Major Nelson on X1 Power: "I can't wait for the truth to come out."

[Shinobi-san]

Machiavellian said:

Shinobi-san said:

Machiavellian said:
Remember the Cell SPUs are considered co-processor and they were used to offload task away from the PS3 GPU.  You already see the results of how well that worked out for Sony (at least for their 1st party). Depending on what task the co-processors are used for, they can make a solid difference in performance.  Without any info on what the processor do, I will not speculate as to the difference they can make only that the possibility exist.  

Lets break down the difference between the X1 GPU and the PS4.  I will limit this to just the CU or Sharder cores as that’s were the big difference in TFlops comes into play.  The X1 has 12 CUs and the PS4 has 18.  CUs are generally used to process shaders.  Since that’s a parallel process, the PS4 can execute more shaders at one time then the X1.  Now from the Hotchips convention some interesting things came out and I was processing the information and thinking about design.  MS has made it so that all parts of the system knows what is happening to a segment of code in memory.  If that is so, then MS can leverage specialize co-processor to handle specific gaming code where it would be more process intensive for the GPU or CPU to handle.  

 

Do not forget that GPUs are designed as an add on to a system not as being the main CPU.  What I am getting at is that for efficiency within a closed box, some of the things the GPU does might make more sense to handle with specialize hardware then within the GPU.

As for Albert statements, he is a PR guy so you will always take their comments with a grain of salt.  Interesting enough him commenting on Neograf where he knows his comments are going to get a lot of push back.  Most PR people know where to pick their battles so him making his statements on Gaf says that either he loves contention or MS has a few tricks up their sleeve.

As for Tflops, do you know that the PS3 is stated as having 2.1 TFlops compared to the 1.8TFlops of the PS4.  As a measuring tool for the performance of these 2 devices, the Tflop number really might not be the difference maker.

the tflops number becomes more reliable and worthwhile comparing when the systems have near identical hardware...

We not comparing a Nvidia GPU core to an AMD core...they both have GCN cores. They both have the exact same CPU. Those are the two core components of the system.

And yes we dont know that much about the co-processors but again it cant replace conventional CPU's or GPU's....using the cell as an example is not exatcly a positive thing. The SPU's were only good and well utilised under very strict circumstances. When devs started coding on the PS3 they flat out ignored the SPU's. These co-processors seems to have a set function...so it doesnt give the dev much flexibility either.

But agian i feel like you tring to argue that these co-processors will have a great impact on increasing overall system performance. Whereas, i see it more as a way to maximise the efficiency of the system. And when comparing the two systems...the PS4 also has co-processors or is that a non issue? I've said this three times now but you never actually address that. Do you not agree or ?

I will agree that the Tflops makes much more of a difference when comparing like for like hardware.  In this case the GPU in the X1 definitely a much weaker part to the one in the PS4, no doubt about it.  The CPU is something different.  Thought the parts are the same, the difference between bandwidth and access are different in the X1 to make for a much superior CPU.  The area where thing changes is the other parts of the system we cannot get a direct comparison.  The Shape Audio processor pretty much mean audio is free on the X1.  Meaning the CPU doesn't contribute any processing for Audio. Thus its free to do other things.

I am not sure why you keep saying co-processor cannot replace conventional CPU/GPU.  They are not used that way, they are used to offload specific task that a CPU or GPU would do but more efficiently.  The Shape Audio processor is a good example.  Its designed to offload all audio processing from the CPU where it has specialize co-processors of its own to handle audio task.  By itself its a very powerful audio combination that will handle audio functions much better and faster than the CPU.  

The reason developers did not ramp up to the Cell fast enough was Sony fault.  It was Sony that did not provide enough information to developers on how to use the Cell chips until after the PhyreEngine.  At the beginning of the PS3 gen, Sony was very poor with providing tools, libraries and sample code on how to use the Cell.  When multiplat games started to look worst from developers, it kicked Sony into gear that they needed a multiplatform game engine to get developers up to speed faster on the Cell.  

You are right that the PS3 does have co-processors of its own I just do not think they are used for anything besides Audio and decompression and compression video.  I have no clue what MS is doing with the additional processors, No one does.  What I am stating is that without any info, there are many possibilities. 

My argument is that MS made a statement that is hard to define just by only comparing the GPU specs.  There are many different ways to skin a cat and co-processors are one.  You cannot dismiss their importance until we find out what they are used for.  I have a few ideals light lighting, physics, and animation but we shall see.

You seriously believe Xbone has a much more superior CPU than the PS4? I mean seriously? This conversation was pretty level headed until that comment. Not sure where that comes from. But im not even going to address that because then thats a whole other argument that i think is pointless. You do realise we dont even know the clock speed of the PS4 cpu right?

@ bolded im saying they can only be used for the task they are meant to be used for. It does not replace a GPU or CPU which can be freely assigned to any task thats appropriate. Also, under no circumstances do i think the SPU's were better than either a GPU for GPU specific tasks or  a CPU for CPU optimised tasks. Like I said using the cell as an example is a bad one. And like you said we dont know much details about these co-processors....so how do you know it does the job better than what the CPU could have done it? We literally have no details on it. Other than the fact that certain tasks are offloaded to it to free up CPU and GPU resources. Thats it. I mean the entire premise behind it explains it all...why would they try and free up resources for the GPU and CPU if those other processors did a much better job?

This makes zeros sense to me. I think we overblowing this. And indeed when we get more info on these co-processors we can revisit the topic.