vivster said:
JRPGfan said:
or atleast (6/4.2) -1 = 42,8%
The PS4pro does have that "rapid path math" thingy, which if they code for, could speed up things abit.
Differnce between the two is probably less than 42%.
|
Of course it is. Real world performance is probably 30% if they're lucky. Just don't tell the peasants.
|
Rapid packed math isn't a doubling in real-world performance anyway, so you are right there.
It's all well and good to throw out theoretical numbers that will never be achieved in the real world...
But, not all packed "math" can actually be packed together, only compatible math, which is the big caveat that most people don't mention.
In a worst-case scenario you could have performance gains as low as 20% or as much as 70%.
But never a doubling.
Plus it can't be used for everything anyway, it is lower precision.
rbej said:
vivster said:
Of course it is. Real world performance is probably 30% if they're lucky. Just don't tell the peasants.
|
Buhaaaaa. You are joking with 30%.
Rise Of Tomb Raider, F1 2017, Killing Floor 2 - every this game have 2x more native pixels on X1X and textures in higher resolution with better AF filtering. This is 30%??. Time to wake up blind Pony. This is different higher than PS4 Pro vs PS4, when PS4 Pro have only 1.4-2x more native pixels than PS4 and this same textures and assets.
|
Er. He was talking about half precision and single precision floating point math.
You do realise there is more to rendering than the amount of flops, right?
Fact is... Half precision floating point on the Playstation 4 Pro exceeds the Xbox One X's Single Precision. And that's fine.
It doesn't make the Playstation 4 Pro faster than the Xbox One X.