It is... nobody can deny that... even MS.
/thread.
Came back to see what kind of crazy conjecture both sides have come up with in this thread. Looks to be the same blah, blah, blah about theoretical performance and the one developer said BS. You can NOT definitively say anything because it is all guess work by ANYONE. Devkits are not final hardware, and mocking up similar spec PC's is not an answer either. Both systems have been customized in ways that we may never really know without actual parts designs in hand. It seems obvious from the generic info provided to date, that the PS4 is more powerful somehow. The key word is "seems". Like I and many others have stated, you will definitely have to wait until a minimum of after launch to even have a partial understanding of the true power of each system.
1.84 TFLOPS vs. 1.31 TFLOPS is fairly cut and dry I would say.
http://metro.co.uk/2013/08/28/sony-jim-ryan-interview-giving-away-stuff-on-day-one-its-kind-of-odd-3939866/
Just read all of that. PS4 is stronger by a non trivial amount.
| JoeTheBro said: http://metro.co.uk/2013/08/28/sony-jim-ryan-interview-giving-away-stuff-on-day-one-its-kind-of-odd-3939866/ Just read all of that. PS4 is stronger by a non trivial amount. |
"What is your greatest weakness?"
"..." *deep in thought*
[whispers] "...Kryptonite."
#dead
"I don't think its cut and dried that the PS4 is more powerful than the Xbox One."
Than your wrong. It's a fact. Enough said...
| Soundwave said: 1.84 TFLOPS vs. 1.31 TFLOPS is fairly cut and dry I would say. |
That's not the whole picture at all though.
Peak bandwidth is now 204 GB/s+68GB/s=272GB/s for XB1 vs 176 GB/s for PS4. Plus, the XB1's bandwidth as we all know is feeding a weaker GPU that needs less BW to begin with, so the RELATIVE BW edge is even greater. For every Gigaflop XB1 now has something like DOUBLE the peak BW as PS4.
People seem to overlook that a lot but BW is one of the most important GPU specs right up there with computing power.
Anyways the situation is any given system has so many complex parts looking at one spec or even ten specs is not going to give us the full story. XB1 has other advantages, a better audio chip that will take load off the CPU for example, and the CPU edge will be even greater if/when devs start using the cloud for at least AI.
And MS said there are 15 helping processors in XB1. Some of that is probably PR, but it could be a situation where those processors take so much load off the GPU that the GPU only has to focus on one thing and becomes much more capable.
I absolutely cannot wait for the first multiplatform comparisons!


fallen said:
Peak bandwidth is now 204 GB/s+68GB/s=272GB/s for XB1 vs 176 GB/s for PS4. |
Ars Technica
Similarly, the company claimed that there was more than 200GB of bandwidth within the system. Again, the number had no context or clarification and if rumors are to be believed, it suggests some rather creative accounting: 68 GB main memory bandwidth, 102GB bandwidth to an embedded SRAM buffer for the GPU, and 30GB bandwidth between the CPU and GPU. While that does add up to 200GB, there are no two parts of the SoC that can communicate with each other at 200 GB/s.
AnandTech
According to their data, there’s roughly 50GB/s of bandwidth in each direction to the SoC’s embedded SRAM (102GB/s total bandwidth). The combination of the two plus the CPU-GPU connection at 30GB/s is how Microsoft arrives at its 200GB/s bandwidth figure, although in reality that’s not how any of this works.
TechPowerup
Xbox One chip is a slightly different beast. It uses the same eight "Jaguar" 1.60 GHz cores, but a slightly smaller Radeon GPU that packs 768 stream processors, and a quad-channel DDR3-2133 MHz memory interface, which offers a memory bandwidth of 68.3 GB/s, and holding 8 GB of memory. Memory between the two subsystems are shared in a similar way to PlayStation 4, with one small difference. Xbox One chip uses a large 32 MB SRAM cache, which operates at 102 GB/s, but at infinitesimally lower latency than GDDR5. This cache cushions data-transfers for the GPU. Microsoft engineers are spinning this off as "200 GB/s of memory bandwidth," by somehow clubbing bandwidths of the various memory types in the system.