After my recent rant decrying the Wii as a failure, I’m sure I’ve pissed off many a Nintendo fan. Let me assure you that’s not my intention. I still love Nintendo as a fantastic game company that simply needs to step its hardware game up. Nintendo needs to compete in the modern electronics market and that requires them to be more accommodating to consumer expectations of the day. So with rumors erupting that Nintendo will soon launch a new console before or at E3, there’s no way I won’t join in on the speculation. Yes, this will be a specs post, but we’ll also weigh the culture of Nintendo’s next console as well.
So, what exactly is going to run this thing? Nintendo has established that it is perfectly fine with selling a low power, high margin machine. Regardless, most rumors state that the new console will support HD graphics and is either “as powerful as”, “slightly more powerfully than” or “significantly more powerful than” the 360 and PS3. Picking the first option would make economic sense for Nintendo, since the company could use commodity IBM parts and easily achieve parity with Microsoft and Sony. A “slightly more powerful” system could allow for increased efficiency or higher memory capacity in order to overcome some deficiencies in current systems. And of course there is the final option which could see Nintendo try to hit parity with today’s highest end machines (don’t hold your breath). Most chances are that we’ll see something in between “slight” and “significant” performance increases if Nintendo plays their cards right. No console in 2011 could be sold at an affordable price and still have equal specs to today’s strongest gaming rigs.
What would this hardware be specifically? It’s hard to measure anything exact. What we can be almost certain is that Nintendo will continue to use an IBM solution. IBM swept the console CPU market last generation but they delivered a pretty excellent product with the PPE core used in the 360′s Xenon 3-core design and the PS3′s CELL primary core. IBM could easily adapt the PPE for the Wii’s purposes, delivery essentially a quad-core version of the Xenon or modifying the design with out-of-order processing or faster clock speed. Other options include other a custom chip based on the POWER6 or POWER7 or other POWER4 derivatives like the PowerPC 970 (aka G5). That processor hasn’t seen action in a long time but refabrication at the 45nm node could make for an interesting multicore chip.
Some have rumored that Nintendo will use another PowerPC meant for the embedded market called the 476FP. IBM built the chip for high performance server work and network appliances, so it has pretty strong performance in tight power requirements. The 476FP also supports 16-core SMP design so it would not be outside the realm of possibility if Nintendo used a 4 or 8-core SoC at around 2Ghz. The caveat to this is that the 476FP architecturally doesn’t look vastly different from the ARM Cortex A9 in today’s dual core tablets. Besides from clear backwards compatibility with the PowerPC ISA and energy efficiency, the 476FP isn’t drastically different from the Wii’s current PPC750-based Broadway and may suffer from per-core performance. Only aggressive core counts would make it a major contender.
On the graphics side, Nintendo will also continue to use an AMD (née ATI) solution. Nintendo has a good relationship with the company and ATI has struck an excellent balance between performance and efficiency in their GU cores. Nvidia may hold the crown for pure performance with Fermi, but I doubt any next generation console could use that architecture. Rumors place the Wii 2′s GPU at around the R700 generation, in the HD4000 era. That’s not bad and if the console had the horsepower of an HD4870 it would be a major jump from current generation consoles while keeping the wattage down. Why not Evergreen or Northern Islands? Simply because these architectures were probably not done by the time Nintendo commenced R&D on the new console. RAM should be in the area of 4GB, but I highly doubt it will. I would expect the system to have about 1GB of main system RAM and probably give the GPU 512MB to 1GB of its own RAM. At least the company will be moving to a modern Unified Shader Architecture.
So this post is spec heavy, it’s meant for that group of my audience that wants some black and white specs to chew on and assess. For everyone else, here’s what else the Wii 2 needs. A much better online system. No friend codes, no system codes. Give me gamer tag and a Mii to associate. Let the Mii plaza be a place where we can communicate freely with friends and other players, trade Miis and game data, maybe even have voice chat. It’s time for modern social networking. Keep the Virtual Console and WiiWare, that’s been a great addition but keep improving it. In general the Channel interface was actually really smart and flexible, but let developers submit applications. The iPhone has done very well here, if Nintendo had a flat-out app store, they could bring in new developers. And for the love of god, introduce real storage, DLC and patches. You could fix many a broken game with patches.
As for control, the Wii remote is pretty good. Don’t do this HD screen BS. Don’t make the controller some specialized secondary game machine. Improve the sensors and IR tracker for better accuracy, contour the shape for comfort, improve the speaker, add actual memory storage so you can bring along Miis and game saves in the controller. You made the right choice with motion control and developers are getting better. Just don’t innovate for the sake of innovation.
Nintendo went out on a major limb last generation with the Wii. It paid off financially and publicly for a long time, though it was ultimately not sustainable. It’s time to reel their console strategy back into the mainstream and bring the fight to Microsoft and Sony with their expanded audience. You’ve proven the casual audience exists but you can support them while making your next console something developers really want to sink their teeth into. It’s time to get back into the fight Nintendo!