Originally posted by: otispunkmeyer
Originally posted by: Cookie Monster
Just a quick summary
G71 (7900ultra)
90 nm low k
32 pixel pipelines
32 pixel shaders
32 TMUs (texture mapping units)
16 ROPs
10 Vertex shaders?
256bit GDDR3 (ext.)
380~ transistors?
core clock 750mhz
memory clock 1900mhz.
R580 (X1900XT)
90 nm lowk
16 pixel pipelines
48 pixel shaders
16 TMUs
10 VS
256bit GDDR3 (512bit int. ring bus)
350~ transistors? (R520 has 321~ transistors)
core clock 695mhz
memory clock 1550mhz
AS you know this is the specs for the X1600XT (R530)
90 nm lowk
4 Pipelines
12 Pixel Shaders
4 Texture Mapping Units
4 ROP's
5 Vertex Shaders
128Bit Ext. Memory Interface, 256Bit Int. Ring Bus
So if you compare this with the R580, it will be R580 = R530 x 4
i honestly fail at these things
so nvidia is "going" to have 32 pixel piplines, i presume then they have 1 pixel shader and 1 TMU each = 32 pixel shaders and 32 TMUs which to me seems nice an balanced (but im uneducated so im very probably wrong) and it makes sense....ie 1 fits into 1
then they have 16 ROP's which i know means despite being able to produce 32 pixels in one go, they can only draw 16 of them. but this isnt a problem since "apparently" that situation rarely arises
Ati is "going" to have what i see as crazy. 16 pixel pipelines....but then 48 pixel shaders? so they put 3 pixel shaders in the pixel pipeline? (3 fits into 1?) why doesnt that make it 48 pixel pipelines?
then they got 16 TMU's and 16ROPS.
can some one explain.