PS4 vs. XBOne performance compared, using representative PC hardware
The Xbox One and PS4 have very similar hardware specs, except for two key areas: The GPU, and the memory architecture — not the amount of memory, but the way in which games access that memory. At launch, the different memory architectures probably won’t be significant — but after three or four generations, as developers master the eighth-generation console hardware, one of the consoles might come out ahead. When it comes to the GPUs, however, there’s a clear-cut leader: the GPU in the PS4 has 50% higher peak performance than the GPU in the Xbox One. As you probably know, though, peak performance isn’t an indicator of real-world frame rate improvements or image quality — that’s why Digital Foundry set out to replicate the Xbox One and PS4, using PC hardware, to see whether the PS4 actually has the edge.
Now, these being consoles, we don’t have their exact hardware specifications, but we have a very good rough idea. Both the Xbone and PS4 have an eight-core AMD Jaguar-based CPU at 1.6GHz, with 8GB of RAM. While the Xbox One’s GPU is comparable to the Radeon HD 7790, the PS4 GPU is more like the HD 7870 — which has 50% more compute units, and thus 50% higher peak teraflops.
Instead of using this exact CPU and GPU setup, though, Digital Foundry opted for a super-fast CPU (Core i7-3770K @ 4.3GHz), 16GB of RAM, with the Xbox One being represented by a Radeon 7850, and the PS4 represented by a Radeon 7870 XT. This might seem a bit silly at first glance, but because the console GPUs have two fewer compute units than their desktop equivalent (the 7790 has 14 compute units, while the Xbox One has 12), a direct comparison with PC hardware isn’t possible. The GPUs that Digital Foundry chose maintain the PS4′s 50%-more-compute-units advantage, ensuring that the performance difference between the simulated consoles is still representative (though you should discard the actual frame rates). A fast CPU wand lots of RAM was used so that the test was purely about GPU performance.
With these representative rigs set up, Digital Foundry ran a variety of game benchmarks and recorded the different frame rates posted by the simulated consoles. Depending on the game, the PS4 was between 17 and 33% faster than the Xbox One. Not quite 50%, but still a fairly sizable lead for the PS4.
The real-world performance difference between the PS4 and Xbox One is likely to be much more complex than Digital Foundry’s basic test. The PS4 and Xbox One have radically different memory architectures, which will probably give the PS4 the lead early on, but performance might swing towards the Xbox One as developers learn how to use its ultra-fast SRAM effectively. Of course, with the PS4 having a unified memory architecture, the performance crown could swing even further in Sony’s favor.
It’s also worth noting that, ultimately, there might not be any real-world difference at all. Most games are developed with cross-platform, lowest-common-denominator compatibility in mind. Developers are unlikely to create a game that runs well on the PS4, but chugs along on the Xbox One and PC. As long as both consoles can run modern games at 1920×1080 (1080p) @ 30 fps, which the Xbox One and PS4 are more than capable of, you probably won’t notice any difference at all.