michael_stutzer said:
Most PCs (and PS4) have the power required for those multiplayer games. Xbone doesn't have enough hardware power to play modern games. Hence the cloud processing is required to play those games. This game is 6vs6 with probably small-medium sized maps. If the cloud processing wasn't there the game would be 3vs3 without AI on even smaller maps. The other platforms does not need cloud processing. Xbone needs it. |
Oh boy...where to begin?
Listen, I don't want to come off like a d***che, so I'll just politely say that your understanding of the MS Cloud service and what it will potentially add to the Titanfall gaming experience is, um, very very optimistic but not based in any known reality.
A server (any server, regardless of it's horsepower) that is accessed by your Xbox remotely via an Internet connection (even the world's fastest multi-gigabit fiber optic connection) is going to be orders of magnitude slower in sending data back and forth than what your console is capable of doing internally between its GPU/CPU and on-board DDR3 RAM. That's just the way it is.
There is NO POSSIBLE WAY that the MS cloud servers are going to perform any real-time AI or graphics processing in-game for Titanfall. Graphics and AI processing requires data I/O throughput between the GPU/CPU and memory that is easily 1000x times faster than any Internet connection is capable of providing now, and for the forseeable future.
To illustrate the difference in data speeds I'm talking about here, I'll use an analogy about reading a book:
Picking up a paperback copy of The Hobbit and reading it page by page is like Titanfall running on the Xbox: your brain and eyes are the Xbox "GPU/CPU" and the text in the book is the graphics data that you "read" and process. You can read the book quite efficiently this way because your CPU (brain/eyes) can very quickly read the data (book text) because you have the book in front of you.
Now, suppose you want to offload, say 20% of your reading "processing" to the "cloud". In this case the cloud is your town library, located 3 miles from your house. The cloud (library) also has a copy of The Hobbit.
In this cloud-processing scenario you pick up your local paperback copy of the book in your house and start reading normally, just like the first scenario. However, since you're now offloading 20% of your reading processing to the cloud you only read up to page #8, then you put down the book, get in your car, drive to the library, find The Hobbit on the shelf, turn to page 9, and read it and page 10. Then you put the book back on the shelf, get back in your car, drive home, and pick up your copy of The Hobbit and continue reading pages 11-18. Then, get back in your car for a trip to the library to read pages 19 and 20. Etc, etc, etc, etc.
You can see how that second scenario would take you WAAAY longer to finish Bilbo's story, right? The "latency" of driving back and forth to the library just to read 2 out of every 10 pages is just not efficient. As silly as my analogy is, the relative speed differences between the library commute vs. reading the book at your house is about the same as the difference between cloud processing via Internet connection vs. the Xbox GPU/CPU retrieving data from its 8Gb of DDR3 RAM.
And that, gentle readers, is why real-time AI and graphics processing in the cloud just ain't a reality.
