You can't scale flops exactly. ANd 933 Gflops IS the number - its calculated by 3 (2 MADD + 1 MUL) * 1300 Shader Clock * 240 SP = 933 GFlops
But on the other hand, apparently G80 MUL wasn't working 75% of the time so it was closr to 2 * 1500 * 128 = 350ish Gflops rather than the theoretical 581
Also, thats why you cant compare numbers directly over different architectures, even within derivatives (such as G80 Ultra vs. G92 GTX, where clearly the G92 GTX architecture has some memory bottlenecks and performs up to G80 ultra performance at higher resolutions, and even loses some when AA and other things are on at higher res)
http://www.fudzilla.com/index.php?op...=7523&Itemid=1
If fud is right, we're looking at 600mhz for the core, which seems pretty high considering how large the die is, definitely going to run very hot. That, and that means what the inquirer has said is impossible to happen because 1296mhz just can't happen now for the shaders (unless we're looking at something like a 2.xxxx shader domain lol) and it never could for the core, plus I think its safe to say gt200 won't be spanked by the r700
Yep as LordEC911, CJ had already stated those numbers before Fud ever got a hand on them, and more than a few sources besides TheInq has quoted them.
And that's how the math for counting flops has been done for some time. Lower clocks doesnt mean anything bad - they're just saving us from excessive heat and power draw and the fact that these shaders might have the full 3 operations / sec and you have 240 of them means that you will still get tons more performance from shaders
Same price as 8800GTX launched at IIRC, and any other previous high end pieces of hardware.
And with 50% morre efficient shader units, do they mean 50% improvisation in performance per transistor, or 50% higher transistor count, as in more raw power? If the latter was true, wouldn't the maths have to be adjusted for the flop-count, making it well beyond 1 Tflop?
Last edited by Nuker_; 05-26-2008 at 02:17 PM.
And how much performance will able to bring "the missing MUL component" ?
This post could probably tell you something: http://forum.beyond3d.com/showpost.p...postcount=1548
Damn power consumption is just increasing and increasing.![]()
All systems sold. Will be back after Sandy Bridge!
80 ROPs![]()
BTW what do you guys mean the MUL wasnt working 75 percent of the time? Do you mean the third shader can do additional operations now?
Seems we made our greatest error when we named it at the start
for though we called it "Human Nature" - it was cancer of the heart
CPU: AMD X3 720BE@ 3,4Ghz
Cooler: Xigmatek S1283(Terrible mounting system for AM2/3)
Motherboard: Gigabyte 790FXT-UD5P(F4) RAM: 2x 2GB OCZ DDR3 1600Mhz Gold 8-8-8-24
GPU:HD5850 1GB
PSU: Seasonic M12D 750W Case: Coolermaster HAF932(aka Dusty)
Yep essentially the shaders can do all full 3 operations rather than the 2 previously.
Also, it's not 80 ROPs, its 32 ROPS.
Like G80 architecture, its 4 ROPS per 64-bits of memory bus hence 512-bit bus -> 32 ROPS, 448-bit bus -> 28 ROPS
And it looks like the ratio is 24 SP's per 8 TMU's as well which isn't quite the 2 to 1 for G92 but better than the 4 to 1 on G80 (though it was 32TA, 64TF)
[Review] Core i7 920 & UD5 » Here!! « .....[Review] XFX GTX260 216SP Black Edition » Here!! «
[Review] ASUS HD4870X2 TOP » Here!! « .....[Review] EVGA 750i SLi FTW » Here!! «
[Review] BFG 9800GTX 512MB » Here!! « .....[Review] Geforce 9800GX2 1GB » Here!! «
[Review] EVGA GTX280 1GB GDDR3 » Here!! « .....[Review] Powercolor HD4870 512MB GDDR5 » Here!! «
whats the story with the 260 shader clock, some sources say 1240 others say 999
Isn't it kinda weird that there is virtually no reviews with definitive numbers for a graphics card that is scheduled for a July release date?
will we get some benchmarks this week ?
For my part I know nothing with any certainty, but the sight of the stars makes me dream.
.
.
Bookmarks