tviceman
Diamond Member
On a side note, many here thought GK110 wouldn't even see a Geforce release. And when it did, many didn't think it would ever see GK110 in the $400 price range.
I think that might be a tell-tale sign. GK104 is arguably the most successful chip for Nvidia since G92. Who's to say Nvidia might not push the mid-die strategy even harder, releasing a slightly more competent chip to be it's Geforce highest end, while stripping out graphic-specific transistors in it's big daddy chip to save on die space and/or to more efficiently compete with Intel's looming threat? That, to me, seems very plausible. I'm not chip engineer, but how many transistors and mm^2 could have been saved in GK110 if it was compute only? Perf/watt would have gone up, die space would have gone down, Knights Landing would have looked even more retarded.
I think that might be a tell-tale sign. GK104 is arguably the most successful chip for Nvidia since G92. Who's to say Nvidia might not push the mid-die strategy even harder, releasing a slightly more competent chip to be it's Geforce highest end, while stripping out graphic-specific transistors in it's big daddy chip to save on die space and/or to more efficiently compete with Intel's looming threat? That, to me, seems very plausible. I'm not chip engineer, but how many transistors and mm^2 could have been saved in GK110 if it was compute only? Perf/watt would have gone up, die space would have gone down, Knights Landing would have looked even more retarded.