New rumors about the now relentless fight between Sony and Microsoft for the supremacy of the new generation of consoles, would seem to agree with those who have preached calm and patience in the aftermath of the amazing Microsoft press conference in which the American company brought out the heavy calibers such as hyper-performance hardware and high-level system resources to reverse the trend that saw it come out defeated by the previous generation. In the following weeks there was a lot of talk about the potential of XBox Series X but the perplexities of the experts are not at all lacking and, as we will see, it seems that once again they have not been denied.
In an interview with Vigiato.net, Ali Salehi, Crytek's rendering engineer, said he had the opportunity to see the upcoming consoles very closely and what approach is needed in order to develop video games for these new platforms, revealing some pretty interesting observations on Xbox Series X, consoles that would have bottlenecks that limit its performance, effectively favoring the PlayStation 5 system.
Let's say right away that Crytek is not new to this tendency to create some controversy on the web, however before talking about the actual hardware differences, Salehi talked about the platform in general, stating that: “The developers are saying that the PS5 is the console for the which developed code more easily at peak performance. As for the software, coding for PS5 is extremely simple and has so many capabilities that developers are much more free to program than ever before. In short, I can say that the PS5 is a better console. "
But why would it be difficult to get maximum performance on Xbox Series X then? In this regard, Salehi stated that the Xbox Series X generally performs below the 12 TFLOP declared by Microsoft, as there are bottlenecks that limit the GPU and it is necessary to achieve perfect conditions of use of resources in order to use all 12 TFLOPs so ostentatious by the American company. Put simply, it is enough that the code optimization is not good enough and the performance drops significantly, making the implementation of such powerful and sophisticated hardware useless.
What the layman don't know is that the GPU is made up of various elements inside. The CU (Control unit) I'm only part of it and they are those who are assigned to process control. In the meantime, only if all other parts of the GPU are put in a position to perform at their best, without being limited, without memory bottlenecks, CUs get all the data they need in a very short time and are capable of performing 12T floating point operations per second. All of this, however, is part of the so-called "ideal case“, Or that case in which any hardware / software limitations are no longer valid, however in reality it is a very difficult condition to obtain.
To better understand this, it is useful to analyze for example the hardware of the Xbox Series X ram. Microsoft has designed a design of the RAM dividing it in fact into two parts, Same mistake made with Xbox One. One part of the RAM has a high bandwidth, while the other is noticeably lower. This duality of RAM was a source of enormous problems in the previous generation of consoles, because implementing the code of an application (for example a video game) with a similar hardware design could be very challenging. In fact, the total amount of elements that must be allocated to the most performing part of the RAM is so high that it quickly saturates the bandwidth and forces the use of that portion of memory that is able to operate with lower performance. This feature is already in itself an obvious bottleneck that developers are called upon to deal with if they want to develop optimally. Let's not forget that consoles coming out at the end of 2020 will support 4K and this will be a source of further problems. After analyzing these elements, it is clear how the GPU will be strongly conditioned by the RAM architecture.
The problems of XBox Series X however, they do not seem to end here: the Control Units are those components of the CPU that have the task of managing all the actions necessary for the execution of one or more instructions and, although there are 36 on PS5 and 56 on XBox Series X, the 36 of Play Station 5 work at one much higher clock frequency compared to the Microsoft counterpart. The choice made by Sony, or to increase the clock speed, has some advantages not only on the memory, but also and above all on the GPU burner and any other part of the GPU. The efficiency of the graphics processing unit in fact depends on the clock speed, an element not related to the count of the Control Units or Tflops.
Again, the Play Station 5 system has been shown to work more efficiently. The lower clock rates on the XBox Series X would lead to the console getting Tflops decidedly lower than the 12 declared by Microsoft, a figure reached in absolutely ideal and theoretical cases.
To these purely hardware bottlenecks, like some parts of the slower GPU due to lower clock rates and lower-bandwidth memory, the software problem must also be added, as Microsoft runs a Windows 10-based operating system on Xbox Series X, while Sony implements a lighter operating system and designed exclusively for the console.
It is therefore obvious that the more system resources needed for the operation of the operating system, the less they have available to run applications and, also in this case, the "lightness" path followed by the Japanese company seems the most functional among the two.
In light of these statements, as we have said on more than one occasion, despite the fact that the Xbox Series X has on paper a clear advantage over the PlayStation 5, in terms of hardware characteristics, it is clear that if this advantage it cannot be optimized even on the software side, it's actually like not having it. Salehi's statements have been removed from the network, probably they have greatly upset the Microsoft top management, however they are absolutely not new things and that have already happened in the past. In the generation Ps3 / XBox 360, it seemed that Sony had the most powerful console but then in practice the applications were easier to develop on the Microsoft platform, for example.
Despite this, however, we must not be influenced by certain analyzes because until the products are launched on the market there will never be an effective response. Let us remember that both consoles will have the x86-64 AMD Ryzen "Zen2" chipset which is a very well known chipset and that the main aspect on which the game will be played in the new generation of consoles will undoubtedly be theapproach to streaming with its high-speed data stream.
