Sony PS6, Microsoft neXt Series - 10th gen console speculation [2020]

[rntongo]

I'm not saying that 32GB of VRAM are impossible, but ignoring costs, it could end up like the 12GB's of the Xbox One X, which is, mostly unused. Nvidia's low-mid end GPU's feet's are firmly planted on low VRAM amounts, and that's a big chunk of the market. Then you got current gen consoles and Switch 2.

The forward looking bet is probably following in Nvidia's footsteps, and that is, AI hardware to upscale textures and/or reduce VRAM consumption.

My prediction is 32GB of unified RAM for the PS6, not 32GB VRAM since thats not been a thing on consoles for a while(Xbox 360 standardized unified memory so its not going to change). And no it wont be similar to the Xbox One X's 12GB bump because that was a midgen refresh. There wasnt a major architectural shift especially a developer ecosystem driven rethinking in how memory was to be utilized. Devs were still targeting the base 8GB slower memory when designing games. Top it all off it still had a spinning HDD which meant the memory hierarchy was bottlenecked by disk I/O. So even with a 384 bit bus, you'd only get about 70% of the 326GB/s bandwidth utilized due to mismatch between disk throughput and actual software design considerations related to the base hw.

Even more important, if you read my prior posts, the major thing devs clamored for(during the 8th gen) was for an introduction of SSDs like SATA, not more memory or memory bandwidth. Memory bandwidth wasnt the biggest request because the main bottleneck was the disk IO and then the CPU. Which is why Sony and MS went above and beyond to provide NVMe SSDs with highly customized hw compression blocks this gen. Again if you read my prior posts, Sony (and to some extent MS) 100% determines console specs by taking developer input and making long term bets on a 7-8 year console cycle, they dont base console specs on PC consumer averages.

This gen, whats different is devs have signaled a need for more physical memory, even MS's impressive virtual memory optimizations with Sampler Feedback Streaming wasn't enough especially with the Series S. With the PS6, Cerny already announced in the PS5 pro technical seminar that the focus of the PS6 will be not just a larger GPU( rasterization block) but also larger die area dedicated to hw acceleration for RayTracing and AI Upscaling.

Now add a more powerful CPU with at least 16 CPU cores to issue instructions/fully utilize the aforementioned 3 hw accelerators for graphics and you have a serious requirement for memory and memory bandwidth. You need to feed the CPU, and these 3 accelerators with enough memory bandwidth. As I discussed earlier with the example of the base Xbox One, cutting corners with ESRAM, couldnt beat Sony simply going wider+ faster memory. There is just no way around this. For a console supposed to last 7-8 years the last thing you want is to cut corners on memory.

Remember also the PS6 pro will likely have to use the same amount of unified memory. Pro models have always used the same amount of memory as the base console because of fundamental Software design constraints of the base model, its unlikely devs will spend the extra time to fully utilize extra physical memory on a pro model(but they can use the extra memory bandwidth on a pro model). So I just dont see anything less than 32GB of memory for the PS6 which will also be 2GB of memory for the PS6 pro. You cant handicap the PS6 pro like that.

 

[function]

Been having a few thoughts about a next gen lineup.

I know I'm not the only person to think about the same chip being used in both a handheld and a separate tv based console. But today I've been thinking more about the suitability of a 'handheld' chip in a conventional console, and how the console might differ.

At the simplest level you would use the console in the handhelds' docked performance level. But why stop there? With sufficient power and cooling a (for instance) 15W mobile chip could run at 60W or more, and roughly double performance over the 15W mode. More aggressive memory clocks could help boost bandwidth, or perhaps it might be possible to make a chip that could support both LPDDR5 and DDR/GDDR (I think AMD did something like this with an early APU, but the GDDR support was never used on it).

MS's strategy of Series S and Series X didn't work out how they wanted, but it wasn't fundamentally a bad idea they just executed poorly. Sharing a chip between a handheld and a Series S like console would be a good way to manage development costs, benefit from economy of scale, and potentially give two binning options for the chip.

Of course, MS might not make another console, and Sony don't seem keen on a Series S like cheaper home system, so maybe there'll be no-one to try this next gen.

 

[rntongo]

Been having a few thoughts about a next gen lineup.

I know I'm not the only person to think about the same chip being used in both a handheld and a separate tv based console. But today I've been thinking more about the suitability of a 'handheld' chip in a conventional console, and how the console might differ.

At the simplest level you would use the console in the handhelds' docked performance level. But why stop there? With sufficient power and cooling a (for instance) 15W mobile chip could run at 60W or more, and roughly double performance over the 15W mode. More aggressive memory clocks could help boost bandwidth, or perhaps it might be possible to make a chip that could support both LPDDR5 and DDR/GDDR (I think AMD did something like this with an early APU, but the GDDR support was never used on it).

MS's strategy of Series S and Series X didn't work out how they wanted, but it wasn't fundamentally a bad idea they just executed poorly. Sharing a chip between a handheld and a Series S like console would be a good way to manage development costs, benefit from economy of scale, and potentially give two binning options for the chip.

Of course, MS might not make another console, and Sony don't seem keen on a Series S like cheaper home system, so maybe there'll be no-one to try this next gen.

In a mobile first world, launching two separate home consoles at once is bound to be a nightmare. You're right, it was executed badly(although I think it was also fundamentally flawed). You bring up an interesting angle. I wonder how feasible it would be to have a shared APU considering the targets for a home console and targets for a portable device are quite different. I think the APU in a portable device would need smaller dies with fewer CUs. Also dont know how the memory configuration would work for such and APU, interoperability between LPDDR and GDDR would be a nightmare. Seems building a separate device with 3nm APU 8 core Zen CPU with 12-20 UDNA CUs,12-16GB of RAM would be more feasible. It can share the fundamental CPU and GPU architecture with the base PS6. And then build a PS6 with a 16 core Zen CPU 80+ UDNA or RDNA CUs and 32GB of GDDR7 RAM.

 

[Khalid]

In a mobile first world, launching two separate home consoles at once is bound to be a nightmare. You're right, it was executed badly(although I think it was also fundamentally flawed). You bring up an interesting angle. I wonder how feasible it would be to have a shared APU considering the targets for a home console and targets for a portable device are quite different. I think the APU in a portable device would need smaller dies with fewer CUs. Also dont know how the memory configuration would work for such and APU, interoperability between LPDDR and GDDR would be a nightmare. Seems building a separate device with 3nm APU 8 core Zen CPU with 12-20 UDNA CUs,12-16GB of RAM would be more feasible. It can share the fundamental CPU and GPU architecture with the base PS6. And then build a PS6 with a 16 core Zen CPU 80+ UDNA or RDNA CUs and 32GB of GDDR7 RAM.

What If there were 3 consoles for either Sony or MS. Obviously it’s a tall order to ask developers to develop 3 versions of their games to run on these consoles but you can solve this by having the high tier console pushing for most advanced APU while the handheld use a smaller APU. Then you take the very same handheld APU and package it in a smaller home console like a Series S. This way developers only need to develop Two versions of their games while console manufacturers can provide Three options for the consumer. Is that approach feasible?

 

[Charlietus]

Is there a list of current gen only titles that aren't on PS4 and Xbox One? Those would be the ones that would need patches. And right now, considering how long cross gen has gone on, it's probably under 100?

 

[Johnny Awesome]

Nothing wrong with the S. It's almost 50% of Series sales. It may yet shine even brighter in this tariff world we are living in.

 

[raytracingfan]

Doing the math, 32GB GDDR7 on a 256-bit bus (I doubt there will be a bus width increase with the PS6, the PS4 also had a 256-bit bus) requires a clamshell design with 2GB modules or for 4GB modules to be invented. We can't assume that 4GB module will be invented and available with adequate volume and pricing at the time the PS6 launches, and a clamshell design would increase costs. So it might have only 24GB GDDR7 (using 3GB modules and retaining the 256-bit bus). If Sony believes that 24GB is insufficient, they could add some additional DDR5 memory. The PS5 Pro already has 2GB of DDR5 for system tasks, the PS6 could have more. 24GB GDDR7 + 8GB DDR5 is a plausible way to hit 32GB memory total at a reduced cost. Only about 12.5GB of the base PS5's memory is available for games, if the entire 24GB is available on the PS6 due to the system only using the DDR5 that would be double the available memory, enough for a true generational leap.

 

[Mitchings]

Posted elsewhere, but these are my latest crazy ramblings..

More what Id like to see than what I expect, but just trying to think in a holistic way within each range and over the course of the gen:

~2028


~2032


I know I know, first thing people see is a 'lite' console, but consider the much lesser gulf here in areas that matter vs XSS/XSX and how it'd scale. It'd also act as a lower cost replacement to the Pro and may even be low enough to entice more base PS5 users or potential PS5 buyers over sooner. PS5 is probably gonna hamstring the PS6 for a good 3-4yrs on most titles anyway, perhaps a console that's less powerful but architecturally on par may be a preferable means of hamstringing; in that it might lower the barrier of entry to get enough people over, in turn incentivising next-gen only development, even if the effect is small, to the tune of a few big games and maybe a ~year. But a ~year is a big chunk of a 7-8yr gen.

Spoiler: More Ramblings:

"In my eg. above, it'd be modular/scalable with chiplets, where each console is fundamentally identical in featuresets and the main board etc., they'd share the same CPU with exception to some extra V-cache added on the main system, then some more for the Pro. The Lite would use 1 Graphics Core Die while the main system would use 2 of the same GCD. Only to make up a little of the gulf, the higher yield units will be fully utilised in the lite and be fully activated to offer 57.5% of the power rather than just 50% (plus with lower thermals, they could just peg the continuous boost scheme higher, or even go the opposite way and use it to provision with a more conservative thermal design), meanwhile the 2 GCDs in the main system will have a 10% redundancy, using lower yield versions of the same chips. Then the MCD's containing the memory bus and infinity cache will scale in quantity with the chips. On the latter Pro, it'd be the same approach albeit with upgraded GPU architecture, add one GCD, increasing the power by ~50%, add an MCD to increase L3 cache and the memory bus by 50%. Keep the CPU CCD same from the beginning to the end on the same process to save money, keep the MCDs on ~N6 for the first console lineup, only upgrading if/when it's viable to do so. By saving in all of these areas, you can focus on using a cutting edge process on the GPU GCDs where it's most valuable, getting backside power delivery + full gaafet will allow it to be pushed harder and/or use less power, which has knock on performance gains and/or cost savings. Not to mention if both are fully designed for, there may be an advantage in terms of topology, area and other optimisations.

Given that it's chiplets, if they have an excess of high yield PS6 Lite level chips, low yield PS6 base level chips or any even lower yield chips that can't be used, they can sell them back and AMD can repurpose them for any number of non standard products, same goes for the CPU dies and the MCD dies. When it's all modular, little has to go to waste.

Unlike the XSX/XSS gulf, the divide here for the base system and lite would be 58% GPU power, 72% Usable RAM, 70% Memory Bandwidth, everything else the same. You'd save money on 1GCD vs 2, you'd save money on no 3D V-Cache, you'd save money on half the storage, you'd save money on cooling and PSU, you'll get a smaller, slightly lighter chassis and you'd save on weight/size for shipping. You could use an identical board, just with less/more VRMs, RAM chips and different density NVMe chips. But all the core features and architectural advantages are there, and the trade-off could -- as an eg. -- be something along the lines of a PS6 base game running 1296-1536p with full fledged upscaling and lots of rays cast, vs say 1080-1152p with slightly less comprehensive scaling and less rays cast. It lowers the barrier of entry significantly but the downgrade isn't excessive.

Also, XSS/XSX are two distinct APUs with ~half the total volume of PS5's APUs (and obviously much less by themselves). Here they'd have the economies of scale on their side not just by being the likely market leader again, but the dies will be shared between everything, with exception to the eventual Pro using an updated GCD die."

 

[Charlietus]

Doing the math, 32GB GDDR7 on a 256-bit bus (I doubt there will be a bus width increase with the PS6, the PS4 also had a 256-bit bus) requires a clamshell design with 2GB modules or for 4GB modules to be invented. We can't assume that 4GB module will be invented and available with adequate volume and pricing at the time the PS6 launches, and a clamshell design would increase costs. So it might have only 24GB GDDR7 (using 3GB modules and retaining the 256-bit bus). If Sony believes that 24GB is insufficient, they could add some additional DDR5 memory. The PS5 Pro already has 2GB of DDR5 for system tasks, the PS6 could have more. 24GB GDDR7 + 8GB DDR5 is a plausible way to hit 32GB memory total at a reduced cost. Only about 12.5GB of the base PS5's memory is available for games, if the entire 24GB is available on the PS6 due to the system only using the DDR5 that would be double the available memory, enough for a true generational leap.

I thought that 4gb modules were already available. That makes something that I struggled to believe pretty much an impossibility. The only realistic possibility that remain are 18gb with a 192 bit bus or 24gb with a 256 bit bus, all with 3gb modules. Additional DDR5 ram to free up GDDR7.

 

[Charlietus]

Posted elsewhere, but these are my latest crazy ramblings..

More what Id like to see than what I expect, but just trying to think in a holistic way within each range and over the course of the gen:

~2028
View attachment 13461

~2032
View attachment 13462

I know I know, first thing people see is a 'lite' console, but consider the much lesser gulf here in areas that matter vs XSS/XSX and how it'd scale. It'd also act as a lower cost replacement to the Pro and may even be low enough to entice more base PS5 users or potential PS5 buyers over sooner. PS5 is probably gonna hamstring the PS6 for a good 3-4yrs on most titles anyway, perhaps a console that's less powerful but architecturally on par may be a preferable means of hamstringing; in that it might lower the barrier of entry to get enough people over, in turn incentivising next-gen only development, even if the effect is small, to the tune of a few big games and maybe a ~year. But a ~year is a big chunk of a 7-8yr gen.

Spoiler: More Ramblings:

"In my eg. above, it'd be modular/scalable with chiplets, where each console is fundamentally identical in featuresets and the main board etc., they'd share the same CPU with exception to some extra V-cache added on the main system, then some more for the Pro. The Lite would use 1 Graphics Core Die while the main system would use 2 of the same GCD. Only to make up a little of the gulf, the higher yield units will be fully utilised in the lite and be fully activated to offer 57.5% of the power rather than just 50% (plus with lower thermals, they could just peg the continuous boost scheme higher, or even go the opposite way and use it to provision with a more conservative thermal design), meanwhile the 2 GCDs in the main system will have a 10% redundancy, using lower yield versions of the same chips. Then the MCD's containing the memory bus and infinity cache will scale in quantity with the chips. On the latter Pro, it'd be the same approach albeit with upgraded GPU architecture, add one GCD, increasing the power by ~50%, add an MCD to increase L3 cache and the memory bus by 50%. Keep the CPU CCD same from the beginning to the end on the same process to save money, keep the MCDs on ~N6 for the first console lineup, only upgrading if/when it's viable to do so. By saving in all of these areas, you can focus on using a cutting edge process on the GPU GCDs where it's most valuable, getting backside power delivery + full gaafet will allow it to be pushed harder and/or use less power, which has knock on performance gains and/or cost savings. Not to mention if both are fully designed for, there may be an advantage in terms of topology, area and other optimisations.

Given that it's chiplets, if they have an excess of high yield PS6 Lite level chips, low yield PS6 base level chips or any even lower yield chips that can't be used, they can sell them back and AMD can repurpose them for any number of non standard products, same goes for the CPU dies and the MCD dies. When it's all modular, little has to go to waste.

Unlike the XSX/XSS gulf, the divide here for the base system and lite would be 58% GPU power, 72% Usable RAM, 70% Memory Bandwidth, everything else the same. You'd save money on 1GCD vs 2, you'd save money on no 3D V-Cache, you'd save money on half the storage, you'd save money on cooling and PSU, you'll get a smaller, slightly lighter chassis and you'd save on weight/size for shipping. You could use an identical board, just with less/more VRMs, RAM chips and different density NVMe chips. But all the core features and architectural advantages are there, and the trade-off could -- as an eg. -- be something along the lines of a PS6 base game running 1296-1536p with full fledged upscaling and lots of rays cast, vs say 1080-1152p with slightly less comprehensive scaling and less rays cast. It lowers the barrier of entry significantly but the downgrade isn't excessive.

Also, XSS/XSX are two distinct APUs with ~half the total volume of PS5's APUs (and obviously much less by themselves). Here they'd have the economies of scale on their side not just by being the likely market leader again, but the dies will be shared between everything, with exception to the eventual Pro using an updated GCD die."

Even if I don't agree with the specs, that's very detailed and rambly . The first thing that i would cut is 8 cores from the CPU. 99% of games barely use 7 cores effectively, so 16 would be a waste, probably. Then, the SSD can pretty much remain the same as PS5, as no games uses it still to it's potential. Aside from that, two chips have leaked, one for a home console, and one for a portable. So, I don't think that there is going to be a "lite" necessarily.

I could see the lite for 549$ as the only PS6, I don't see how they could sell that at 449$ without bleeding themselves.

 

[McHuj]

I thought that 4gb modules were already available. That makes something that I struggled to believe pretty much an impossibility. The only realistic possibility that remain are 18gb with a 192 bit bus or 24gb with a 256 bit bus, all with 3gb modules. Additional DDR5 ram to free up GDDR7.

Actually the split memory pool is not a bad idea. Using lpddr5 you could plop down a single die with 16 GB of memory. Yes, they’d have to go back to a split memory pool, but I don’t think it’s a big deal.

 

[Charlietus]

Actually the split memory pool is not a bad idea. Using lpddr5 you could plop down a single die with 16 GB of memory. Yes, they’d have to go back to a split memory pool, but I don’t think it’s a big deal.

PS5 has 512mb of DDR4, and the pro has 2gb of DDR5. So they would have to just put in a bigger chip, like 4GB.

 

[chris1515]

Even if I don't agree with the specs, that's very detailed and rambly . The first thing that i would cut is 8 cores from the CPU. 99% of games barely use 7 cores effectively, so 16 would be a waste, probably. Then, the SSD can pretty much remain the same as PS5, as no games uses it still to it's potential. Aside from that, two chips have leaked, one for a home console, and one for a portable. So, I don't think that there is going to be a "lite" necessarily.

I could see the lite for 549$ as the only PS6, I don't see how they could sell that at 449$ without bleeding themselves.

Some developer wanted 16 cores for PS5. I suppose Naughty Dog, ID software or Guerrilla Games would be happy to have 16 cores...

How PS5 Was Built: Mark Cerny Reveals That Developers Wanted a 16-Core CPU in New Video

Wired has shared a new video with Mark Cerny that provides new insight on the PlayStation 5 and some of the interesting decisions that Sony had to make during the course of its development.

www.thefpsreview.com

How PS5 Was Built: Mark Cerny Reveals That Developers Wanted a 16-Core CPU in New Video

hey don't do the CPU for the average devs but for those who want to push the system.

The Jolt physics engine used by Guerrilla Games scale well until 16 cores

https://jrouwe.nl/jolt/JoltPhysicsMulticoreScaling.pdf

 

[QPlayer]

I could see the lite for 549$ as the only PS6, I don't see how they could sell that at 449$ without bleeding themselves.

The next consoles will cost at least $800. ! I was right about the $700/€800 PSPro. It's another matter that they sell for $630 these days, because that way they can sell some, not many...

But the price of the next consoles won't go down... at today's component prices!

 

[Charlietus]

Some developer wanted 16 cores for PS5. I suppose Naughty Dog, ID software or Guerrilla Games would be happy to have 16 cores...

How PS5 Was Built: Mark Cerny Reveals That Developers Wanted a 16-Core CPU in New Video

Wired has shared a new video with Mark Cerny that provides new insight on the PlayStation 5 and some of the interesting decisions that Sony had to make during the course of its development.

www.thefpsreview.com

hey don't do the CPU for the average devs but for those who want to push the system.

The Jolt physics engine used by Guerrilla Games scale well until 16 cores

https://jrouwe.nl/jolt/JoltPhysicsMulticoreScaling.pdf

I believe that consoles shouldn't be made with the 0.1% of developers in mind. If you start to take in to consideration all the wants of the developers you end up with stuff like the PS5 SSD. Made the 0.1% of developers that will maybe use that happy, but they could have saved some 10-20$ on the BOM and nothing would have changed.

 

[Charlietus]

The next consoles will cost at least $800. ! I was right about the $700/€800 PSPro. It's another matter that they sell for $630 these days, because that way they can sell some, not many...

But the price of the next consoles won't go down... at today's component prices!

Before the tariffs, Sony was probably making around 200$ per Pro unit as pure profit. The Pro isn't indicative of future prices, and they aren't that dumb. They know that making a 800$ PS6 as the only option would kill their business model.

 

[Lurkmass]

The Pro isn't indicative of future prices, and they aren't that dumb. They know that making a 800$ PS6 as the only option would kill their business model.

It's not like they have any option with 'standardized' hardware which is to evolve to more powerful systems and it's in their business model to provide subsidized access to public by design ...

If developers want X levels of performance then that's Sony's job to justify their customers a higher price to make it happen ...

 

[Charlietus]

It's not like they have any option with 'standardized' hardware which is to evolve more powerful systems and it's in their business model to provide subsidized access to public by design ...

If developers want X levels of performance then that's Sony's job to make it happen ...

They will hit a price point that makes sense for them, barring any craziness with tariffs or something else.

Let's not forget that Cerny doesn't see big raster power increases in the future. And, as much as some people fear it, it might even mean that in raster games PS6 isn't that much more powerful than a PS5 Pro.

PS5 Pro is a smaller chip than the original PS5, I feel like many people don't see how cheap Sony has gone this time around.

Like, developers can ask for 16 cores, 32gb of ram, 40 teraflops and ReRam and Sony would shrug and tell them that it's not in the budget. Otherwise, they will prepare for a PS6 that sells half of the PS5.

 

[Lurkmass]

They will hit a price point that makes sense for them, barring any craziness with tariffs or something else.

Let's not forget that Cerny doesn't see big raster power increases in the future. And, as much as some people fear it, it might even mean that in raster games PS6 isn't that much more powerful than a PS5 Pro.

PS5 Pro is a smaller chip than the original PS5, I feel like many people don't see how cheap Sony has gone this time around.

AAA game developers and Epic Games probably wants at MINIMUM ~4080 tier performance level so Sony can sell that for $800USD or take bigger losses ...

It's either the above happens or somehow Sony tries to sell alternative rendering techniques (beyond just either raster, neural, or RT) to them ...

 

[Charlietus]

AAA game developers and Epic Games probably wants at MINIMUM ~4080 tier performance level so Sony can sell that for $800USD or take bigger losses ...

It's either the above happens or somehow Sony tries to sell alternative rendering techniques (beyond just either raster, neural, or RT) to them ...

That second alternative is what will happen, I don't have many doubts about that. What developers want and what they will get is probably very different. But even in the absurd scenario where PS6 has less raster power than a PS5, developers would still develop for it. It's not like they are going to cross their arms and refuse to work.