Predict: Next gen console tech (9th iteration and 10th iteration edition) [2014 - 2017]

[DieH@rd]

No matter when new consoles come, a new production process is always "just around the corner". It's better to not think about it.

PS5 will most likely arrive in 2019 or 2020, and it will use "safe" 7nm tech. There is still headroom for some interposer/3D shenanigans, but even that is most likely too pricey for mass production.

 

[Silenti]

No matter when new consoles come, a new production process is always "just around the corner". It's better to not think about it.

PS5 will most likely arrive in 2019 or 2020, and it will use "safe" 7nm tech. There is still headroom for some interposer/3D shenanigans, but even that is most likely too pricey for mass production.

If it is late 2018/ early 2019 - 10nm?

 

[DieH@rd]

It looks like everyone will skip 10nm, the same way they skipped 20nm when they jumped from 28nm to 14/16nm.

 

[HBRU]

I think this 5 nm 2021 viable process of IBM & Globalfoundries will give huge advantages

 

[Prophecy2k]

I think this 5 nm 2021 viable process of IBM & Globalfoundries will give huge advantages

Sure, but the process won't be mature by the time Sony/MS want to launch their PSNextBoxes.

So they'll launch at 7nm and keep 5nm in hand for the inevitable "slim" versions.

 

[Entropy]

It looks like everyone will skip 10nm, the same way they skipped 20nm when they jumped from 28nm to 14/16nm.

TSMC (at this point in time) doesn't offer a High Performance variant of 10nm, but does/will do so for 7nm. So if you have a 150+Watt power budget for your chip, and can fit within practical size limits, you'll likely be better off using a 16nm variant during 2017.
7nm is far more likely for a traditional console chip. We know that 7nm risk production started this April, and that one of TSMCs test chips was a Cortex A72 produced at a HP variant of the node. Which implies that GPUs/CPUs/console SoCs could conceivably go to production on the node relatively soon.
If it makes sense to do so from other perspectives than pure feasability of course.

 

[HBRU]

ah... So no 10 nm.... Console not going to upgrade soon and even a launch of a slim Ps4 pro is difficult. Well good news, so will have a REAL upgrade in 2021

 

[DieH@rd]

There is no need to be dramatic. 7nm is a next target, and we will get our hands on it relatively soon [in 2 years].

 

[Silenti]

Hoping we hear something at E3 time-frame wise. I have been considering investing in an after-market controller, probably the Scuf. AT that price though, I would want it to be good for longer than 18 months or so.

Silly question re: 7nm - How quickly will they be able to increase their production quantities? I assume demand for the process will be from more than just console makers. So what kind of time frame for the 10k's of thousands that would be needed for the first late summer/ fall release? Just assuming 2019 - will there be enough capacity?

 

[Megadrive1988]

I'm not expecting Sony to say anything about a PS5 until 2019 at the soonest --be it E3 2019 or the 25th anniversary of PlayStation (December 2019) or even early 2020.

I think it would be unwise, and unwelcome for Sony to release another console before Fall 2020.

 

[Deleted member 11852]

I'm not expecting Sony to say anything about a PS5 until 2019 at the soonest --be it E3 2019 or the 25th anniversary of PlayStation (December 2019) or even early 2020. I think it would be unwise, and unwelcome for Sony to release another console before Fall 2020.

You really think Sony are going to roll PS4 for seven years? I don't see it. The one seven year generation (for Sony) was forced upon the industry and consumers by a brutal combination of technology (and lack of it) and the economy, which tanked everywhere. For Microsoft it was an eight year generation.

 

[Entropy]

Silly question re: 7nm - How quickly will they be able to increase their production quantities? I assume demand for the process will be from more than just console makers. So what kind of time frame for the 10k's of thousands that would be needed for the first late summer/ fall release? Just assuming 2019 - will there be enough capacity?

It's not a silly question, just difficult to answer.
Apple will want to use 7nm for their 2018 iPhone SoC, which assuming normal cadence means going into full volume production mid 2018. A console SoC would probably be larger and even though designed with redundancy for yields, size still matters, so economic prudence would seem to suggest aiming for production a ways into 2019, letting Apple pay for most of the initial production process tweaking and letting tools improve a bit. While the Switch bucked the trend, consoles seem to preferably be released in time for christmas shopping, so heading towards the mid of 2019 might be a good time for reserving production wafer starts.
But - who knows, really?

 

[dobwal]

You really think Sony are going to roll PS4 for seven years? I don't see it. The one seven year generation (for Sony) was forced upon the industry and consumers by a brutal combination of technology (and lack of it) and the economy, which tanked everywhere. For Microsoft it was an eight year generation.

I'm think it all depends on how this gen pans out. If the PS4 continues to outsell the Xbox and Switch doesn't bother the PS4 sales, what's the hurry? With Scorpio launching this year and after the Pro, I doubt MS will be in the mood to release before Sony. Delaying a year to gain a performance advantage only to potentially give up that advantage by launching next gen earlier isn't going to serve MS well.

If the Scorpio is able to make a serious dent in gobbling up high end users, Sony may feel the need to respond with new hardware sooner. But that only leaves an opening for MS to do a late launch with stronger hardware again. I bet Sony is targeting to release along side with the next Xbox. With its first party titles and the good will it created with the PS4, Sony can shoot for parity in terms of hardware and launch schedule and still win out over the Xbox.

 

[Prophecy2k]

You really think Sony are going to roll PS4 for seven years? I don't see it. The one seven year generation (for Sony) was forced upon the industry and consumers by a brutal combination of technology (and lack of it) and the economy, which tanked everywhere. For Microsoft it was an eight year generation.

Isn't that the point of the mid-gen refreshes though? So that gamers who want to access newer hardware can get some of the benefits of that earlier.

I take PS4Pro and Scorpio to be an indication that the PS4/XB1 gen may even be longer than the PS3 gen.

 

[Entropy]

Isn't that the point of the mid-gen refreshes though? So that gamers who want to access newer hardware can get some of the benefits of that earlier.

I take PS4Pro and Scorpio to be an indication that the PS4/XB1 gen may even be longer than the PS3 gen.

Could be. It's a complex mix of market evaluation, longer term business strategy, litographic process tech, consumer trends and so on.
Focussing on the technical aspect alone is a fallacy.

 

[Prophecy2k]

Could be. It's a complex mix of market evaluation, longer term business strategy, litographic process tech, consumer trends and so on.
Focussing on the technical aspect alone is a fallacy.

Sure, but alas we lowly consumer plebs aren't privy to much insight on the majority of the above. We can only speculate on Sony and MS's next-gen strategy based on the individual market moves they make (e.g. launching the Pro/Scorpio) together with publicly available information about available lithographic process technology.

 

[Entropy]

Sure, but alas we lowly consumer plebs aren't privy to much insight on the majority of the above. We can only speculate on Sony and MS's next-gen strategy based on the individual market moves they make (e.g. launching the Pro/Scorpio) together with publicly available information about available lithographic process technology.

Yup. And it can be both fun and interesting. One just have to beware (oneself or others) mistaking such speculation for anything other than it is.
We are speculating on the progression of a game where we can only observe snippets of what's going on. The major decisions are made in the boardrooms. Lowly consumer plebs, or for that matter developer code monkeys just have to roll with the punches.

 

[Megadrive1988]

On one hand, there are people expecting Sony to respond quickly to Scorpio by releasing PS5 in 2018, or saying absolutely no later than 2019. At the other end of the spectrum, there are people who don't want to see PS5 until 2022 at the very earliest. I do see valid points for either a short generation, or a very long cycle, but that said, I'm kinda in the middle. 2018 seems just too soon, while 2022 (even 2021) is so far away. I'd be happy with either 2019 or 2020 being the time for PS5. Sony is going into E3 2017 with over 60 million PS4s sold through to consumers worldwide since launch.

Keep in mind that we had solid info on Orbis / PS4 spec around Spring of 2012, we have nothing like the equivalent about PS5.

 

[Deleted member 13524]

You really think Sony are going to roll PS4 for seven years? I don't see it. The one seven year generation (for Sony) was forced upon the industry and consumers by a brutal combination of technology (and lack of it) and the economy, which tanked everywhere. For Microsoft it was an eight year generation.

The PS3 didn't get a mid-gen upgrade with >2.5x higher GPU performance. Nor did the X360 get a mid-gen upgrade with >4.5x higher GPU performance and 50% more RAM.

Mid-gens are proof that the 8th gen will be very long.
7/5nm may phase down the 1.8TFLOPS PS4 and 1.3 TFLOPs XBone as home consoles, but then they may have the headroom to transition those performance targets into handhelds.

On one hand, there are people expecting Sony to respond quickly to Scorpio by releasing PS5 in 2018, or saying absolutely no later than 2019.

Sony has a >2:1 marketshare advantage over Microsoft this gen and Scorpio's visuals over the PS4 Pro will be equivalent to PS4->XBone at best. This means it's not a generational difference, mostly because CPU performance gap isn't that big and especially if devs make heavy use of FP16 in the Pro (which has 30% higher throughput than Scorpio).
They won't need to react.

 

[Tkumpathenurple]

The PS3 didn't get a mid-gen upgrade with >2.5x higher GPU performance. Nor did the X360 get a mid-gen upgrade with >4.5x higher GPU performance and 50% more RAM.

Mid-gens are proof that the 8th gen will be very long.

Not really. They're an indicator of that possibility, but not proof. They're only proof that people want newer hardware for prettier visuals and that the console manufacturers think there are enough of those people to warrant release.

I'm still of the opinion that we'll see the PS5 in 2019: enough time for the PS4 to have lasted the length of a traditional generation; establishes a 3 year cadence, which gives security to launch purchasers of the base and mid-gen models; in time to take advantage of a fairly mature/stable 7nm fabrication process; likely too early for Microsoft to launch another Xbox, so should give Sony a year alone on the market.

As long as it's backwards compatible, I think the stars have aligned perfectly for the PS5 to be wildly successful.