-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
Next gen PS5 and next Xbox launch speculation - Secret sauces spicing 2019
- Thread starter Phoenix Splash
- Start date
- Status
Here it is ... Variant (2) opts for a smaller die size ...
Revision 3 changes:
Memory speed GDDR6 14GB/s chips for all consoles
Reduced CU count for PS5 in general compared to Xbox Pro
Added Variant (2) with lower CU count but higher clock speed
Your estimates are based on a 331mm2 Radeon 7 (vega) using HBM memory. S much wrong with this.
1. Next gen is based on Navi. Which we at least know lack some of the things that go into your typical Vega GPU.
2. Next gen APUs should be expected t be at least 340mm2 (just taking the average of the PS4pro is 320mm2 and the XB1X 360mm2)
3. The GPU you used as your base is using 16GB of HBM, a chunk of its die space is eaten up by memory interfaces. Next Gen APUs is not using HBM.
It makes no economic sense yet. Sony would rather sell their console 100 dollars cheaper.
SSD prices in big bulk are going to be way cheaper in 2020 for Sony than 100$ per unit. You can buy 1TB SSDs now for 110$~ and they have to have some sort of Storage in there and I'd be surprised if the difference for a 1TB HDD and 1TB SSD in late 2020 was very big. Of course it is notable when you plan on selling 100m consoles then the delta adds up
Yup,but i expect some sort of flash storage, 64-128 GB,for OS & fast loading games.Plus 1TB HDD,of course.
Fuck i'm so tired of seeing graphs, tables and back and forths between members over tweets that read like high level calculus.
Can someone please leak something already lol
Can someone please leak something already lol
https://www.pcgamesn.com/xbox-one-games-on-pc this is weid, but some have been guessing ms might do something like this. I wounder what cpu/gpu resource usage compared to the windows version is like.
While the cost difference isn't too high, the failure rate might be higher since you'll have to go from a single platter to two platters.
Two platters will also add more vibrations, though probably won't matter.
If there is headroom for more storage, they'll probably add it.
(If this is what you meant.)
Last edited:
Albert Penello - Insight on specs and price planification for the hardware
I'm not going to comment on the specs of course. But there are couple things in here that I felt Brad took a pretty hard stance on which would be counter to my experience.
First, Sony and Microsoft know exactly the prices and specs they intend to launch, and they know it before a contract with AMD is ever signed. An enormous amount of diligence is done on a process like this internally - and figuring out the pricing and specifications of what you can build are, like, fundamental to the whole process. It's literally Step 1.
The reason is they have modeled the entire architecture and are building system components in parallel - not in series. So they have to know roughly where everything will land so that the motherboard, cooling system, case design, fan speeds, radio antennas, and countless other components all land to support the intended price and performance targets. Margins on console are super thin so there is not a ton of room to make major changes late in the program.
Now - things can change. But those changes are almost always in the margins. In the case of the Xbox One for instance, the entire case and cooling system was way overdesigned (obviously given the size!) which allowed the team to increase the clock speeds after the initial parts were tested. This was not part of the plan, and had the case been designed to precisely hit the target there would not have been the headroom to change the clock speed. And on top of that, there was a huge amount of time spent calculating the cost of that change - because even something as small as a 10% clock increase could have more than a 10% yield implication both at launch, and over the long term. So these things are not taken lightly.
So I think it's important to know that specs and prices are set pretty early in the process. Yes, things can change and evolve, but it's generally small tweaks because the implications of doing a major change late in the process are very risky. This is why any idea that Xbox One X made any change or reaction based on the Pro shows a fundamental misunderstanding of the timelines HW works on.
I don't think you need an SSD. Much in the same way you don't need a UHD Blu-Ray player. If I'm going to pay more, it's going to be for stuff that makes actually playing the game better, not just loading times. Dolby Atmos and Vision support would be something I pay more for if that was a sound and HDR system I paid for in an instance.
Far as a I recall, some people said it was a guy that just had some time on his hands with Visio.
Appreciate the input. Essentially, it sounds like there is a deal of flux, but the current department of leaks aren't in the know on what that range actually is and are spitballing.
Far as a I recall, some people said it was a guy that just had some time on his hands with Visio.
Appreciate the input. Essentially, it sounds like there is a deal of flux, but the current department of leaks aren't in the know on what that range actually is and are spitballing.
I think because of the architecture similarities, better scaling in game engines etc cross gen is going to be a big thing going forward. Maintaining things for the old Cell/PPC architecture while trying to build things on X86 was a big burden on developers, so it was in their best interest to cut ties with the previous gen as quickly as possible.
Now that the next gen is going to be just simply more powerful X86 boxes, I don't see them ditching things as fast this time around– especially because indy games and things like Fortnite, Apex, Rocket League etc don't require next gen power to run. I can also see Sony trying to keep the PS4 on the market as long as possible and as cheap as possible so people who can't afford a PS5 at least invest in a library on PS4 that can easily move over to a PS5 someday.
Always good to see someone commenting who was an actually integral part of a console launch process!!
Last edited:
Indeed.And if Sony,as Matt told us,initially wanted 2019 launch,PS5 specs were probably locked long time ago...?
Well if that's true, and they pushed out a year, then they either changed their mind about the spec, or something they wanted to do could not be done in time or on cost. I doubt a 1 year slip means they are going to ship the same thing as they had planned a year earlier.
Wow Albert. Awesome post and very informative on the whole process of preparing for a new gen. Much appreciated.
Edit:
Also Albert, without giving anything away. From what you know so far, do you think the jump from this gen to next gen will be big enough to WOW gamers or can we expect a modest leap?
Interesting,thanks!
Can you tell us,please,at what point before console launch specs are usually fully locked(except minor clock changes that you mentioned): 1-1.5 year or so?
I give it ten pages and a few people will forget, then everyone else will be trying to find Albert's post.
What if that process involves a 6TF GPU and 12GB Ram. This forum will be the farthest thing from calm LOL.
Hm. It's hard to say. Moore's Law is slowing - the '90's and early '00's were an amazing period for silicon development - the rate of change from Gen 6 to Gen 8 was pretty remarkable.
But Moore's Law talks about density, not about cost. At this point, with 16nm in market and 7nm on the way - while you do have a rough doubling of the density it's not that much more cost effective. And rumors about 5nm and lower will likely be more expensive to produce than 7nm or even 16nm. So there will be a slowing in how much more performance you can squeeze into that 350mm2 space in a console, not just from a raw compute perspective but also due to cost.
This doesn't really affect the PC because they just have so much more area to play with on the cards - and they can charge a lot more $$.
And I mentioned this earlier - there are a lot of market forces that used to benefit consoles (like Hard Disc drives and spinning physical media) where game machines were just a small fraction of production, but today consoles are the largest (and maybe soon the only) recipient. It's totally conceivable that during the next generation of consoles, standalone Blu-Ray and DVD players fade nearly away.
Really if you step back, you can sort of observe this yourself. There were many who suggested the early XBO and PS4 games didn't look that much better than late gen 360 or PS3 games. And further you hear this between the Pro/X and PS4/S versions of games. Depending on TV, viewing environment, and frankly the viewer themselves, it can sometimes be hard to tell the differences when you're not looking side-by-side.
The flipside is that I think there are some good rendering techniques that are "cheaper" from a compute perspective directly in the new HW, and developers are clearly taking advantage of different reconstruction techniques. So it's possible that developers who decide to give up native resolutions for really good reconstruction techniques + innate speed improvements + improved HW techniques could give games a pretty good visual leap.
It also depends on where you are coming from. If you're on the base consoles then probably. Do I think it will be as big as the PS1 -> PS2, or Xbox -> Xbox 360? That I'm less sure of.
But game developers are amazing. Metro, RDR2, GoW - these games on the premium systems are pretty outstanding and unleashing them on the new stuff may surprise me.
Without a doubt, consoles as we know them today are going to change. Not because Sony, Microsoft of Nintendo are bad companies who don't like consumers or want to get out of the console business, but simply because the industry that supplies consoles is changing and the factors that go into making these machines is not the same.
Okay Albert Penello, since you got time today...
Simple Question: Would you market a machine as 1080p in 2020?
Simple Question: Would you market a machine as 1080p in 2020?
I don't like the word "locked" because they really aren't "locked" until mass production. That's usually (and man this one I'll have to think about) not until about maybe ~6 months before release. Units that are near-production quality for testing and validation are more like ~9 months from street date. At that point, small tweaks may be happening so there could be very minor differences. So I take the word "locked" very seriously.
But, you can make very small changes right about a year before launch. The variables for a full-next generation console are MUCH different then what we were dealing with on Scorpio, which we were pretty confident in early on (and which allowed us to announce so early), in not the same situation.
This is the nuance of my early post. The specs for these machines are largely targeted very early, and any changes that do happen that late in the game are usually pretty small. Does that make sense?
Albert Penello, nice calming post.
Good to know Sony may be changing up a gear and Microsoft are on-plan. More importantly games should be a whole lot better.
Good to know Sony may be changing up a gear and Microsoft are on-plan. More importantly games should be a whole lot better.
This x100. Chasing native 4k is such a waste when you can checkerboard up from lower resolutions and use processing resources to push higher quality visuals.
Many thanks for your response Albert. You're a legend.
To this forum? No way. LOL.
That said, I think 1080p is totally a viable target resolution in 2020. TV manufacturers will probably stop shipping 1080p sets in the next few years except for the very smallest sizes.
I don't generally like this "Screen size vs. Viewing Distance" argument for the most part, but I do believe it's true that if you look at where the mass market is buying sets (usually between 42in - 50in), these non-HDR 4K sets make it pretty tough to tell the difference between 4K and 1080p at any sort of reasonable room size. So a very well designed 1080p game, with all the lighting and post-processing turned to 11, may look pretty damn good.
The hard part is actually explaining all that to someone with the required nuance.
You won't get it. Compared to current consoles and even with 9-10tflops it will be a monster because of the new GPU architecture and Ryzen CPU.
First Pro specs were leaked 18 months before the release of Scorpio. 18 months before is pretty early for a mid-gen console. First Scorpio reveal was done after the Pro was leaked.
This brings a natural follow up question: who would you consider to be the launch day consumer outside of folks on forums like this that you could explain it to?
It would be insanely interesting to see how that all works. Like the conversations and discussion the people in charge (or better said: those that make the decisions) talk with each other and decide upon what path/route they go down. How it all comes together. What specific hardware they decide on makes it into the final product etc.
Like from a documentary perspective or being a spy fly on the wall looking over Mark Cernys shoulder.
Like from a documentary perspective or being a spy fly on the wall looking over Mark Cernys shoulder.
I would just like to say thank you for all the invaluable insights you have given us
This seems to be something I've seen posted lately - someone suggested there were rumors of a Pro shortly after the PS4 console launch.
I never heard or saw this. My first recollection of hearing about the Pro was the Kotaku leak in March 2016. We announced Scorpio 3 months later, with board layouts, specs, and even a console ID (shadowed) that even DF will remind people were remarkably similar to the shipping product.
If there was something earlier than that - can you link it? I'd love to see it.
Thank you! It's nice to be able to talk about stuff like this. I'm just as curious what's going to actually ship as everyone else!
Well that assumes someone is planning to actually do something like this.
But if someone WERE planning to do this, I think the question is how much does it cost? If it's cheap enough, 1080p may not matter if it plays the newest games.
Possible to know what lead to the engineers to discover that they can increase the clock by 10%?
It's been a bit of a back and forth on whether one would outright create a 1080p machine based on rumored specs for Next Geneeation.
I was not sure how you'd market a machine dubbed as next generation after 4 years of 4K advertisement across both consoles and how you could sell it beyond just price point itself given where the tv market has headed and consumer marketing in general for games themselves
A 1080p console option wouldn't surprise me. The pc market still has a ton of 1080p players due to that being the most popular monitor size, and hundreds of millions of people still have a 1080p TV set.
If the games are comparable for both, I can see it being a decent seller.
If the games are comparable for both, I can see it being a decent seller.
If Penello has done anything, it's probably relieved concerns of Sony releasing an under powered console. Which is an odd fear to begin with. Either they speced so high that they needed a year to have price come down, or they decided to push out a year to increase things
Agreed. Biggest leap I expect to see is in VR, if we get foveated rendering and maybe 4x the resolution. You combine that with even a modest boost in raw console power, and it's almost like 2 generation jumps at once.
I feel like any '1080p' system would still be marketed as 4K as much as was legally allowed, since it would have 4K output and 4K upscaling, and probably reconstructed or native 4K on plenty of BC games.
This is a really well done and interesting post.
Really thanks Albert for your trustworthy contribution ))
Albert Penello , speaking of console makers determining specification, wouldn't this upcoming generation have been the trickiest to plan due to the high uncertainty of the timeline for 7nm and 7nm+ in earlier years? Is it probable that Sony and/or Microsoft would have done two different designs depending on those timelines, or would they focus on a single node no matter the developments?
Edit: Also, many thanks for your input, and insight, on the topic of console development!
Edit: Also, many thanks for your input, and insight, on the topic of console development!
Last edited:
- Status