware for the shaders.
You can just imagine how realistic this shader model will be if NV30 and Radeon 9700 PRO cinematic breakthrough VPU’ won't have support for it.
Best thing of all if you go out and buy Radeon 9700 Pro, NV 30 or any future DirectX 9.0 card you will be able to call it DirectX 9.1 hardware. Nice isn’t it.
If you ask why Microsoft wants to introduce Pixel Shader 3.0 before hardware supports it, it is to give time for programmers to learn about this new shader model before hardware and possible speed up its acceptation and transition process which is indeed a painful path.
http://www.theinquirer.net/?article=5329
We wrote about this API at Cebit time in March 2002 but without as many details as we now have, from a very reliable source close to one of the top two graphics companies. So it's either ATI or Nvidia. Guess away.
DirectX 9.1 will not bring the new shaders that should be implemented in hardware that will be launched at the end of next year.
Just to remind you, Microsoft decided at the last minute to accept Pixel Shader 1.1 as a standard for DirectX 8. This was a win for Nvidia since ATI's Radeon, later called Radeon 7200, had support for pixel shader 1.0 but not for 1.1.
ATI did a similar thing as Nvidia with DirectX 8.1, when it introduced Pixel shader 1.4 with more instructions and improvements meaning the Nvidia Geforce 3 card supported just DirectX 8.0 but not DirectX 8.1.
This time DirectX 9 hardware will have full support for all DirectX 9.1 feature. To clarify things for you, Radeon 9700 Pro and NV30 when it comes out will be DirectX 9.1 hardware. You won’t see any new pixel shader supported in hardware than version 2.0, which is used by Radeon 9700 PRO.
Apart from bug fixes and improvement of existing things in DirectX 9, Microsoft will introduce brand new Pixel shader 3.0.
So the new Shader 3.0 will not be supported by any hardware available in first half of 2003. We have good reason to assume that R350 and NV30 or the Spring refresh codenamed NV35 will not have support in hardware for the shaders.
You can just imagine how realistic this shader model will be if NV30 and Radeon 9700 PRO cinematic breakthrough VPU’ won't have support for it.
Best thing of all if you go out and buy Radeon 9700 Pro, NV 30 or any future DirectX 9.0 card you will be able to call it DirectX 9.1 hardware. Nice isn’t it.
If you ask why Microsoft wants to introduce Pixel Shader 3.0 before hardware supports it, it is to give time for programmers to learn about this new shader model before hardware and possible speed up its acceptation and transition process which is indeed a painful path.
http://www.theinquirer.net/?article=5329
- posting it up
