With the full hardware specification for the PlayStation Neo out in the open, several questions spring to mind, principally - why now: as the impending arrival of the Nintendo NX caused a rethink at Sony headquarters? And perhaps more curiously, how is Microsoft going to respond?
The way things look right now, it seems that the biggest winner of the upcoming console three-way is AMD. After all, the chances are that the technology of each new piece of hardware features the same core CPU and GPU components produced by the same engineers. Which begs the question - at the technological nuts and bolts level, are we going to end up with three almost identical new consoles? And is that such a bad thing?
Nintendo, Microsoft and indeed Sony itself will be looking at the phenomenon that is PlayStation 4, wondering why it was so successful. Part of the formula is undoubtedly the strength of the core technology. Looking back, Nielsen polling data suggested that "better resolution" was the number one reason buyers purchased PlayStation 4 over its competition. On the one hand, that nugget of data may seem quite unbelievable but on the other, consider this - given two extremely similar pieces of hardware sold at similar price-points with much the same library of software, which would you buy? "Better resolution" may well be the closest the survey had to a response highlighting the spec differential.
And with the upcoming current-gen redux, you can be sure that Microsoft - and perhaps even Nintendo - will not want to be hamstrung by lower-spec hardware when the possibility exists for a competitor to produce more powerful kit at the same price. That said, by and large, developers have done a good job adapting to the differences between PlayStation 4 and Xbox One. 900p resolution works effectively in most cases as a 1080p surrogate, while mostly subtle reductions in visual effects quality mostly go by unnoticed. And of course, the 30fps cap in the majority of console titles is a great leveller. As things stand, we only really have one game that allows itself to run on both systems identically, with an unlocked frame-rate - Io Interactive's Hitman, quite possibly the closest thing we actually have to a current-gen hardward benchmark.
The video speaks for itself, but the headlines are clear enough - there's a 30 per cent boost to PlayStation 4 performance in GPU-bound areas, and a 10 per cent improvement on Xbox One in NPC-heavy CPU-limited test scenarios. It's a fascinating, unique look at the spec differential between the two consoles. It's just one test of course (a more rounded look at two consoles at a hardware level would encompass many more titles) but it serves to highlight the relative strengths and weaknesses of the two machines. It's the only PC-like hardware benchmark we have, and it shows very PC-like results - one console has a faster CPU, one has a more powerful GPU. By and large, the machine with the more capable graphics is preferable judged purely on PC terms. And again, given the choice between the two, at similar price-points, with similar libraries, which would you buy?
Which brings us back to the PlayStation Neo hardware spec. Assuming a late 2016/Q1 2017 release, what we are looking at is the most powerful hardware design AMD is currently capable of producing at a console-friendly price-point. GPU-wise, things aren't 100 per cent clear, but assuming a 14nm FinFET design, the 36 compute unit tally looks very much like AMD's upcoming Polaris 10 GPU. Until the later Vega architecture comes online, the firm has nothing more powerful in its arsenal - Polaris 11 is now confirmed as a smaller GPU designed primarily for the laptop market, where AMD is promising 'better than console' performance. It's going to be interesting from a gaming notebook perspective, but not 'next-gen' enough for an Xbox One or PS4 successor.
The situation actually looks fairly moribund on the CPU side. With its upcoming Zen CPU chips seemingly aimed at the server market first with no sign of any low-power variant, it seems that the only viable x86 architecture available for a console is an upclocked version of the relatively weak Jaguar CPU we already have - exactly what Neo has. This makes things tricky for Microsoft. The move to x86 architecture in the current generation effectively rules out access to any other more powerful parts from other manufacturers. Other than AMD, only Intel can produce x86 processors - and moving to ARM for an Nvidia SoC would introduce more problems than it would solve. The only advance here would be to move to a 12-core Jaguar or Puma solution - possible, but may cause issues for the existing interconnect fabric.
In effect, unless Microsoft sinks a vast amount of money into a fully custom AMD design, or reverts to CPU and GPU cores from different manufacturers (highly unlikely), the best it can do is to produce a machine with a similar hardware spec to the Neo, or else sit back and wait a couple of years for more advanced hardware to come along. This does not seem likely.
Driving the move to new hardware is the availability of the 14nm FinFET production process, allowing for a 2x increase in transistor density compared to the 28nm chips found in PS4 and Xbox One. Processor clock-speeds can be improved too - hence the 2.1GHz Jaguar cores in Neo vs PS4's 1.6GHz, plus the 911MHz GPU speed, a curiously arbitrary 111MHz increase over the older hardware. But the key point here is that new AMD architectures are rolling out in step with new fabrication technologies, effectively meaning that innovations in hardware will roll out alongside the means to cost-effectively shrink transistors.
If the mid-gen console refresh is commercially successful, we should expect new hardware with each major node refresh - 14nm will be followed by 10nm, but it seems that 7nm will be the next major step for high-end graphics hardware. Factoring in the inevitable delays, plus the wait for these processes to become commercially viable, it suggests that new console hardware could arrive at three to four year intervals - and as long as AMD is in pole position with all platform holders, the cycle of very similar hardware manifesting in all plaforms could persist for a long time to come.
Of course, the question is to what extent a commonality in hardware is actually a bad thing. After all, a level playing field in spec means an easier time for developers, while the platform holders can concentrate elsewhere on services and features that make their consoles distinct and unique. And that's why it's the NX that we're particularly eager to see. Assuming Nintendo has chosen similar parts to Sony (not a foregone conclusion - remember that the smaller, less capable Polaris 11 should still handily outperform PS4), criticisms of under-powered tech will be a thing of the past. But perhaps more importantly, Nintendo has a proven track record of handing in distinctive hardware designs where the concept trumps the importance of the raw horsepower available.
There's a good chance that new Xbox hardware of some description will manifest at E3. The form it will take - a new console with a hardware boost, or a 'slim' revamp of the existing Xbox One - remains to be seen. But if it is the former, it presents an interesting challenge for Microsoft: can it beat the Neo spec, and assuming it does end up looking very similar, what else can the platform holder offer to challenge PlayStation's supremacy?
Will you support the Digital Foundry team?
Digital Foundry specialises in technical analysis of gaming hardware and software, using state-of-the-art capture systems and bespoke software to show you how well games and hardware run, visualising precisely what they're capable of. In order to show you what 4K gaming actually looks like we needed to build our own platform to supply high quality 4K video for offline viewing. So we did.
Our videos are multi-gigabyte files and we've chosen a high quality provider to ensure fast downloads. However, that bandwidth isn't free and so we charge a small monthly subscription fee of €5. We think it's a small price to pay for unlimited access to top-tier quality encodes of our content. Thank you.Support Digital Foundry