We don't know how big data will be in either case but this test was to show gpgpu and we are not given any details on it regarding Xbox numbers.

What you and others are doing is guessing what the slides are to show us but sadly don't. That's why I say there is data missing and infering anything off this is close to impossible in comparison of hardware.

Sure, it does.  System overhead is some amount of time on a core per frame.  Some of it will be in a burst like some 'ms' at the end of a frame; some of it just background tasks.

If you're doing a 5ms CPU benchmark then system overhead should be unnoticeable since the are 4 cores.  System will take one, and the benchmark should run on another with little contention.  You would see the overhead when you try to pack all 4 cores like in a full game.

The point of doing these measures like "number of dancers in 5ms" is for budgeting.  You have some budget per frame, say for Ubisoft's 30fps games the amount of CPU time per frame is (4 * 33.3ms - System_Overhead).  Their benchmark tells them how many dancers they can do per ms, but you have to count system overhead to determine number of dancers per frame.

Actually this is indeed an interesting topic to talk about as this gen we have full fledged multi-tasking operating systems with a specific scheduler at hand in both systems which, when measuring and judging performance, plays a significant role outside of pure specs which influences performance possibly greatly.

But something seems wrong with what you are trying to say: Both systems take 2 of their 8 cores for operating system purposes so I don't see why the remaining 6 cores should suffer from that, other than the scheduler does some bad work but even then this is intransparent for the remaining cores which are for game purposes.

Anyway, these are details we will probably never get to know but I do know that dev tools nowadays also consider schedulers of the current gen consoles.

In that Benchmark X1 Marginally faster CPU not because of the higher clock speed (1.6ghz vs 1.75ghz); it is due to the fact XOne CPU was using ESRAM, which only works in theory on that Benchmark(ESRAM is mostly used by the GPU). In reality, real game application, Devs already said PS4 CPU is slighty faster than XOne CPU, and the Combo CPU + GPU is massively in favour to PS4.

The Cell in the PS3 works like a hybrid CPU/GPU, is responsible of plenty of heavy task such as polygons processing, and other graphical tasks, this is why it was so powerful. A lot of its power is used for things which are usually done by the GPU.   XBox 360, in the end, was a more impressive piece of hardware, with a far better GPU, more powerful and advanced, a more balanced architecture and friendly to develop for.  CPU alone means nothing. It's the whole architecture which really matters.

Right, I don't think the other 6 cores would suffer from that.  But you're assuming that they both take 2 cores, but we don't know that.  The running assumption is that XB1 has been reserving more core time than the PS4.  How can that be if they both reserve 2 cores?  Maybe PS4 reserves 2 cores and XB1 reserves 2.5? Then overall you see how a PS4 game could do more CPU work even if the CPU is slightly slower.

Yeah, you're right , :)

............they will obliterate their competitor on graphic department and on first time ever the most console sold with the more powerful spec and  i hope we will see great games in the future with great graphic, something that could only been dreamed of when it was happend during PS2 vs Xbox original era.

:)

More info about the console capability.

http://www.neogaf.com/forum/showthread.php?t=916219

It is new and seems like PS4 (and Xbone too) have a addictional GPU compute module (outside the main GPU) that can do media tasks.

Interesting.

This was new to you? I don't expect to see much of this anytime soon, if ever.  It's totally up to both MS or Sony to use this and especially probably only for media play purposes.

But yes, at least interesting :)