didnt easily compete with tahiti? huh? When one speaks of performance there are many metrics and scales you could go by. Like raw fps, calculations per watt, or even frame times. But when you talk about the GK104 and its ability to compete with tahiti *over a span of time*, I would think that there would be little doubt (i mean who could argue) that nvidia competed quite well with their gk104. When its all added up (and it should be) the gk104 was a hugely successful chip for nvidia.
If this were not true and the gk104 did not easily compete with tahiti-then AMD would currently be owning the market, outselling nvidia. Right now nvidia sells nearly 2 gpus to AMDs 1, if they were struggling to compete with the gk104 it would be quite the opposite scenario.
Performance can be many things. Performance can be measured in many ways. But when its all put together and we are talking about products, the ability to compete......well this becomes something pretty specific. It is true that a products ability to compete is the sum of many factors, the accumulative performance of many scales. Nvidia was able to compete quite well with the gk104 and i would bet money that this is one of their most successful architectures ever. One that they are extremely proud of. I just dont see how anyone would think that nvidia was not able to compete well with their gk104 this round. As far as competing goes, it would be scary if they done much better. Especially scary for AMD
I made it quite clear that my premise was based on the fact that Sirpauly used the words
easily competed implying throughout the entire life-cycle and the entire range. While Nvidia's release of GTX680 did increase price/perf and did marginally win the corwn for Nvidia it didn't last long. For the 1st time in a few GPU generations AMD not only matched but exceeded the performance of the equivalent Nvidia cards for most of the year. It took Nvidia one year after 7970GE was released to match high end Tahiti performance with the GTX770. Like it or not HD 7970 GE was ~10% faster than GTX680 out of the box and that performance delta remained similar or better as we went down the range. If we add in GPGU compute Tahiti leaves GK-104 in the dust. When GTX580 offered a similar performance delta compared to HD 6970 all the Nvidia fans hailed it as an amazing piece of kit or, "Fermi done right". Tahiti does the same thing to GK-104 and we get the revisionist "GK-104 was arguably still faster overall" when it clearly wasn't.
I am not in any way referring to market share here, simply from a stock for stock standpoint Tahiti was overall faster than GK-104 out of the box. No "arguably" about it.
Once again taking Tahiti vs Gk-104
7870XT > GTX660
7950 BE > GTX660Ti
7970 > GTX670
7970 GE > GTX680
Nvidia have utter crap power consumption = "who cares, it's performance that matters".
Nvidia have marginally better power consumption = "These new Nvidia cards are so power efficient"
Nvidia have superior GPGU = "Nvidia give more features"
Nvidia have inferior GPGU = "Who needs GPGU anyway"
Nvidia are 10% faster = "Nvidia are untouchable for speed"
Nvidia are 10% slower = "They are arguably faster"
Nvidia release a card ~27% faster for a massive premium (Titan vs 7970GE) = "You want the best you must pay for it"
AMD release a card ~27% faster for a lower price (7970 vs GTX580 3GB) = "Overpriced junk, wait for Nvidia to cream them".
I'm sick of reading revisionist history always attempt to make Nvidia the winner in all areas when the truth was sometimes the opposite.