Baseless Next Generation Rumors with no Technical Merits [post E3 2019, pre GDC 2020] [XBSX, PS5]

Status
Not open for further replies.
are you serious?
Am I serious that I think a game tester would use a devkit for testing? Yes, of course. The only reason they wouldn't is if one is not available, and then what's the point in testing? And yes I believe that a tester would know the specs of what he's testing, that may not be a fact but it is a truth.

Do I believe in the veracity of that post? Why should I? I take it with a healthy grain of salt like all such leaks. Would you believe it more if it was 9.2TF or if it came in a spreadsheet?

Frankly, unless you are fully committed to the PS5 is a lesser console narrative there's really nothing in that "leak" to believe is not possible or extraordinary.

That said, there is one thing in it that is interesting in that we haven't discussed it at all, the 500GB nand vs 1TB nand for Xsx. I could certainly see where Sony would be having doubts at being at that much of a storage deficit even if they believe it puts them at a lower price point. Could there be a 2013 like last minute doubling of nand? It would make sense in why they are delaying an announcement while trying to get a new deal done for nand. It even makes sense in the 8TF narrative where maybe they feel like they could deal with one of them but both is too much and doubling nand would just be a contract negotiation. I still believe that GDC is more or less a hard stop for an announcement of some kind. They could announce a price at a later date (but its getting late for that too).
 
Play testing will involve both gameplay testing and technical testing. Generally speaking, play testers will be running through gameplay and game bugs, but technical tests reliant on the hardware performance will mostly be operated by the developers during development at this point. Devkits are likely rare and in the hands of developers. Final testing on next-gen retail hardware probably won't happen until nearer title launch, and it doesn't need to either. If this supposed tester is working on a launch title, I don't expect testers to get access to PS5 platforms to play the game until a few months before release, so not until late Summer. If the game isn't releasing until later, I doubt play testers will be playing on PS5/XBSX hardware before the consumer device is available.

That doesn't discount a play-tester from knowing specs - they could be friends with a dev who told them. ;) It's just unlikely that they have hands-on with devkits.
 
That said, there is one thing in it that is interesting in that we haven't discussed it at all, the 500GB nand vs 1TB nand for Xsx. I could certainly see where Sony would be having doubts at being at that much of a storage deficit even if they believe it puts them at a lower price point. Could there be a 2013 like last minute doubling of nand? It would make sense in why they are delaying an announcement while trying to get a new deal done for nand. It even makes sense in the 8TF narrative where maybe they feel like they could deal with one of them but both is too much and doubling nand would just be a contract negotiation. I still believe that GDC is more or less a hard stop for an announcement of some kind. They could announce a price at a later date (but its getting late for that too).

what?
 
If one storage is faster, then it wouldn't be surprising for it to have less capacity. Otherwise it'd cost more and this rumour is placing everything else at XBSX's level, the whole machine would cost XBSX's level, which clashes with the stated goal of PS5 will be the fastest selling PS ever.
 
I think at one point in time these rumours were perfectly fine for me. Until MS started showing die size and we started to land closer to 400mm^2+. And then as a group when we started to figure out the potential costs or BOM. I think we can naturally agree the costs are going to be high. As in; we know it won’t be $399. And we are looking at likely a $499 price point or possibly higher.

7nm waffers IIRC are at less than $10 000 (or were during late 2019). We also know the defect density of the node was pretty low on 7nm large dies, with estimated yields above 91%.

I think someone made the calculations for a 350-400mm^2 SoC and it would be around the $130-150 mark. It's really not that much more expensive than the 2013 SoCs.
This is without going into the fact that console makers guarantee fabs contracts that span for multiple years with tens of millions of units so they should get better wafer prices than pretty much anyone else.

I think there's too much assumption of console BoM + assembly based on discrete PC GPU prices, which honestly is just a wrong parallelism.
Consumer dGPU vendors will always charge as high as they possibly can for their hardware, because their main (and only?) source of revenue is hardware sales. AMD didn't price the 5700XT only based on how much the cards cost to make, they priced it as high as they possibly could to maximize margins considering the competition and value perception.
Console makers OTOH tend to price their hardware as low as they possibly can, because they get their revenue mostly from software sales. They're highly dependent on userbase size so that more consoles in gamers' homes = more software revenue.



And we are looking at likely a $499 price point or possibly higher.

seems to be in direct conflict with Sony’s aim of wanting to transition their PlayStation population as quick as possible. Hard to transition people of the price is high.
(...)

there’s a lot hints that suggest XSX will be pricy. And if PS5 is logically as powerful in every way, I can see transitioning quickly as being impossible.

Why exactly are the "quick transition" statements being translated as "the console will cost $399"?
That seems like quite the big leap in logic to me.

I honestly read it as BC allowing people to quickly transition their retail + digital library to the new console (as has been officially confirmed), plus of course the console not costing a fortune.

Besides, why are people assuming there's this large proportion of PS4 users who somehow already set a hard price threshold of $400 for the PS5? Or that the PS5 will be slower to adopt at all if it comes at $500?
Where's the data to support this?

For all I know, even if it came for $600 the PS5 could be supply constrained for their first 20-30 million units, after which Sony/AMD would make the transition to a newer node that drove the long-term BoM costs down.

We're not suffering from the 2008 crash anymore, like we were in 2013. Worldwide inflation happened in the meantime. Honestly, it seems to me that counting on a $400 price tag for the PS5 - and assume its possible performance/features from it - is a bit of a long shot at this point.


That doesn't discount a play-tester from knowing specs - they could be friends with a dev who told them. ;) It's just unlikely that they have hands-on with devkits.
This doesn't make much sense to me.
If there are no production model consoles laying around during the development of games that come out on release window, where are play-testers supposed to test said games if not on devkits?



Selection bias mate, just let it go:D.
I know. It's just that sometimes the "arguments" used to shoot down (admittedly baseless) rumors are so poor that they might as well just write I hate this rumor because it challenges my beliefs and hurts my feelings.
 
Last edited by a moderator:
Not that I believe that leak but any QA tester that is a formal full-time employee of a studio that has dev kits would very likely know the specs. Even if there is no need for them to know they are probably working with tools on a daily basis that would be able to find/show that info which they are bound to come across. Which like Shifty Geezer said would be the only types of testers that are working with the dev kits at this stage.
 
7nm waffers IIRC are at less than $10 000 (or were during late 2019). We also know the defect density of the node was pretty low on 7nm large dies, with estimated yields above 91%.

I think someone made the calculations for a 350-400mm^2 SoC and it would be around the $130-150 mark. It's really not that much more expensive than the 2013 SoCs.
I thought the estimates were closer to $100-110 at launch, so that’s a 20-50% jump in cost there if any of the estimates are even close at all. :confused:
 
7nm waffers IIRC are at less than $10 000 (or were during late 2019). We also know the defect density of the node was pretty low on 7nm large dies, with estimated yields above 91%.
On a 12" wafer, a 400mm^2 is approximately 136 Dies.
So with a yield of 91%, then it's about 123 dies.
10,000 / 123 = $81 for silicon costs. I'm not sure how to factor in the other aspects here though.

10,000 sounds pretty cheap though, not sure how much a 12" wafer is


Why exactly are the "quick transition" statements being translated as "the console will cost $399"?
That seems like quite the big leap in logic to me.
I think quick means cheaper, the cheaper the price point the more purchases that will be made.

I think $399 represented an optimal solid price/performance value point. And I think that's why we're so focused on that number in particular, but also why we saw 4Pro also launch at that price as well.
I also agree that $399 just seems very unlikely at this point in time given the specs we know so far.
 
I think someone made the calculations for a 350-400mm^2 SoC and it would be around the $130-150 mark. It's really not that much more expensive than the 2013 SoCs.
IIRC last gen BOMs were $150 for SOC plus RAM. Yeah, tear-down was $180 for SOC and RAM for PS4, and we know RAM wasn't cheap.

This doesn't make much sense to me.
If there are no production model consoles laying around during the development of games that come out on release window, where are play-testers supposed to test said games if not on devkits?
For multiplatform games, the other platforms. Obviously for a PS5 exclusive, play testing will need to use PS5 devkits if that's the only platform it runs on, but then you're talking someone at a Sony first-party who'll be afforded the luxury of early hardware for testing. SOCs not being mass-produced certainly aren't cheap and won't be in the laps of the lowest-tier of the development process.
 
I thought the estimates were closer to $100-110 at launch, so that’s a 20-50% jump in cost there if any of the estimates are even close at all. :confused:
So it's a $20-50 difference in the SoC. That seems doable to me considering:
- People are probably willing / able to pay more for the PS5 than they did for the PS4;
- Sony is coming from a much healthier financial place from the PS4 than they did from the PS3, meaning they can risk subsidizing their hardware;
- Subscription services can help pay for the hardware as well, especially now that many people have Plus and online is behind a paywall.


On a 12" wafer, a 400mm^2 is approximately 136 Dies.
So with a yield of 91%, then it's about 123 dies.
10,000 / 123 = $81 for silicon costs. I'm not sure how to factor in the other aspects here though.
I think the calculations I saw involved packaging and transport, but that still seems lower than what I saw.


IIRC last gen BOMs were $150 for SOC plus RAM. Yeah, tear-down was $180 for SOC and RAM for PS4, and we know RAM wasn't cheap.
Is 16GB GDDR6 in 2020 cheaper or more expensive than 8GB GDDR5 in 2013?
 
I think the calculations I saw involved packaging and transport, but that still seems lower than what I saw.
I think for consideration 28nm was sort of the last time Moore's law of doubling really worked with respect to transistor doubling in the same area but being at the same price.
So I think if we're thinking about $150 for the original xbox ones for instance, 7nm at the same die are should cost more than 28nm at the same die area. And it's larger by ~40mm^2.

So you're probably right the $80 is pretty low, curious to see how much everything else will cost to put it together.
 
Status
Not open for further replies.
Back
Top