HollyGamer said:
sc94597 said:
HollyGamer said:
sc94597 said:
Yet an r9 270 still slightly outperforms a PS4, anyway. So your point doesn't hold in this particular case. Usually you cant compare gflops when the other parts of the gpu hardware are quite different. The r9 270 and PS4 gpu are both based on the same chip though. And in all other areas the r9 270 is slightly better. The difference is that of running games at medium at 1600 *900 and running them at medium-high 1080p with some AA. If you overclock the r9 270, which literally only takes a button click, then you see even greater performance gains for the same price you paid.
|
It's different from one game to another tho, i have the problem with ACU, watch Dog with My Nvidia 880M GTX, i understand it can be slightly benefit on greater specs, but it seems most games run almost the same graphic because most developer (multyplatform games) build on the same engine that can be use for multyple configuration and specs.
|
Not for the r9 270 though. Your 880M GTX is much more of a different card with different advantages and disadvantages than an r9 270 is from a PS4. All are within the same ballpark performance, but the R9 270 is repackaged Radeon HD7870 with a higher clock speed, while the PS4's GPU is an underclocked HD 7870.
http://gpuboss.com/gpus/Radeon-R9-270-vs-Radeon-HD-7870
They are essentially the same gpu in everything but clock speed. The only benefit the PS4 has is that its memory is unified GDDR5, and that means it won't be constrained by VRAM in future games, but r9 270 should be fine with 2gb of Vram for the rest of the generation.
|
Well, i said many times, PC need to run Direct X, and many coding layer for compability, and the design is very general, PS4 benefit not just unified RAm but GPGPU technology, ACE unit 8x then the 7870 , low level code on metal level, and optimization which is imposible on PC.
Read my comparison on Ferrari VS GT Club, you cannot compare the spec directly interms of GLOF (TFLop) or GPU clock etc.
|
I was saying in this case we have real-world performance advantages, however. Empirical data to draw from. So while, yes, you can't just compare Floating Point Calculations, there is an advantage nevertheless, which we can assume comes from the extra power (clock speed) considering the archicture of both chips is nearly identical. Optimization is not impossible on PC, it is just - as you said - more general. If it were impossible, Nvidia (or AMD) wouldn't release drivers with descriptions of "100% faster speeds in video game x" as they do. Or developers wouldn't release patches that improve game performance. As for GPGPU, most games don't even use it, and the ones that do seemingly provide no graphical advantage for performance. I highly doubt many developers will off-load resources and time in multiplatform games which have already been coded to add GPGPU features. In exclusive games you have a point, but not multiplats.
