Chances of NextGen Console Launch in 5 years?

7nm on AMD is coming the end of 2018. Navi Gpu will launch the beginning of 2019 with upwards of 20TFlops power. 2020 ps4 with that power, 2021 xbnext with 25-30Tflops, both with have some for of HBM and 512+ gb/s busses.
 
PS5 in 2019 for me is a given. If we assume same rate of improvement.

2013 : PS4 (1,8 TF)
2016: PS4Pro (1,8 TF * 2,3 = 4,2 TF)
2019: PS5 ( 4,2 TF * 2,3 = 9,7TF)

However, moving from jaguar to Zen will provide a very nice power increase from Jaguar. Probably Sony will produce a console in the 10-11 TF range.
 
PS5 in 2019 for me is a given. If we assume same rate of improvement.

2013 : PS4 (1,8 TF)
2016: PS4Pro (1,8 TF * 2,3 = 4,2 TF)
2019: PS5 ( 4,2 TF * 2,3 = 9,7TF)

However, moving from jaguar to Zen will provide a very nice power increase from Jaguar. Probably Sony will produce a console in the 10-11 TF range.
By this logic Microsoft should have a 27TF Xbox in 2020 :runaway:
 
Oct-rate uint4 shading

With a throw back to black-and-white stylistic gaming or throw back to moiré pattern gaming they could roll with "dutriginti-rate bit shading".
 
Keep seeing arguments that consoles not hitting 60fps is a primarily a cpu problem and not a developer choice. Zen in next gen consoles isn't a gaurantee of 60fps. Developers are still prioritized graphics in in single player aspects of games.

The only way I can see 60fps becoming standard is if 30fps starts detracting from the increasing realism of games. Though I still think animations are currently more of a problem in that aspect.
 
Keep seeing arguments that consoles not hitting 60fps is a primarily a cpu problem and not a developer choice. Zen in next gen consoles isn't a gaurantee of 60fps. Developers are still prioritized graphics in in single player aspects of games.

The only way I can see 60fps becoming standard is if 30fps starts detracting from the increasing realism of games. Though I still think animations are currently more of a problem in that aspect.
If VR is the next big thing, then frame rate will matter. Eventually VR will be the "thing", but we aren't there yet. When it happens I don't really know. The issue is that it takes away from consoles being in the living room. there are some issues to figure out.
 
Well, power is king. we already here a lot of talk about PS5 crop up suddenly, why? well it's obvious, because now Xbox One X has the power crown and this isn't really a tenable situation in the long term for the other side. At least, over the longer term.

So yes, some form of more powerful PS is inevitable now. Within the next 5 years.

And then the shoe will presumably be on the other foot and same rules apply...

The interesting thing is, if PS5/Xbox Two breaks BC with PS4/pro/Xbox One, either/or, both?. And how all that plays out one way or another.

Scorpio is so powerful though, it does make it difficult to overshadow by a significant margin within the next 2-3 years. Lets say Sony decides they want PS5 in 2018. Can they markedly beat 6 TF and 12GB RAM? Not that I know of, it would be pretty difficult squeezing 8-10TF and 16GB RAM in a console anytime soon for a decent price.

Of course you dont even really need a decent margin, lucky for Sony. 7 Teraflops would be enough in a vacuum.

But yes, DEFINITELY at least one new console within 5 years. That's easy. The more interesting question to me is if we are truly going to get iterative consoles every 2-3 years, BC/FC forever, if one side does and the other doesn't do BC/FC forever, (lets say PS5 is traditional console without BC with prior gens, while Xbox 2 is a "Scorpio plus" that maintains FC/BC back to XOne) which does the market choose? and just how all that plays out.
 
7nm on AMD is coming the end of 2018. Navi Gpu will launch the beginning of 2019 with upwards of 20TFlops power. 2020 ps4 with that power, 2021 xbnext with 25-30Tflops, both with have some for of HBM and 512+ gb/s busses.

PS4 had 1.8TFlops near the end of 2013 when AMD released a 3.7TFlops GPU for the PC on January 2012 (and the PC had the 5.6TF 290x by end of 2013), 20TFlops on the PC could easily end with a sub 10TFlops PS5, I doubt that Sony wants the 300W style monster GPUs added to their console, specially with people so keen on keeping the same price ($399)

starting from the Pro, I think it will be difficult to have something that looks like a proper new gen for regular prices in 2020, it's probably realistic by 2020 to have a 3x faster CPU around the same power, but the rest, not so sure.
 
Scorpio is so powerful though, it does make it difficult to overshadow by a significant margin within the next 2-3 years. Lets say Sony decides they want PS5 in 2018. Can they markedly beat 6 TF and 12GB RAM? Not that I know of, it would be pretty difficult squeezing 8-10TF and 16GB RAM in a console anytime soon for a decent price.

A process shrink allowed the PS4 Pro and Xbox One X to exist, I don't see why you don't think the next one will allow another significant increase in power. The TF increase from the previously most powerful PS4 on 28nm to Pro on 16nm three years later was over 2 times. Why would you expect the One X or even Pro on 16nm to PS5 on 7nm increase to be less?
 
PS4 had 1.8TFlops near the end of 2013 when AMD released a 3.7TFlops GPU for the PC on January 2012 (and the PC had the 5.6TF 290x by end of 2013), 20TFlops on the PC could easily end with a sub 10TFlops PS5, I doubt that Sony wants the 300W style monster GPUs added to their console, specially with people so keen on keeping the same price ($399)

starting from the Pro, I think it will be difficult to have something that looks like a proper new gen for regular prices in 2020, it's probably realistic by 2020 to have a 3x faster CPU around the same power, but the rest, not so sure.

And the fact that AMD was stuck on the same node and architecture for so long has nothing to do with this?
 
On paper, an 8x increase in pure grunt might sound impressive. Unfortunately I do not think it buys you an awful lot of visual oomph these days. Certainly not the kind of leap you'd expect from a generational leap. And that's if we didn't have to waste a huge chunk of said power solely for pushing more pixels in the first place.
 
And the fact that AMD was stuck on the same node and architecture for so long has nothing to do with this?

given that 28nm GPUs only started with the 7970 in January 2012;
when the consoles were launched I think it was not really a factor, it became unusual later in 2014 when Nvidia made Maxwell still on 28nm and so on, you would expect 2014 to be 20nm time for GPUs.

again late this year we are going to see Xbox One X with 6TFlops while the top AMD PC GPU is going to be around 13TFlops, and the Xbox One X have the benefit of higher price and being sold as premium version of the console, a regular new gen PS5 would likely have to be priced lower, so the 8TFlops for a PS5 in 2020 that digital foundry mentioned as target for 4K sounds more realistic than something higher, and with 8TFlops it's hardly going to look like a typical new gen compared to even the Pro, but OK against regular PS4
 
If PS5 was just 8TF, I'm not sure if it's worth Sony releasing it, better of waiting.
The trouble is it's not just about TFlops, the 1X has enough memory for 4k assets if we except that as a fact for now.
So an 8TF console, with the same assets, I'm just not sure how much of a visual difference there would be at 4K?

Although there would be a big jump from 4Pro, but I think people buying 4Pro would be ok with the fidelity untill a big enough one over 1X becomes available.
Especially if your talking a full gen reset.

Cerny has said need about 8TF, so maybe he was just throwing that out there.
Maybe with CBR (which multiplies the performance), and HBMCC (32GB m.2/SSD) and 12-16GB low cost HBM, may be able to release a $399 console?
I still don't think 8TF would be enough, but willing to be told otherwise.

Anything faster I suspect would result in a $499 price point in a 2020 time frame. Large GPU and Zen cpu.

This all goes for an X2 also.
 
Maybe this feeds into my preferred outcome: regular and Pro at launch.

A cheap base PS5 with a performance lead over the X1X, but sold at a loss for the mass market, and a Pro model for the people that bought the PS3 at launch, sold at a profit or break-even.

Sure, the PS4Pro has some architectural improvements that weren't around in 2013, but brute force and extra silicone could have come close enough to making it viable then, especially considering you could have built a PC in 2013 that would still outclass the X1X.

I'm certainly of the opinion that the PS5 will launch in 2019: too early for XboxTwo, long enough to make the Pro exactly a mid-gen console.

If I recall correctly, the 7nm node is going to be in use next year and ought to be developed enough for widespread use the year after. Also, I would expect some very low power iteration of Zen to be well developed within a couple of years, considering that it's already doing quite well on that front.

From what I've read on here, there's contention as to the form that the memory will take - GDDR6 or some form of HBM - as well as the amount, but it seems quite widely accepted that it will be one of those and with a goodly chunk more than is in the current consoles.

So there's a new process node coming next year, a new CPU architecture is already on the market, new memory types are due to usurp GDDR5, and GPU's are always developing on a performance per watt curve.

So yeah, every component will be begging to be assembled into a new console within the next couple of years.
 
Back
Top