I wouldn't necessarily say that, without knowing the costs associated with each console or the actual specs (but we'll go with the agreed upon narrative right now), we can't say whether Sony are being realistic or conservative. Likewise we can't say whether MS are being overambitious.
On the face of it, MS are betting big on attempting to win back some of the consumers they lost this generation. It's entirely possible that they're eating a relatively large loss on each console sold in the hopes of winning back more market share. If so, that could bite them in the butt if not enough consumers come back and/or they can't cost reduce as aggressively as they hope to be able to.
If I were to rate each maker on the specs that many seem to believe is real (I'm still not convinced we have the right numbers for both consoles) then I see...
- Sony having the more realistic approach to building their console, with an eye towards balancing the most powerful console they can make WRT how much money they expect to make during this next generation. Basically they are working within their established budget (expectations being that they'll retain all or most of their users from this generation).
- Microsoft are over-achieving with their console. They have large cash reserves and the board has OK'd dipping into it heavily in order to win back consumers lost during the last generation. The plan is to grow and establish a large software ecosystem (PC, Console, and streaming with Netflix like game subscriptions alongside software purchases) with consumer confidence that whatever they buy, they'll be able to run it forever on whatever hardware they get. This could be seen as a stepping stone towards that, so they may be willing to take a loss, break even, or just have a small profit for this generation.
There's also many other narratives that we could come up with. Lockhart if it is released changes the dynamics of having a very expensive console (Anaconda). PS5 could be more powerful than people are currently saying it is. Perhaps there's some breakthrough in technology (whether fab related or other tech related) coming along that we don't know about that will change the cost calculations of the consoles this generation.
Basically...WE DON'T KNOW.
And I see far too many people just buying into the narrative that
these are the specs for this generation of consoles (and then either patting themselves on the back, panicking or whatever) rather than these just being data points that might or might not be valid when the consoles release to retail.
Regards,
SB
Yes Nextgen Consoles are very medicore , to much standard Tech , and hold back the Game Development and here the answers why:
Only 9-12 Tflops for the GPU , this is not enough for a generational Jump and for a Lifetime of 5-7 Years. They need 20 Tflops at a minimum. In 2021 GPUs coming to the market for the PC with a much higher Tflop number. And those Monster Consoles become very quick tiny Machines.
Only 8 CPU Cores , not enough , in 2021 we see 16 and 12 Cores are coming to the Massmarket for an affortable Price , so Multiplatformgames on Nextgen Consoles are cleary downgraded Ports with less Features, because they are still limited by 8 CPU Cores.
Only 16 Gbyte Ram , this is only 2 X more than last Generation and cleary not enough for nextgen Content/Games. And you cannot replace a big Ram Memory with a SSD , because they are not fast enough , not enough Bandwidth, and they have a limited Lifetime. When SSD Chips running hot, the Datarates becomes slower or going down, so you didnt have permanent 2 Gbytes /s .
UMA Memoryarchitektur with a small 256 Bit Bus shared by CPU and GPU combined with high Latency Memory GDDR6. This ist the same Bottleneckarchitecture from previous Generation. Not enough for 4K, Raytraycing ,60 Fps and Rgba32 Puffer. And cleary a Design Failure. Because CPU and Raytraycing needs low latency Memory to run well , so why they put high Latency Memory into nextgen Consoles???
No additional Caches, Secret Sauces etc. to solve the Bandwidth and Latency Problem. And VRS , Checkerboarding or other Compression Technics destroys Graphicquality.
No Interposer with HBM Memory or other wide I/O solutions. HBM2/3 has higher Bandwidth, Datarates and wider Bus with lower Latency, Powerconsumption in a smaller Package than GDDR6.
No Tensorcores, no Deep Learning functions to make Gaming AI, NPC etc more intelligent, for a really new Gaming Experiences or clean Resolutionscaling Technics.
No FPGAs to Update the Hardware for new GPU/CPU Functions in the Future.
No In-Memory-Computing. IBMs new Chiptech can make Calculations 200 x faster han with traditional Cpu/Gpu Computing. It make sense to put some parts of the CPU and GPU direct into the Memory. I would put the Rops and Raytraycingunits direct into the Framebuffer to push up the Calculation/Renderspeed.
Focusing more and more on Foward and Back Compatibility , for difffernt Consoleiteration, Pro Models etc. , focusing on Play Anywhere Ecosystem, focusing more and more on High Level Programming , Times where Game Companys coding to Metal like in the old 32/16 Bit era are definitiv over. So i dont believe in Consoleoptimations anymore. This makes Consoles total uninteresting.
