[my highlighting]The bandwidth "requirement" for the 06 ST fill rate test is pretty clear cut, if you test it out:
It would be nice to see what the drop off values are like with chips that have more blenders than the G80. I wish more review sites would concentrate on analysing one variable over one test, rather than doing several over numerous.Code:G80 @ 1280 x 1024 Core/RAM ST Theor %Diff 576 / 1000 6905 6912 -0.1% 576 / 950 6905 6912 -0.1% 576 / 900 6905 6912 -0.1% 576 / 850 6903 6912 -0.1% 576 / 800 6898 6912 -0.2% 576 / 750 6883 6912 -0.4% 576 / 700 6725 6912 -2.7% 576 / 650 6342 6912 -8.2% [B]576 / 600 5877 6912 -15.0%[/B] 576 / 550 5354 6912 -22.5% 576 / 500 4939 6912 -28.5% 576 / 450 4421 6912 -36.0%
That's as high as i can get my GTX280 as far as bytes per pixel is concerned with default ROP-Domain of 700 MHz. At 8,02 bbp (e700/m1404), I'm getting about 83 percent blend efficiency.
Lowering the e-clk to 500 MHz yielding a neat 16 GPix/s of INT8 blending, I can get almost 98 percent utilization.
So, wrt to fullspeed blending - even in INT8-case - GT200 seems to be quite BW starved. If need be, I can post the whole results.