Next Generation Hardware Speculation with a Technical Spin [2018]

[iroboto]

So we can't discuss the credibility of the sources of rumors?

That’s the difference between era/gaf and here.

Anonymity is more important than vetting of information. You keep pushing and people will just leave. Over there I guess being a leaker is form of internet power. This isn’t a forum for leaks or rumours breaking or what not. And if it were we shouldn’t be pushing for proof.

Sources should stay protected here for the sake of technical discussion so that people can discuss and learn.

 

[anexanhume]

This is pretty good for 3d stacked memory on high power devices: dissipating through the other side of the board to a heat pipe connecting back to the main heatsink.

Also allows to have the memory at a lower operating point than the SoC with an independent heatsink.

My concern is that this approach wouldn't be sufficient cooling capacity for a full-on console with a TDP north of 100W. That's why I'm struggling to place what it's actually for. I appreciate the hearty push-back from this thread and bemoan my lack of a mechanical background to even do some napkin math to determine what level of dissipation this thing could actually handle.

I think Intel's recent slide on Foveros shows the reason this is the end-goal though. Energy per bit is definitely way down in a true 3D stacked approach, which is going to help lower that overall power anyway (I/O on the active die is dissipating less to run its own I/O, so it is a bit of a positive feedback loop). AMD has talked about active interposers in white-papers (to the end of providing the optimal amount of interconnect paths), but I doubt it would be ready for 2020 consoles. To that end, I think Intel is leapfrogging AMD a bit with their IO die approach.

 

[Globalisateur]

My concern is that this approach wouldn't be sufficient cooling capacity for a full-on console with a TDP north of 100W. That's why I'm struggling to place what it's actually for. I appreciate the hearty push-back from this thread and bemoan my lack of a mechanical background to even do some napkin math to determine what level of dissipation this thing could actually handle.

What if it's used to cool down the APU of a console from both sides ?

 

[Shifty Geezer]

That’s the difference between era/gaf and here.

Anonymity is more important than vetting of information. You keep pushing and people will just leave. Over there I guess being a leaker is form of internet power. This isn’t a forum for leaks or rumours breaking or what not. And if it were we shouldn’t be pushing for proof.

Sources should stay protected here for the sake of technical discussion so that people can discuss and learn.

In addition, on this forum we discuss ideas. Generally speaking, it doesn't matter where an idea comes from or how realistic it is as long as it leads to meaningful discussion. We only stamp out wild rumour like the MisterXMedia stuff when it's known noise designed for fanboys. The most important thing in such discussions is that the ideas are taken through critical analysis to a natural conclusion, either consensus or a disagreement on validity and a natural acceptance that those of different ideas don't need to be convinced otherwise, because there's nothing at stake over what people think.

B3D does not exist to find the truth and oust the witches; it exists to keep us entertained through the discussions we have about our hobby.

 

[anexanhume]

What if it's used to cool down the APU of a console from both sides ?

That's certainly a possibility and is expressly stated in the patent application. For that, you'd think that the typical top down approach would be preferable, with your highest dissipating die on top. Then, your memory is cooled with this-through the board heatsink. My guess is that this approach is untenable due to the z-height of said stacked memory, or that non-stacked and horizontally arranged memory would be too large of a void in the BGA pattern. Similarly, you could try and do a ledged design where there's actually a "pocket" in the package to house the RAM, but I bet that has manufacturability/reliability concerns. Finally, one can imagine just including an outright cut-out in the PCB to let the memory protrude through, but manufacturing and reliability concerns remain present here.

Maybe the compromised paths (bottom with the through-hole method, top with additional die in the way) is still comparable to cooling with a single-die, topside heatsink approach.

 

[mrcorbo]

I would ask:

Is it worth the potential thread derail to establish the credibility of this particular fact?". In this case, the fact is the name of the Xbox Scarlet dev kit.

If the fact is important enough, is discussion of the credibility of the source best introduced in the way it was?

If there's enough supporting evidence that the source lacks credibility to have a genuine discussion about it wouldn't that be best done in it's own thread? Then, any time that source comes up, you can link to it so it doesn't have to keep coming up in threads going forward?

 

[BRiT]

Who cares if a person is a PR MouthPiece or not? Being a PR MouthPiece is not a bad thing. Cerny is a Sony PR MouthPiece. Should we discard anything he says because of that?

Now do you see why its such a waste attacking sources instead of focusing on the topic?

 

[MrFox]

We should discard PR spin from actual technical info.

 

[BRiT]

Agreed, which still has nothing to do with identifying the source of the information or labelling them as PR MouthPiece. As Shifty stated, discuss the idea not the person.

 

[Shifty Geezer]

Also, this is a lot of noise over a rumoured codename! The codename has no technical value, and one is more likely to get an accurate codename from a PR mouthpiece than someone else. If we had a leak on technical specs, it may be worth arguing over the validity of the source (not really, but at least there'd be some reason to want to know how reliable a technical source versus one leaking a codename).

 

[mrcorbo]

Well, the difference between a PR mouthpiece and an independent journalist is the independent journalist would be more inclined to obtain all of the information as opposed to being content with releasing the information that the company they were covering want to get out in order to support whatever narrative they are trying to create. So, while specific individual facts may be reliable, the conclusions you draw from them may be unreliable due to the omission of other facts.

In this specific case, though, it doesn't seem very relevant.

 

[iroboto]

Well, the difference between a PR mouthpiece and an independent journalist is the independent journalist would be more inclined to obtain all of the information as opposed to being content with releasing the information that the company they were covering want to get out in order to support whatever narrative they are trying to create. So, while specific individual facts may be reliable, the conclusions you draw from them may be unreliable due to the omission of other facts.

In this specific case, though, it doesn't seem very relevant.

Yea. Unfortunate problem with internet journalism as a whole today, this type of behaviour isn’t just specific to a few individuals, the drive for information now verify later is valued more than accurate complete perspective reporting.

Anyway, code name is not a big deal as Shifty writes. Even mulling through some older released documents it’s not like we saw the hardware mentioned by its code names.

 

[Shifty Geezer]

Well, the difference between a PR mouthpiece and an independent journalist is the independent journalist would be more inclined to obtain all of the information...

It's been a long time since journalism wasn't "publish first, ask questions later." 90% of the information out there is rapid-fire quotes of the latest soundbite with zero corroboration. You need specialist, trustworthy sources with a reputation to expect some degree of validation to have happened before publishing, and that's as likely from an affiliated source as a neutral party.

If people are using the term 'PR mouthpiece' to mean 'raving fanboy lunatic posting their delusions', they shouldn't.

 

[Jay]

Just to add, even getting multiple independent collaborating sources doesn't mean a leak will always end up being what is relalised. It's the nature of what a leak is especially the earlier it happens.
Also how it's reported can be heavily influenced by the kind of technical knowledge someone has, and their ability to parse what they've heard.

 

[Deleted member 13524]

This thread is purely for discussing the next-gen main boxes. All other possibilities - multitiered SKUs, VR devices, portables, etc - warrant their own threads to ensure this one doesn't become unwieldy.

My point was to consider something with the power of a base PS4 being used as a controller for the PS5, and in which ways such a powerful client+controller could be used for games.
Say, rear-view for racing games, super-detailed 3D zoomable maps in adventure/FPS games, an inventory with detailed 3D models, etc.

The PS5 gamepad is rumored to have an LCD, so developing for a second screen would be standard for all games. Sony already tested the waters for this with the PS4 (second screen through the android/iOS app already exists in a bunch of games) so it actually makes sense for them to pursue the idea.

But with the PS4 Go being used as "super controller", developers could go even further with local 3D rendering taking place. Or even if they didn't want to put a lot of effort into the "premium 2nd screen" functionality they could just scale up the UI and people would get a gamepad with a 10" screen.

 

[BRiT]

How much would such a super controller cost? How receptive would the base be with paying $X for said super controller? Or is this super controller entirely optional for the PS5? Does it only function with PS5?

 

[anexanhume]

How much would such a super controller cost? How receptive would the base be with paying $X for said super controller? Or is this super controller entirely optional for the PS5? Does it only function with PS5?

The patents show a screen with detachable dual shock sides. I think such a device would definitely be optional, but I also think the plan is to have touchscreen for the standard DS5. They need some sort of touch area for DS4 compatibility, and I think e-ink should be a possibility here - especially since color e-ink is now a thing.

 

[Shifty Geezer]

An optional super controller device with detachable sides would be a mobile PlayStation. Still not relevant for this thread. Discussion of a PS4 Go or somesuch belongs in a different thread. This thread is for determining the hardware inside the main boxes - RAM, CPU, GPU, special sources, etc. plus ports and things. Not sure even controllers belong here. eg. https://forum.beyond3d.com/threads/gamepads-with-led-touchscreens.60809/

In the box, standard, default controllers, possibly. There's not a great deal of hardware to discuss there though and it's more a matter of ergonomics and gimmicks.

Note that a topic like a patent can be linked in multiple threads if it has relevance to them. The patent here would be discussed in terms of its value to a console, whereas in a handheld platform thread, its value to handheld PS.

 

[Ike Turner]

Here's a better "source" :
Deciphering Xbox Scarlett: Anaconda, Anthem, Danta, Lockhart, Maverick, and Roma:

https://www.thurrott.com/xbox/19539...conda-anthem-danta-lockhart-maverick-and-roma

 

[Shifty Geezer]

I give up...