Hellbinder
Banned
You guys can take this or leave it. I am not trying to infer anything here, be a fanboi (like I usually am ) etc..
I have heard this whispered on the wind.. And I have no link or source to offer. I am just throwing this out here so that there is a record of it once the Nv30 goes public..... official like... Just on the outside chance that some of this is correct.
1. The Nv30 is a 4x4 architecture. /although in 32bit mode it operates at 4x2
2. It was initially designed with a 128bit memory controler.
3. It *could have* but not definitly been upgraded to a 256bit controler. However it is very likely that it is still at 128bits. (thus the absolute necessity for DDR II)
4. Much of its great speed is actually due to some interesting approaches it takes to accomplish stuff, not due to an abundance of pixel pipelines.
5. Its huge transistor count is not due to an 8x2 design like we have all assumed. It is becuase of the complexity of the pixel/vertex shaders and becuase they have included a True Fixed Function TCL unit for backwards compatability (on top of its very advanced vertex shaders)..
The end.
No I did not make the above info up.. However thats all i am going to say. It will be interesting to see how this all pans out in a few more weeks.
I have heard this whispered on the wind.. And I have no link or source to offer. I am just throwing this out here so that there is a record of it once the Nv30 goes public..... official like... Just on the outside chance that some of this is correct.
1. The Nv30 is a 4x4 architecture. /although in 32bit mode it operates at 4x2
2. It was initially designed with a 128bit memory controler.
3. It *could have* but not definitly been upgraded to a 256bit controler. However it is very likely that it is still at 128bits. (thus the absolute necessity for DDR II)
4. Much of its great speed is actually due to some interesting approaches it takes to accomplish stuff, not due to an abundance of pixel pipelines.
5. Its huge transistor count is not due to an 8x2 design like we have all assumed. It is becuase of the complexity of the pixel/vertex shaders and becuase they have included a True Fixed Function TCL unit for backwards compatability (on top of its very advanced vertex shaders)..
The end.
No I did not make the above info up.. However thats all i am going to say. It will be interesting to see how this all pans out in a few more weeks.