Support NeoGAF

[Theonik]

That move makes sense for a post launch model, less so for a launch configuration.
Sony would need to allocate a maximum for OS features that they intend to release over time to games. When they cancel features and/or shrink the OS they can't pass the benefit to games in the same way they can with just putting it on the main RAM.

 

[Shin]

That's not normally how it works, Shin.

Normally the OS reservation, cordoned off at the start of the gen, includes the amount of memory required for the OS services, functions and features, as well as an allocation for future feature additions. That's one of the reasons why the OS reservation on PS4 and XB1 is so much.

Honestly though, i'm highly skeptical that the OS reservation next-gen will need to grow that much. It shouldn't really need to be more than the 3GB on the XB1X, since 4k game streaming is possibly one of the larger consumers of the OS memory footprint. I can't see the PS5 OS offering any features beyond than, which will require much more in terms of memory.

If anything, I wonder whether Sony will opt to implement a reasonably large pool of cheap DDR4 on the PCB for OS features. Since it's such a waste to carve off such a large portion of your expensive high performance memory for OS functions that simply don't require that level of performance.

I know you reserve a chunk of the total pool during designing phase in case you need it in the future and free that up when you don't/years go by.
Just giving an example as to how less 16Gb is, most likely it will be around 3Gb because even MS that's doing a lot more with their OS only reserved that.
DDR4L should be more than fast enough to handle OS and background tasks, if anything such an approach would come in a lot cheaper, that would leave you with 16Gb for just games.
If PS4Pro is anything to go by then it's not far fetched that they will take that route (DDR5 is slated for 2018? not that I see them using that just curious).

To be fair I can only read what is in front of me and it is confusing and contradictory! We have Pachter saying 2019 half step, Thong saying 2018 >10TF and Matt saying not 2018 and be more realistic about specs.

None of this is making logical sense to me right now. And that's before we get into wondering about dev kits etc. I'm sure Sony would change things a bit (compared to the leaks of 2012) to try and limit leaks!

The only thing that we can take as a fact is that PlayStation is discussing which direction to take internally and that there will be a PS5 and to some extent when things started moving the rest are opinions.
As to what Matt knows or doesn't know, I wouldn't be surprised if the pitch went something along the lines of what would you guys like to see in a next-gen system and mentions of BC.
That node shrink is important and as long as they are leading and sales aren't declining they have no reason to release it even if I'm willing to spend $1000 on a 12-16TF PS5, it has to make sense from a business point of view.
I don't know how much the wait will save Sony, but a HDMI license fee per electronic is $0.15 now this might not seem much but that's already 9M at 60M consoles sold, that's one of the cheapest things now imagine the rest.

 

[Graphics Horse]

That move makes sense for a post launch model, less so for a launch configuration.
Sony would need to allocate a maximum for OS features that they intend to release over time to games. When they cancel features and/or shrink the OS they can't pass the benefit to games in the same way they can with just putting it on the main RAM.

If they are repeating the PS4 family model entirely then yeah, but I wouldn't know if that is the only option, wasn't to Microsoft but they might have gone a bit nuts.


Edit, but basically my thinking was sticking a single 2GB chip of ddr4 would be possible, which is still not going to consume the whole OS, so doubling that in a revision would still give an equivalent benefit

 

[Shin]

I'd be more curious as what they do with DualShock 5, light bar disable option in OS since that was requested a lot.
Make it easier on developers like MS did with a FPS counter on the SDK and other shiznit?
Touchpad made of LCD with whatever info on it, clickable still and retain it's current features?

We should make our own general PS5 thread since people are interested in the subject.
There's a lot to cover even if it will be more of a wishlist, idea box/general discussion, sucks that developers and platform holders decide everything behind the screen.

 

[TheThreadsThatBindUs]

I know you reserve a chunk of the total pool during designing phase in case you need it in the future and free that up when you don't/years go by.
Just giving an example as to how less 16Gb is, most likely it will be around 3Gb because even MS that's doing a lot more with their OS only reserved that.
DDR4L should be more than fast enough to handle OS and background tasks, if anything such an approach would come in a lot cheaper, that would leave you with 16Gb for just games.
If PS4Pro is anything to go by then it's not far fetched that they will take that route (DDR5 is slated for 2018? not that I see them using that just curious).

16GB of RAM is still quite a bit of RAM. And especially if Sony does something similar with the PS5 as with the Pro, i.e. having a second large cheap pool of memory for the OS functionality.

The way AMD and NVidia have been designing GPUs around larger local caches and conserving memory bandwidth, I can very much see a situation where games next-gen aren't even needing to use the majority of the RAM allocation they have; outside of simply storing as much data as they can locally to the execution units to improve load times.

Even with next-gen assets, e.g. textures, meshes and lighting data, I think we'll be unlikely to see any developers complaining about a lack of RAM quantity next-gen, provided a minimum baseline of 16GB.

I think memory bandwidth will be more important, in terms of them being able extract the most out of the hardware that is there. It's just a shame that HBM seems so far to be out of reach of the next-gen consoles.

 

[TheThreadsThatBindUs]

I'd be more curious as what they do with DualShock 5, light bar disable option in OS since that was requested a lot.
Make it easier on developers like MS did with a FPS counter on the SDK and other shiznit?
Touchpad made of LCD with whatever info on it, clickable still and retain it's current features?

We should make our own general PS5 thread since people are interested in the subject.
There's a lot to cover even if it will be more of a wishlist, idea box/general discussion, sucks that developers and platform holders decide everything behind the screen.

I'm pretty annoyed this gen., that devs didn't make much use of the touch pad.

I felt that it was excellent in Assassin's Creed Black Flag, for navigating the map, and things like that. It's such a shame more games didn't do more than simply using it as an extra button.

 

[00ich]

I'm pretty annoyed this gen., that devs didn't make much use of the touch pad.

I felt that it was excellent in Assassin's Creed Black Flag, for navigating the map, and things like that. It's such a shame more games didn't do more than simply using it as an extra button.

It has been like this every generation since the PS2
-Pressure Sensitive Buttons
-motion sensors (six axis)
-back touchpad

 

[TheThreadsThatBindUs]

It has been like this every generation since the PS2
-Pressure Sensitive Buttons
-motion sensors (six axis)
-back touchpad

But...

Pressure sensitive buttons (at least on the triggers) have done wonders for racing games at least.

Motion sensors still get some mileage from time to time, and at least have been leveraged in the PS3 gen with Move games, and PS4 gen with some VR games. Also, sixaxis steering is great too.

I assume you mean, the back touchpad on the Vita? If so then you might as ell as just include the entire Vita itself on that list of underutilized tech... lol.

The DS4 touchpad, however, is actually a genuinely great idea, and I would have loved to have seen more analogue "mouse-like" functionality added to games where it made sense (e.g. cursor control of menus, map and inventory screens). It just feels like such a letdown as I imagine this functionality would have been trivial to add (especially with most AAA games available on PC and so playable with a M&KB anyway).

 

[Theonik]

If they are repeating the PS4 family model entirely then yeah, but I wouldn't know if that is the only option, wasn't to Microsoft but they might have gone a bit nuts.


Edit, but basically my thinking was sticking a single 2GB chip of ddr4 would be possible, which is still not going to consume the whole OS, so doubling that in a revision would still give an equivalent benefit

I guess it would just end up like the flash memory on the PS3. Launch models might have extra ddr4 RAM and they just remove it for cost savings down the line when it becomes clear they don't need it. But these moves make more sense when you know how much memory you need and not before that. Otherwise you must commit to a lot of extra memory since you can increase the RAM for games but not take it back.

 

[THE:MILKMAN]

Matt if you are around do you know if Sony (and MS for that matter) are for sure targeting 7nm and won't do anything for the next consoles until that node is ready?

 

[onQ123]

Matt if you are around do you know if Sony (and MS for that matter) are for sure targeting 7nm and won't do anything for the next consoles until that node is ready?

It wouldn't really make sense for them to try to make a next gen console before 7nm /10nm especially after the fact that PS4 Pro & Xbox One X is already using 14nm / 16nm.

 

[THE:MILKMAN]

It wouldn't really make sense for them to try to make a next gen console before 7nm /10nm especially after the fact that PS4 Pro & Xbox One X is already using 14nm / 16nm.

Sure and I agree but the thing is that 7nm looks like it isn't happening in high volume until late 2018 or early 2019 for Apple who will have the first wave of wafers from TSMC anyway. Add on time for the process to mature and any delays (these seem to always happen) and it means 7nm might not be available until 2020 at least for Sony/MS.

If that is the case why is there even any talk at all about PS5 and why have two analysts wasted their time saying PS5 is coming in 2018/19?

 

[TheThreadsThatBindUs]

Sure and I agree but the thing is that 7nm looks like it isn't happening in high volume until late 2018 or early 2019 for Apple who will have the first wave of wafers from TSMC anyway. Add on time for the process to mature and any delays (these seem to always happen) and it means 7nm might not be available until 2020 at least for Sony/MS.

If that is the case why is there even any talk at all about PS5 and why have two analysts wasted their time saying PS5 is coming in 2018/19?

Your timeframe expectations are off... woefully.

7nm availability is NOT the critical path for PS5/nextbox launches.

 

[THE:MILKMAN]

Your timeframe expectations are off... woefully.

7nm availability is NOT the critical path for PS5/nextbox launches.

When do you think 7nm might be viable for a 75-100W 300+ mm^2 SoC in 500,000-1,000,000 units a month quantities (presumably from TSMC)?

And are you saying PS5/Xbox Next might not use 7nm? I'm genuinely interested in what you and others think about this.

 

[TheThreadsThatBindUs]

I've said it before, as have others.

Manufacturing ramp up from early 2019, ready for a late 2019 launch.

7nm manufacturing availability won't be the reason these consoles are targeted at a 2020 launch, if they do release in 2020.

If the PS5 is 2020, it's because they bit the bullet on something un-conventional in the design, e.g. HBM or something even further out of the left field (like an MCM design).

 

[mrklaw]

They are doing this in the Pro's larger amount of ddr3 if you weren't aware, so yes that is very possible.

I imagine the core OS needs to be on shared memory. But cache could be a good contender for a separate pool to at least prevent OS reservation from increasing too much - 4k DVR, that sort of thing

 

[THE:MILKMAN]

I've said it before, as have others.

Manufacturing ramp up from early 2019, ready for a late 2019 launch.

7nm manufacturing availability won't be the reason these consoles are targeted at a 2020 launch, if they do release in 2020.

If the PS5 is 2020, it's because they bit the bullet on something un-conventional in the design, e.g. HBM or something even further out of the left field (like an MCM design).

I'm not optimistic that 2019 is realistic looking back at the timeline of TSMC's 16nm deployment. 16nm went into risk production in the late 2013-mid 2014 time frame and the first product to launch with these chips were for Apple in mid 2015.

I've read risk production just started for 7nm so we're looking at late 2018 or early 2019 for the first mobile chips in volume for which we just know Apple have first dibs.

Adding a year more for Sony/MS to get a share seems wise?

 

[Shin]

You're forgetting that there's at least 3 different companies with multiple Fabs on different nodes.
That's like saying Apple or other mobile phone manufacturers will use up all the capacity these Fabs have.
Mobile SoC is a lot smaller and less complex one could say the gain in node shrink makes less of an impact, there's also the "off" nodes which they tend to utilize just fine.
I don't agree that 7nm won't play a big part in PS5's launch or design, from a cost effective standpoint you would want to make use of it.
Otherwise they might have to take the loss to keep $399 for example and make it back through Network Services, when they could have used the new node save money there and take less of a hit.
And let's say R&D started last year and they started poking around (hence why Matt got a heads up), we're going to look at 4 years minimum before it's out that puts us at 2020.

I don't see Xbox launching XB2 or whatever before Sony, they got a console coming out later this year which is basically 2018, their next console is probably 2021.
Most likely they've "lost" this generation and whatever profit they can make they are fine with, until then Sony doesn't have to worry about a thing.
Sales decline might be less since there's PS4Pro people can upgrade to and keep things in check to a certain extent until the next new thing comes out.
Add your price cuts and whatever big name games (GTA6?, TLoU2) and whatever else and things just might look peachy, so I wonder where in that would Sony feel the need to rush to the market?

I should probably add that new process nodes cost a lot of money for the Fabs themselves (equipment) as well as for those first making use of the new process.
You'd want someone else to go first and let cost go down, given that 7nm is due next year 2 years later the process should be more than mature and cheaper.
By that time TSMC might have 5nm available if it's so important to the mobile market, Apple and the GPU manufacturers will be after that, 7nm process will be more freed-up?

 

[Darklor01]

I honestly feel like it doesn't actually matter what Sony puts out for the next console. It will be powerful, it will not be a PC, it will have exclusives, people will or won't buy initially or later. That's pretty obvious to me, and yet, I am pretty sure people will over-react either positively or negatively respective to their perceptions about the system.

It does get tiring sometimes.

 

[Theonik]

At this point we can say with near certainty that a PS5 is not coming in 2018. We are already past the announcement window for a 2018 release and there is a notable lack of the level of information Sony would be sharing with developers if the system was to drop in 2018. This automatically makes a 2019 release the earliest we're going to get but Sony has no reason to rush. It will all depend on what technology is available and how well the PS4 is selling by then.

 

[THE:MILKMAN]

At this point we can say with near certainty that a PS5 is not coming in 2018. We are already past the announcement window for a 2018 release and there is a notable lack of the level of information Sony would be sharing with developers if the system was to drop in 2018. This automatically makes a 2019 release the earliest we're going to get but Sony has no reason to rush. It will all depend on what technology is available and how well the PS4 is selling by then.

Whilst this is almost certainly the case now (Matt says so!) I still think PS5 development will be very similar to what we saw with PS4 Pro i.e. a single final dev kit around a year before release and it might also have a much shorter official reveal/announce to release time too.

Incidentally here is what the PS4 Pro dev kit looks like:

 

[Fess]

At this point we can say with near certainty that a PS5 is not coming in 2018. We are already past the announcement window for a 2018 release and there is a notable lack of the level of information Sony would be sharing with developers if the system was to drop in 2018. This automatically makes a 2019 release the earliest we're going to get but Sony has no reason to rush. It will all depend on what technology is available and how well the PS4 is selling by then.

Yeah 2018 is way too soon, I can't believe anyone thinking it'll launch then, Pro has only been out for just over half a year.

 

[Theonik]

Whilst this is almost certainly the case now (Matt says so!) I still think PS5 development will be very similar to what we saw with PS4 Pro i.e. a single final dev kit around a year before release and it might also have a much shorter official reveal/announce to release time too.

Incidentally here is what the PS4 Pro dev kit looks like:

You don't get final devkits out until a few months before release typically a year if you're lucky. But developers will be briefed earlier than that on what is planned to gather feedback and whatnot. You can use 3 parameters to gauge release dates 1) Tech availability at the time, what can x ship at that timeframe, what is their target price? 2) Info availability. From announcement to release you are looking at 6-8 months at the minimum. Often close to a year, you will usually have devs briefed months before announcement and the information is wide enough in circulation that you can tell some 12 months before whether it's happening. 3) Market indicators. What would the system offer their business, do they need to launch it now? Can they wait? What kind of projections have they given stockholders? Projecting sales declines can indicate they are bringing out new hardware.

Etc etc.

 

[McHuj]

At this point we can say with near certainty that a PS5 is not coming in 2018. We are already past the announcement window for a 2018 release and there is a notable lack of the level of information Sony would be sharing with developers if the system was to drop in 2018. This automatically makes a 2019 release the earliest we're going to get but Sony has no reason to rush. It will all depend on what technology is available and how well the PS4 is selling by then.

Ps5 isn't coming in 2018, but PS4 was announced and released in the same year, 2013.

 

[THE:MILKMAN]

You don't get final devkits out until a few months before release typically a year if you're lucky. But developers will be briefed earlier than that on what is planned to gather feedback and whatnot. You can use 3 parameters to gauge release dates 1) Tech availability at the time, what can x ship at that timeframe, what is their target price? 2) Info availability. From announcement to release you are looking at 6-8 months at the minimum. Often close to a year, you will usually have devs briefed months before announcement and the information is wide enough in circulation that you can tell some 12 months before whether it's happening. 3) Market indicators. What would the system offer their business, do they need to launch it now? Can they wait? What kind of projections have they given stockholders? Projecting sales declines can indicate they are bringing out new hardware.

Etc etc.

What you describe has been the case for consoles up until PS4 where it changed with the adoption of x86 and standard (customised) PC hardware. Sony also like to make their own dedicated dev kits and even for the mid-gen Pro they produced a new dev kit almost a year before launch. PS4 had Frankenstein PC-based dev kits as early as 2011 but it wouldn't surprise me if PS5 doesn't need to do this with much more standardised tech.

Ps5 isn't coming in 2018, but PS4 was announced and released in the same year, 2013.

And for Pro they reduced this further to 2 months. Might not be the case for PS5 but still...

Edit: A related question to Matt or anyone in the know. Above I posted a picture of the Pro dev kit and wondered if this serves both standard PS4 and Pro or are two separate kits required by devs?

 

[Theonik]

Ps5 isn't coming in 2018, but PS4 was announced and released in the same year, 2013.

The PlayStation Meeting was in February 2013 for a November release but the event itself was announced 20 days before hand with heavy hints of it happening months before that and with tons of developer buzz even before that. We simply have 0 indication that points to anything before a 2019 release but most evidence could point to an even later release in 2020.

 

[FLAguy954]

I'll take a full generational leap at the cost of no BC.

I don't have to sacrifice BC to get a full gen leap on PC but opinions.

 

[Bitch Pudding]

I don't have to sacrifice BC to get a full gen leap on PC but opinions.

This time BC will be vital for Sony to transfer their precious 25mn+ PS+ subscribers to the new platform.

Unlike PS3, a huge share of PS4 games have been bought digitally, and PS+ subs have gathered a nice library of free games as well in the meantime.

If their users can transfer those libraries without hassle to PS5 that would be a huge advantage.

It would also mitigate the inevitable "has no gamez" phase at the beginning.

 

[kyser73]

I'm a bit surprised/confused about this. If they are coming later than sooner then logically I think they are waiting for 7nm but then for that I would think 2020 is the realistic time frame looking at the timeline of 16nm and Apple taking the first volume production runs of new nodes.

If the above is close to right then why are we hearing anything about PS5 this early? Did we get as much talk and articles about PS4 in 2009/10? I don't think so?

We haven't heard anything concrete beyond 'Yes we're working on it.' from Sony, the rest is speculation & that weird media flurry thing where a single sentence in an interview becomes a dominant news item for 48 hours - that's the noise.

Why 2019? Sony said at the start of the generation that 7 years was too long, so unless there's a substantial change in buying patterns for PS4 and/or they can't hit their power/price goal thats why many people are arguing for a 2019 launch.

Matt may well have been party to a briefing document that offers a top line roadmap & where Sony estimate the target specs to be. These may change before the dev kits (which themselves will iterate between prototypes & production-model units) are delivered.

You're also conflating the issues Switch is having with a traditional console. Switch has a bunch of components they have to bid against multiple phone companies for (e.g. Screens) which complicates their process.

I'll state my prediction:

November 2019
10-12TF
16-24gb GDDR6
3.0ghz Ryzen
2TB 7200rpm HDD

Plus some nice custom widgets for developers to play with.

I also think they might try & push the price point to $459 as something Matt said earlier about there not being a need/expectation to explode out of the gate sales-wise might mean Sony feel they can push a higher initial price point.

 

[Souldestroyer Reborn]

It's not coming in 2018.

C'mon Matt, that's a bone that has already been passed around 6 different dogs.

Anything with some juicy meat?

 

[Shin]

What kind of projections have they given stockholders? Projecting sales declines can indicate they are bringing out new hardware.

The forecast for FY18 (18m?) is lower than FY17 (20m?), if it's lower for FY19 then it might be an indication.

I also think they might try & push the price point to $459 as something Matt said earlier about there not being a need/expectation to explode out of the gate sales-wise might mean Sony feel they can push a higher initial price point.

I could see $449 happening mainly because of the increase in cost (memory chips/amount), you might save on 7nm but you're also adding more to the box which could raise the price or keep it the same.
With BC being a reality people can rest assured about their games library carrying over and/or look better, that could warrant the purchase even if priced higher (your games look and play better etc etc).

November 2020, Feb briefing (I think this is a safe bet because it's 3 years after Pro, people won't feel shafted as much plus they'll understand it's a completely new generation and that Pro was only an option).

12 - 14TF
16 - 24GB GDDR6
3.2Ghz Ryzen 7 xxxx Gen3* 8 Core
2TB 5400RPM Seagate FireCuda SSHD
$449,- $50 loss per unit ($499 BOM, shipping $10, assembly $8, $518 total)

 

[THE:MILKMAN]

We haven't heard anything concrete beyond 'Yes we're working on it.' from Sony, the rest is speculation & that weird media flurry thing where a single sentence in an interview becomes a dominant news item for 48 hours - that's the noise.

Why 2019? Sony said at the start of the generation that 7 years was too long, so unless there's a substantial change in buying patterns for PS4 and/or they can't hit their power/price goal thats why many people are arguing for a 2019 launch.

Matt may well have been party to a briefing document that offers a top line roadmap & where Sony estimate the target specs to be. These may change before the dev kits (which themselves will iterate between prototypes & production-model units) are delivered.

You're also conflating the issues Switch is having with a traditional console. Switch has a bunch of components they have to bid against multiple phone companies for (e.g. Screens) which complicates their process.

I'll state my prediction:

November 2019
10-12TF
16-24gb GDDR6
3.0ghz Ryzen
2TB 7200rpm HDD

Plus some nice custom widgets for developers to play with.

I also think they might try & push the price point to $459 as something Matt said earlier about there not being a need/expectation to explode out of the gate sales-wise might mean Sony feel they can push a higher initial price point.

But surely Sony don't/can't control when node shrinks happen and can't overcome Apple getting first dibs on the wafers?

This is why I'm questioning a 2019 launch on 7nm (it seems highly unlikely to get the required TF boost on even a mature 16nm?). I looked at what happened with 16nm on the PC tech sites and everything points to 7nm not being available and/or ready for consoles for mid-2019 (latest production start date for late 2019 PS5 launch IMO).

Apple get first dibs and then the big high-end GPUs after that then I assume the likes of Sony and Microsoft. One person above talked about other fabs but TSMC are favourites to launch 7nm first so any of those would mean a even later launch.

 

[RCU005]

I'd be more curious as what they do with DualShock 5, light bar disable option in OS since that was requested a lot.
Make it easier on developers like MS did with a FPS counter on the SDK and other shiznit?
Touchpad made of LCD with whatever info on it, clickable still and retain it's current features?

We should make our own general PS5 thread since people are interested in the subject.
There's a lot to cover even if it will be more of a wishlist, idea box/general discussion, sucks that developers and platform holders decide everything behind the screen.

Dualshock 5 needs to be what Xbox One controller is: Updatable, higher quality, maybe an Elite option?. Also, the lightbar needs to go, they should figure out a way to do VR without that light in the controller. The little strip in the touchpad is more than enough to know which player you are (I actually like that one).

 

[Shin]

Retail: $449
Total BOM Cost: $499
Manufacturing Cost: $9.00
Shipping Cost: $9.00
BOM + Manufacturing + Shipping: $518

CPU/GPU: $190.00 Navi/Zen3 APU
RAM: $100.00 - 16 GB GDDR6
Additional RAM: $20.00 4GB DDR4L
Power Supply: $20.00
Optical Drive: $33.00 UHD
Hard Drive: $37.00
Mechanical / Electro-Mechanical: $35.00
Other (Electronic Content): $40.00

Console Sub Total: $475
Controller: $18.00
Box Contents: $6.00

I used PS3 and PS4 BOM and adjusted it accordingly to expected specs and parts and prices.
From that the only thing we don't know is the price of a 4GB DDR4L module and the APU itself.
Smaller memory jump compared to previous generations due to cost, but beefier CPU & GPU.

 

[THE:MILKMAN]

For me the APU and RAM would be quite a lot lower, Shin. $100-$120 for APU and $50-$70 for RAM. Whatever the BOM of PS5 is the retail price will be at or higher than it. Taking a hit on the hardware ship sailed a long time ago!

 

[Darklor01]

For me the APU and RAM would be quite a lot lower, Shin. $100-$120 for APU and $50-$70 for RAM. Whatever the BOM of PS5 is the retail price will be at or higher than it. Taking a hit on the hardware ship sailed a long time ago!

Didn't MS say they were taking a hit on the XBOX One X?
I forget.

 

[THE:MILKMAN]

Didn't MS say they were taking a hit on the XBOX One X?
I forget.

Unless they got a terrible deal with AMD and whoever is providing the RAM I highly doubt it especially at $499. IIRC even Tom Warren said the BOM wasn't that high and he was surprised by the $499 price point.

 

[Shin]

For me the APU and RAM would be quite a lot lower, Shin. $100-$120 for APU and $50-$70 for RAM. Whatever the BOM of PS5 is the retail price will be at or higher than it. Taking a hit on the hardware ship sailed a long time ago!

Sony will get cheaper prices I'm pretty sure because they'll most likely order to match 60M consoles (that's a safe bet of amount that will be sold regardless if they lose/win the gen).
We don't insight to their purchasing habit but IHS has all the information regarding previous PS consoles (the image below is not from them btw), extra: 4GB HBM2 cost $80.
RAM will most likely be provided by Samsung it's why I'm keeping track of their lineup and press releases of their semi conductor business, because they are the cheapest.

 

[Darklor01]

Unless they got a terrible deal with AMD and whoever is providing the RAM I highly doubt it especially at $499. IIRC even Tom Warren said the BOM wasn't that high and he was surprised by the $499 price point.

It seems the actual Quote was that they weren't making money on it. No comment about losing money on it.

 

[THE:MILKMAN]

That table is a blast from the past. I remember it well from the PS3 BOM days!

 

[Shin]

That table is a blast from the past. I remember it well from the PS3 BOM days!

You could have a look at:

And get an idea how/what with prices, from PS4 2013 launch to it's current price, $250?
Lisa Su said they were paying $30 for 4GB (I believe it was GDDR5) so 16GB at $100 seems fair, $25 per 4GB.

 

[THE:MILKMAN]

Honestly I can't see Sony straying too far from the BOM of the PS4 and Pro for PS5 but that is just me.

Here is the Xbox One S to show how some things reduce in cost substantially over time and even the APU on a brand new node is cheaper even with the lower initial yields.

 

[Shin]

I've noticed that APU tends to hover around $100, though that might change with XBOX, then again maybe not.
They could further save money by going with a internal PSU ($6.00 saved), that would generate more heat, ok so vapor chamber cooling.
The system would run a lot more silent (my laptop uses it), secondly they could run the memory clock speeds much higher due to more efficient cooling.
Memory cost hasn't been dropping so I'm expecting an increase to the bill-of-materials there, which is the main reason I think it will be 16GB GDDR6.
The rest doesn't really change as they are basic components, it's the memory, UHD/no UHD and the APU itself that will make the BOM fluctuate.

As I've said before I'm pretty sure both companies are taking a hit on the low SKU's and off-setting that through services and/or the higher-end SKU.

2019 seems troublesome for them to launch, you'd first probably need AMD to have Navi out of the gate in the form of desktop GPU's.
I've yet to see a console launch with a architecture before it's desktop counterpart, so 2019 seems 50/50 in that regard, lower even if Navi is further delayed.
If they go with a full fledged Vega that gives us 64CU (plus maybe 6 added CU's?, so 70), plus the power reduction itself from 7nm (thus less heat generated).
You'd have 70 (compute unite) * 64 (shader/stream) = 4480 (stream processors) * 2x 1340Mhz (instructions per clock) = 12TF, with a vapor chamber setup.
60% power reduction compared to 14nm, Scorpio runs at 1172MHz on 14nm, while 60% doesn't equal the amount of heat reduction they could still bump clocks to 1.3Ghz+ I reckon.

 

[RaijinFY]

I'm a bit surprised/confused about this. If they are coming later than sooner then logically I think they are waiting for 7nm but then for that I would think 2020 is the realistic time frame looking at the timeline of 16nm and Apple taking the first volume production runs of new nodes.

If the above is close to right then why are we hearing anything about PS5 this early? Did we get as much talk and articles about PS4 in 2009/10? I don't think so?

In 2011 we already had some rumblings about the PS4...

 

[THE:MILKMAN]

In 2011 we already had some rumblings about the PS4...

I'm not talking about casual forum chat but actual articles with details or codenames and such from the likes of IGN, Gamespot or Kotaku. If you have links from 2011 though.....? The earliest specific PS4 article I can find is from March 2012 and then by June 2012 we had the big VGLeaks er leak which proved pretty spot on.

 

[Shin]

Even if we use the rough estimate of 1.8TF gain per year, we're still looking at the higher amount of RAM that will bring up the cost, so $449?

 

[TheThreadsThatBindUs]

I'm not optimistic that 2019 is realistic looking back at the timeline of TSMC's 16nm deployment. 16nm went into risk production in the late 2013-mid 2014 time frame and the first product to launch with these chips were for Apple in mid 2015.

I've read risk production just started for 7nm so we're looking at late 2018 or early 2019 for the first mobile chips in volume for which we just know Apple have first dibs.

Adding a year more for Sony/MS to get a share seems wise?

Add a year to late 2018... and you get late 2019...

 

[Marlenus]

I always thought 2019 or 2020 was doable. It will probably be on 7nm as well which is looking like a much better node than GFs 14nm.

Spec wise I think 12 Tflops of GPU power, likely based on Navi. I could see something similar to full vega spec being small enough at 7nm to work as based on Fiji to Vega scaling I would expect something around 250mm^2 which is in the ball park of Pitcairn and Polaris.

8 core ryzen 2 is a likely CPU target unless Ryzen 3 is out by early 2019. Big boost over the Jaguar cores and will stay relevant for a whole gen like the 2500K has done.

Ram I could see HBM as possible so atleast 32GB imo. If GDDR6 then probably atleast 64GB. Vega supports 16GB from 2 stacks so a 4 stack layout could support 32GB today and in a couple of years time it will be cheaper.

HDD probably a 2TB hdd, possibly a hybrid with a small amount of Nand to make the currently played game faster.

That would be a big boost over PS4 pro, let alone OG PS4 and it seems doable.

 

[THE:MILKMAN]

Add a year to late 2018... and you get late 2019...

It wouldn't be that simple. For one it assumes the best case scenario for 7nm when there is nearly always a delay (Even Intel have delayed their recent node shrinks) Sony can't one day in late 2019 flick a start PS5 production switch when Apple finally "allow" others to take some 7nm wafers.

For a late 2019 launch they would need to start up production in mid 2019 to ramp up and produce a couple of million units or more for worldwide launch and all the distribution logistics that come with this.

If Sony go with GF or Samsung then all articles are saying these will be coming later than TSMC for 7nm so in effect about the same time for availability without Apple taking the first wafer runs.

Actually is it likely that Sony would want to release PS5 on a brand new node just released? Have they done such before? PS3 and PS4 both launched on pretty mature nodes, right?

 

[Futurematic]

As has been said fall 2019 looks tricky, but could spring 2020 be pulled off?