Do we really need PS5 Pro or Xbox consoles optimized for this generation?

PlayStation 4 launched in November 2013 and just three years later Sony followed up with the release of both PlayStation 4 Pro and PlayStation VR – so what are the chances of repeating the same date for the latest generation of consoles? We know a new VR headset is on the way, but what about the PS5 Pro, or its Xbox Series equivalent already? Rumors of enhanced consoles have been circulating for quite some time now, while a recent press event from TV manufacturer TCL indicates that “gen 9.5” consoles will launch in 2023/2024 – so what’s the likelihood of that actually happening?

There are a number of reasons to bet against the arrival of improved consoles — not least because we’re still mired in a long, endless transmission through genes that looks likely to extend into the third year of this generation. The PS4 Pro and Xbox One X are also designed to support a new breed of Ultra HD displays and there’s no evidence to suggest that 8K or even 4K at 120Hz has much chance of justifying more powerful hardware. Perhaps most importantly, there is the issue of economics, the idea that Sony and Microsoft may not find the way to an affordable console in current conditions, while rampant inflation, the cost of living and the specter of recession mean less money in the public’s pocket.

Then there’s the Xbox Series S. The reason for its existence is unusual, and was told in depth by Xbox Chief Architect, Andrew Goossen, in 2020. Simply put, the reason Microsoft released two consoles instead of one is that the idea of ​​no longer cutting The cost of Xbox Series X over time seems viable. Faced with this startling reality, Microsoft launched the call for a cheaper, less capable machine from day one. In fact, we got the mid-generation “Slim” update along with the top-tier model simultaneously because cost reductions over time don’t happen. In fact, in the current climate, inflation makes controllers more expensive. Meanwhile, chip maker TSMC is raising prices for silicon chips and demanding insurance premiums from some of its biggest customers. It’s a far cry from the cost reductions we’ve seen in the past.

Video article by Rich Leadbetter on the question of whether consoles optimized for this generation are desirable or economically viable.

“The Series S has had a huge impact for us. As we design our new consoles for the new generation, we look a lot forward across the generation to think about the future – like, how does this work? – which is why we have to have two consoles at the same time, Andrew Goossen told us in a big Series S interview. “We’re experiencing a huge change in how consoles are designed. I think when we first started building the original Xbox 360 — the smallest device without a hard drive — it cost us about $460. By the end of the generation , cost us about $120 – and this cost-cutting path was primarily driven by lowering the cost of silicon.”

Advances in chip manufacturing technology are advancing – 6nm, 5nm and even 3nm nodes look good. This means that similar to the shift from 28 nm to 16 nm we saw in the last generation, chip designers can still integrate more transistors into smaller chips – which is exactly what made the PS4 Slim, Xbox One S, PS4 Pro and Xbox One X viable . Improved performance and better power characteristics are available, which means we’re still ready to see progress in PC CPUs and GPUs. However, there is a catch: shrinking the chip no longer comes along with a cheaper chip.

“what or what [the new process are] Not bringing in more is a good cost reduction, the cost per transistor – so this has key implications for controller development, because now we’re going to get cost reductions, but they’re slowing down and it won’t be nearly as much as we’ve seen before,” Jossen explained. ” And that was another reason we felt we really had to do the Series S in the beginning because we had to design for the future. For the first time, we had to get an entry level controller initially. Previous generations were kind of easy because at the beginning of the generation, you’d make something really expensive — put as much silicon and perform as you can — and then you would ride the cost-cutting curves down to mass market prices. This doesn’t exist anymore.”

The last generation of the console was expressed in silicon terms. The timeline begins with 28nm firing processors on the left, followed by smaller, more efficient 16nm chips in the middle, opening the door for the Xbox One S and PS4 ‘Slim’. Finally, on the right, the same 16nmFF process used to create the enhanced consoles – more logic in a similar area space to 28nm launch chips.

The image above shows the situation as it was in the last generation. With the PS4 and Xbox One launching in 2013 at 28nm, we saw a pair of fairly large processors. With the move to 16nm FinFET in 2016, the cost-per-transistor cut opened the door to two scenarios – the smaller, cheaper “skinny” controller and the improved “Pro” model. You’ll notice that in general, the improved 16nm processors were the same size as the original 28nm processors. A $399 cost-per-transistor reduction made the PlayStation 4 Pro possible. Using Microsoft’s more ambitious console with its massive chips and 12GB of memory entailed a price hike to $499 a year later, but the same principle was applied.

Now, what if the cost per transistor remains constant? The only way to release an improved console is to raise the price even more in a world where Series X and PlayStation 5 are more expensive than previous generations. In short, the same reasons Microsoft built the Xbox Series S make the concept of launching an improved model that much more challenging. If the platform holder can’t see a way to make the X-Series cheaper in the mid-generation, it stands to reason that a more powerful controller with twice as many transistors wouldn’t be possible at $499. Far from that price, the concept of the console as a flagship product doesn’t start to make sense.

So Andrew Goossen’s comments, of course, are now two years old – was he wrong in his assessment of the cost per transistor? The key is to look at the equivalent markets where the competition is intense – and there is no better example than computer graphics. Even excluding the mining boom, the hashrate recommended by the manufacturers is only moving in one direction. The Radeon RX 5700 XT launched at $399, replaced the 6700 XT at $479, and the next 6750XT upgraded at $549. Goossen talked about the improved power and performance characteristics, and we’ve already seen faster clock speeds and higher power draw – which is good for selling PC components but not particularly compatible with the fixed console design.

If a “slim” processor for a cheaper console isn’t feasible later in the generation, what’s the alternative? Shrink the GPU instead! Here’s how the Xbox Series X processor (left) compares logically to the Series S equivalent (right). They are effectively matched, which prevents a drop in memory controllers and GPU compute units. The S-Series silicon is about 55 percent the size of the X-Series monster, and in our tests, the smaller controller consumed about 40 percent of the power.

You may also have noticed that neither AMD nor Nvidia is able to aggressively undercut the other in terms of price points, usually a sign of healthy competitive markets — and exactly what happened when AMD slashed the CPU space. Price parity is generally achieved because both are constrained by cost constraints imposed by chip manufacturers and memory suppliers.

Perhaps there is a bigger question to address regarding the Pro concept or the enhanced console: Is the mainstream gamer in fact tied to the concepts of more power and higher performance? The truth is, Microsoft’s gamble on the Series S has paid off. Reliable data suggests that, overall, Xbox mini consoles were sold in addition to the Xbox Series X. We’ve had some issues with some of Microsoft’s cuts to the unit — primarily memory and memory bandwidth, not to mention a lack of drive — but the console itself Irresistible.

Meanwhile, the Nintendo Switch is on its way to selling the already hugely successful PlayStation 4, despite launching more than three years later. Furthermore, the fact that the current generation of consoles is still rich in untapped potential – The Matrix Awakens from UE5 shows what’s possible, but we haven’t yet seen a shipping game emerge with the same level of ambition. The CPU, GPU, and even storage have hardly been explored.

The Xbox Series S is the cheapest and least capable console of the current generation – but its build is great, in keeping with its overall appeal. Here is the disassembly of the machine.

This is the case against enhanced consoles – but I would expect there to be updated versions of existing hardware. There have been reports that the PS5 will get a 6nm version of its processor, downsized from the current 7nm, and Microsoft will have the same option. Although cost reduction may be a challenge, it will be possible to use a smaller, more precise control unit and the expensive cooling solutions used by both manufacturers can be reduced. It may seem unlikely that the Series X and PS5 will hit $299 / £250, but there could be some price movement. So, is there any chance at all of an improved $499 console, even if there is a desire to make one?

AMD’s Zen 3 and Zen 4 CPUs look like a huge improvement over the Zen 2 consoles. Meanwhile, if AMD somehow manages to pull off an architecture as good as RDNA 2, it might be possible to boost the graphics – though remember a CPU The graphics multiplier between the Xbox One X and Xbox Series X is barely 2x, although the Scorpio machine is based mostly on the original 2011 South Islands architecture. The improved clock speeds are certainly workable — there’s talk of RDNA 3 hitting 3 GHz (The PS5 keeps the console clock register at a maximum of 2.23GHz) – but that’s another major challenge for power consumption and cooling.

We don’t know about improvements to ray tracing performance that might be interesting but a potential area of ​​interest is in machine learning. Why double your GPU when an AI upgrade similar to Intel’s XeSS and Nvidia’s DLSS can deliver excellent 4K image quality at a quarter of the internal resolution? The integration of silicon to machine learning opens the door to a smaller chip capable of achieving much higher levels of graphics performance — an alternative to spending a lot of money on a much larger GPU. Machine learning is the new frontier in technology and some would say the biggest potential improvement to the consoles we have today, and a huge enabler for a new generation of consoles in the future.

However, there is an alternative view – the idea that the console generation as we know it is dead, a scenario born through the longest intergenerational transition we’ve ever seen. There is a possibility that consoles optimized as a concept will incorporate what will traditionally become next generation – console generations give way to the ever-present cross-generation scenario, a more intrusive version of the mobile upgrade cycle. Recent years of console games have shown that they can work and with the cost of game development increasing, the idea of ​​titles staggered across generations may actually make this scenario inevitable.

What I can say is that console hardware development continues across all major platform owners – but the concept of what a new generation represents may well evolve with the limitations they face. How Sony and Microsoft will respond to the challenge is not yet clear but in the meantime, all eyes are on AMD. The need for backward compatibility means that Ryzen CPU cores and Radeon graphics will remain at the heart of any new console hardware — and AMD’s innovations in the PC space will at least give us an idea of ​​the building blocks that Sony and Microsoft will have to work with in the coming years.