G200-302 Rev A2 begun manufacturing back in September, and the first parts are now finding their way to mass manufacturing. The chip features a die size of 470mm2, 107mm2 less than the original G200 chip. This just goes to show the vast difference between 65nm and 55nm - if Nvidia had the balls to go with 55nm chip back in May, the prices of GTX260/280 parts could have been way cheaper and offer much more flexibility, but we can’t cry over spilt milk. 55nm part is here now, and it will consume much less power than is the case with the 65nm one.
The 55nm GPU consumes roughly 50% less power than it was the case with 65nm one, and this difference is more than massive. When I did quick power checks, the GTX280 at 650/1500/2200 would eat around 266W, while the default clocked GTX280 (600/1300/2200) was specc’ed at 238W.
Well, the 55nm GPU will eat around 170W at 650/1500/2200, meaning that GTX290 just got 100W of power to play with. If you’re into overclocking, you can now start dreaming about clocking those 240 shaders to 1.7-1.8 GHz range (perhaps even 2.0 if water-cooling setup is powerful enough), and achieve massive performance gains, all happening while you’re consuming less power than a stock clocked GTX280.
zu zeigen.
