I made an account to point out how much of an idiot you are for this comment.
A teraflop is unit of measurement and can't be different from one product maker to another. A teraflop is a teraflop no matter what platform or card used. The only difference can be how many teraflops one card can output over the other. You don't understand what or how any of this actually works and should never comment again about anything like this again.
I'm touched that you'd take time out of your busy schedule correcting people's grammar to sign up and let me know what a moron I am. I commend your commitment to pedantry. But you may have missed my point as you fell over yourself to belittle a stranger on the Internet so you could feel the smug satisfaction of another successful nitpick.
I know what a teraflop is, mate. I've known about floating point operations since they were commonly measured in megaflops, when I got a CPU with a floating point unit about 20 years ago.
I was oversimplifying somewhat to explain to the OP why his comparison was wrong and why teraflops on spec sheets are theoretical maximums and a mostly meaningless comparison in gaming. I'll remind you again of my point that a GTX 1060 (4 tflops) leaves an RX480 (6 tflops) eating its dust.
I've considered your advice and am returning it back to you with the suggestion that you stick it up your arse.
Nvidia teraflops are not the same as AMD teraflops. Compare the reported teraflops of say, the GTX 1060 and the RX480, and then compare their gaming performance.
A GTX 680 would still trade blows with an RX480, which is faster than the PS4 Pro by some way. It's not just down to GPUs either, consoles have very weak CPUs and slow hard drives which really puts a bottleneck on game world streaming and is usually the reason why frame rates are often locked at 30.
And it was pretty good