derpysquirtle64 said:
My point was is that stating "just 7 tflops more" as something that doesn't matter, doesn't make any sense. If we had GPUs with hundreds of TFLOPs of processing power, then just 7 tflops more wouldn't have been an issue. But here we have an increase from 13 TFLOPS (2080 Ti) to 17 TFLOPS (3070). Doesn't seem like a small deal. And if we even don't use TFLOPs as a measure, are you sure you watched the same Digital Foundry vid that I did? 3080 has around 80% increase over 2080 in real performance in almost any game. That's kinda huge in my opinion. |
But that 7 tflop advantage is not an 80% gain you are comparing. The 80% advantage from the 2080 to the 3080 wich is a 10.7 tflops to 30 tflops. 3x on paper to only 1.8x on real world performance. From 215 watts to 320 watts so 1.5x. They claimed a 1.9x performance per watt while we see a around 20% real world increase.
All im saying is those paper specs they are claiming are not shown in the real world performance gained showed by digital foundry. And at the end of the day is the real world performance they have to beat, not paper spec and the bar was set along the promised goal lisa su promissed.
Edit: On paper the 3080 is 33.33% better than a 3070. so a quick correlation would sugest the 3070 is 20% than a 2080. We obiously need to wait on benchmarks. but if so, I see the series x matching it.
Last edited by eva01beserk - on 01 September 2020It takes genuine talent to see greatness in yourself despite your absence of genuine talent.







