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

[Shifty Geezer]

Reminder how this thread rolls - This thread is ideally about technical forecasts, looking at technical trends and predicting what's possible in silicon. Ideally people will pick hardware targets like RAM, bus width, die sizes, fabrication tech, and string those together without discussing the business behind their choices.

Basically, don't argue performance targets but general plans for new hardware - there are dedicated business threads for those. Present your idea of the hardware that'll appear using your prediction of business models.

 

[Seanspeed]

You should just rename the thread to 'Specs that can theoretically exist within the next year to three years' in that case.

 

[Shifty Geezer]

It's broader than that. It's what displays will they be targeting, and interfaces, and novel tech. What it's not is another Business Discussion.

 

[rntongo]

These are silly speculations because most people are listing what they want instead of what is the capability of the company -- in this case -- AMD and the time-frame they have to work with.

Given cost, time and technical prowess of AMD, I wouldn't expect anything more performant than a RTX3090 for the next crop of consoles.

I beg to differ. By 2028?? Nothing better than an RTX 3090 that only has 24GB of GDDR6 memory? I'm thinking what Sony releases in 2028 will be certainly more powerful than a 6 year old GPU(Yes more powerful than an RTX 4090). I dont doubt NVIDIA GPUs are going to hold up well but a console releasing with a GPU weaker than a 6 year old high end consumer GPU(RTX 4090) doesnt make sense. They need to sell it for 8 years, cant be selling hw weaker than a 10 year old GPU in 2032 or even worse if its weaker than a 3090; a 12 year old GPU. PS5 launched with the equivalent of a 5700XT but with more memory and if it wasnt for the large initial cost of moving to an SSD, they could have invested more in the GPU. But that wasnt too far from a 2080 which had been released 2 years prior. The biggest cost next gen will go towards memory as well as accelerators for AI and Raytracing. No obscene costs from having to upgrade disk storage. Thats going to play a major factor in ensuring the PS6 has some serious hardware from the get go that Sony can quickly sell at a profit within the first 11 months. Some variant of GDDR7 is almost a guarantee as well which will also ensure to keep the costs down.

So above 32GB of GDDR7 RAM, GPU with Raster performance slightly above RTX 4090, Ray Tracing and AI/ML hw accelerators, 2TB memory bandwidth or more(min 4x memory bandwith), 16core CPU clocked at 5GHz.

 

[Charlietus]

I beg to differ. By 2028?? Nothing better than an RTX 3090 that only has 24GB of GDDR6 memory? I'm thinking what Sony releases in 2028 will be certainly more powerful than a 6 year old GPU(Yes more powerful than an RTX 4090). I dont doubt NVIDIA GPUs are going to hold up well but a console releasing with a GPU weaker than a 6 year old high end consumer GPU(RTX 4090) doesnt make sense. They need to sell it for 8 years, cant be selling hw weaker than a 10 year old GPU in 2032 or even worse if its weaker than a 3090; a 12 year old GPU. PS5 launched with the equivalent of a 5700XT but with more memory and if it wasnt for the large initial cost of moving to an SSD, they could have invested more in the GPU. But that wasnt too far from a 2080 which had been released 2 years prior. The biggest cost next gen will go towards memory as well as accelerators for AI and Raytracing. No obscene costs from having to upgrade disk storage. Thats going to play a major factor in ensuring the PS6 has some serious hardware from the get go that Sony can quickly sell at a profit within the first 11 months. Some variant of GDDR7 is almost a guarantee as well which will also ensure to keep the costs down.

I'd say that 4090 performance is the ceiling. You think that xx60 level hardware in 2028 will be more powerful than a 4090? I can't see that happening.

Also, silicon scaling and costs not reducing and yada yada...

 

[rntongo]

I'd say that 4090 performance is the ceiling. You think that xx60 level hardware in 2028 will be more powerful than a 4090? I can't see that happening.

Also, silicon scaling and costs not reducing and yada yada...

4090 will be too old by November 2028. RTX 60 series cards will be a larger leap from the 40 series than the current RTX 50 series cards. In any case I dont see Sony investing much in raster performance so 4090 should be the benchmark i.e a 6 year old GPU by 2028.

 

[Charlietus]

4090 will be too old by November 2028. RTX 60 series cards will be a larger leap from the 40 series than the current RTX 50 series cards. In any case I dont see Sony investing much in raster performance so 4090 should be the benchmark i.e a 6 year old GPU by 2028.

The 4090 will be old, but that doesn't matter honestly. Consoles trying to compete with PC on raw power is a mistake.

And if you look at the features and the capabilities, even a 4080 is a generational improvement compared to a PS5. A console that has a mature upscaling solution, with dedicated and performant ray tracing hardware and almost a terabyte of memory bandwidth? I wouldn't be disappointed at all.

The target is path tracing at good image quality and framerates, not going against 2028 Nvidia hardware.

 

[rntongo]

The 4090 will be old, but that doesn't matter honestly. Consoles trying to compete with PC on raw power is a mistake.

And if you look at the features and the capabilities, even a 4080 is a generational improvement compared to a PS5. A console that has a mature upscaling solution, with dedicated and performant ray tracing hardware and almost a terabyte of memory bandwidth? I wouldn't be disappointed at all.

The target is path tracing at good image quality and framerates, not going against 2028 Nvidia hardware.

Okay just as some benchmark for raster performance expect slightly better than 7900 XTX raster performance in the PS6. The 4090 and 7900 XTX are not that different in this regard.
Its too early to tell how well Sony/AMD partnership with project Amethyst will turn out.

 

[Karamazov]

next gen should heavily rely on AI upscaling and FG.
If they can offer gaphics not much more detailed than this gen, but with path tracing at decent frame rate, the percieved difference will be very noticeable.

 

[rntongo]

next gen should heavily rely on AI upscaling and FG.
If they can offer gaphics not much more detailed than this gen, but with path tracing at decent frame rate, the percieved difference will be very noticeable.

All this requires significant investment in the memory subsystem... which is going to happen. 2+TB memory bandwith

 

[Charlietus]

All this requires significant investment in the memory subsystem... which is going to happen. 2+TB memory bandwith

The 5090 has 1,8tb of memory bandwidth on a 512bit bus. There is no way that any console gets 2tb. Memory controllers aren't even scaling much anymore with node reductions, so it's even a possibility that consoles go with a 192bit memory interface, down from 256bit.

192bit with 36gb/s gddr7 would be close to a tb of bandwidth, do they even need more than that?

Even those specs feel more realistic for a day one pro console, with a mid priced console coming with lower specs.

 

[Seanspeed]

At the rate things are going, even 4090-like performance is potentially optimistic for a $600 or less base console in 2028.

Performance per dollar improvement dying out is going to kill any chance of the next consoles being impressive. Yea, they're gonna rely heavily on AI, because there's no other choice cost-wise, not because more raw power wouldn't still be supremely useful for elevating all round presentation aspects.

Next gen is gonna be the most disappointing generation ever by a long shot and I dont think it's a hot take to say that. I am not a cynic by nature. Gonna feel like a glorified 'Pro' generation. I think realistically, probably something like 2-2.5x improvement in CPU capabilities, 3-4x in GPU, 2x bandwidth, 2xRAM at most, meaningless SSD improvement. Will make the XB1/PS4->XSX/PS5 leap look huge in comparison. And being AMD, the AI capabilities wont even be great.

I think Sony and MS will be needing to rely heavily on people's short memories and a whole lotta extremely hyperbolic marketing nonsense to convince people these machines are a real generational uplift.

And god save Xbox if they try and release a 'new generation' console next year as some have rumored. lol Might as well start digging a deep pit now to bury them all ET-style after nobody can figure out why they should buy one.

 

[Shifty Geezer]

Next gen is gonna be the most disappointing generation ever by a long shot and I dont think it's a hot take to say that. should buy one.

Predict: Next gen console tech (10th generation edition) [2028+]

But that’s just the cost of the wafers. 7nm vs 5nm has roughly a 45-50% density improvement so the same design on 5nm vs 7nm will require far fewer wafers (assuming similar yields) in addition to benefits in power consumption allowing cheaper cooling. It’s a pretty complex design and...

forum.beyond3d.com

Why have we got two threads on this?

 

[Charlietus]

At the rate things are going, even 4090-like performance is potentially optimistic for a $600 or less base console in 2028.

Performance per dollar improvement dying out is going to kill any chance of the next consoles being impressive. Yea, they're gonna rely heavily on AI, because there's no other choice cost-wise, not because more raw power wouldn't still be supremely useful for elevating all round presentation aspects.

Next gen is gonna be the most disappointing generation ever by a long shot and I dont think it's a hot take to say that. I am not a cynic by nature. Gonna feel like a glorified 'Pro' generation. I think realistically, probably something like 2-2.5x improvement in CPU capabilities, 3-4x in GPU, 2x bandwidth, 2xRAM at most, meaningless SSD improvement. Will make the XB1/PS4->XSX/PS5 leap look huge in comparison. And being AMD, the AI capabilities wont even be great.

I think Sony and MS will be needing to rely heavily on people's short memories and a whole lotta extremely hyperbolic marketing nonsense to convince people these machines are a real generational uplift.

And god save Xbox if they try and release a 'new generation' console next year as some have rumored. lol Might as well start digging a deep pit now to bury them all ET-style after nobody can figure out why they should buy one.

I don't agree with that at all. If they really get to Nvidia levels of ray tracing and ai (maybe rtx 4000 levels?), next gen will be the last where common people will see meaningful differences. Much more than in this generation, where settings upgrades and higher resolution are what we have mostly seen so far.


And I remember seeing screenshots with differences that were even more evident. This is much easier to market to the masses.

 

[Seanspeed]

Much more than in this generation, where settings upgrades and higher resolution are what we have mostly seen so far.

For all the many cross gen games we had for a long time, sure, but there's a significant leap in overall presentation values with plenty of the actual next gen titles we've started getting. Just because we dont have direct 1:1 screenshots to compare with non-existent last gen versions of the games doesn't mean that big leap isn't there. In fact, if you ARE showing direct comparison shots of games with old gen versions, then it's kind of implicitly NOT next gen, is it?

Path tracing is also not gonna feel like such a crazy upgrade when ray traced GI becomes pretty common this generation already, as it's already starting to be. It'll be a nice, but modest improvement. That alone wont remotely justify calling something 'next generation', unless we're seriously watering down that term to be so much less than it used to mean. We've gotten quite good improvements to lighting quality every generation for a while now, but that's generally been accompanied by sizeable improvements in many other aspects of the visuals as well. Without those other sizeable improvements, you're just putting lipstick on a pig. Again, there's still so much room for improvement in visuals outside things that ray/path tracing can help with. Those things cant just be neglected.

 

[Boss]

I don't agree with that at all. If they really get to Nvidia levels of ray tracing and ai (maybe rtx 4000 levels?), next gen will be the last where common people will see meaningful differences. Much more than in this generation, where settings upgrades and higher resolution are what we have mostly seen so far.

View attachment 12819
And I remember seeing screenshots with differences that were even more evident. This is much easier to market to the masses.

The problem you'll find is that there are some people who actually prefer the image on the left... I wish i was joking but I'm not.

 

[neckthrough]

The problem you'll find is that there are some people who actually prefer the image on the left... I wish i was joking but I'm not.

Ugh. I think it's because some people have been conditioned to accept bad lighting.

It's sort of like how we've been conditioned to love movies at 23.976fps. Anything else feels "wrong".

It's also like how my buddy at work thinks the imperial system is more "intuitive" than the metric/SI system.

 

[Charlietus]

For all the many cross gen games we had for a long time, sure, but there's a significant leap in overall presentation values with plenty of the actual next gen titles we've started getting. Just because we dont have direct 1:1 screenshots to compare with non-existent last gen versions of the games doesn't mean that big leap isn't there. In fact, if you ARE showing direct comparison shots of games with old gen versions, then it's kind of implicitly NOT next gen, is it?

Path tracing is also not gonna feel like such a crazy upgrade when ray traced GI becomes pretty common this generation already, as it's already starting to be. It'll be a nice, but modest improvement. That alone wont remotely justify calling something 'next generation', unless we're seriously watering down that term to be so much less than it used to mean. We've gotten quite good improvements to lighting quality every generation for a while now, but that's generally been accompanied by sizeable improvements in many other aspects of the visuals as well. Without those other sizeable improvements, you're just putting lipstick on a pig. Again, there's still so much room for improvement in visuals outside things that ray/path tracing can help with. Those things cant just be neglected.

Current consoles can barely render a convincing effect without having to accept significant compromises. The games that will manage to do it well will probably be in the single digits.

Meanwhile imagine developers having access to great upscaling, ai denoising, SER and much more. Consoles (or maybe graphics in general?) have one last impressive jump left, I can believe in that.

 

[Karamazov]

wonder if it'll be possible to run CP77 path traced in VR with a 5090.

 

[Seanspeed]

Current consoles can barely render a convincing effect without having to accept significant compromises. The games that will manage to do it well will probably be in the single digits.

Meanwhile imagine developers having access to great upscaling, ai denoising, SER and much more. Consoles (or maybe graphics in general?) have one last impressive jump left, I can believe in that.

If you were expecting this gen to usher in perfectly realistic graphics, I dont know what to tell you. Of course there will be compromises. Next gen machines will also require games to be full of visual compromises.

But to say that graphics this gen are nothing but an uplift in settings/resolution is definitely wrong. We're still seeing a reasonable leap in overall presentation aspects compared to last gen, if we look at true next gen titles. And there will be plenty more to come from these systems. Of course cross gen titles wont demonstrate that much, and we definitely had too much of that and for too long, but that wasn't because of the hardware.

Ultimately, we're still quite a ways from having truly perfect graphics, and there's still quite a lot of scope for improvement in many areas. But I think we're gonna run into practical limits of processor gains and affordability very quickly, and likely already feel that pretty strong with the next set of consoles. And unless something radically changes, I suspect those will be the end point for consoles as we know them. Not because they achieve everything we could want, but because it simply wont be possible to do them anymore.