Pemalite said:
Zappykins said:
I would hope they go better than 5 Tflops for next get. Devs wanted 2.5 for this gen. Just doubling it isn't that much of a step at all. I would rather see something at least like last gen 6-10 times more powerful.
It's an effect of dimishing returns. you really need to double the power to see much of a differnce. 5 Tflops would be better, sure, but not really that noticible over current gen IMO.
It it was that big, they why would anyone game on anything but a PC?
|
You can have 1000 Petaflops of single precision floating point performance... But if you take a GPU with 10 Teraflops, but with more rops, texture mapping units, geometry engines, efficiency gains from the likes of better caching, prediction and compression... And bundle it with orders-of-magnitude more bandwidth. The 10 Teraflop GPU will win every time.
The only time flops can really be used as a "Performance gauge" is if all other things are equal, it seems this forum seems to cling to it though despite the fact even though it is highly inaccurate.
Case in point: Radeon 5870.
Even though the Radeon 5870 has 2.72 Teraflops of performance, majority of games it will loose against the Radeon 6950 at 2.253 Teraflops and the Radeon 6970 at 2.703 Teraflops (Usually by a healthy margin). And will even loose against the Radeon 7850 at 1.761 Teraflops, Radeon R7 265 at 1,843 Teraflops and so on.
The Radeon 270X will more than double the Radeon 5870's performance despite having 2.688 Teraflops verses the Radeon 5870's 2.72 Teraflops.
Graphics is more than just single precision floating point.
|
I'm not exactly sure what you were refering too. I was saying I hope next gen's concoles GPU's were more powerful than 5 Tflops.
Yes, you are correct, it's more complicated than simple, "This is 4 Tflopes, and that is only 3.4!" So this one is 60% better (cause 4-3.4=.6 .6 = 60%!!!)
Plus, your CPU and how they talk to each other is imporant too.
So what do you think 'next gens' GPU's should be?
