Portable PS5? *spawn*

[Shifty Geezer]

A portable PS5 isn't an option. That gives Sony two choices for a portable -
1) a new portable with no compatibility with any previous machine, or
2) a portable with PS4 compatibility.

Option 1 starts with no games and devs we hope will make games for it. If devs will, they will.
Option 2 has a starting library of thousands and if those devs were willing to make a game for a new design, they'll be happy to make new games for a portable PS4 with the option of releasing even on PS4 and PS5.

Options 1 makes no sense. Option 2 will get new games just as any new handheld would. I mean, if devs won't make new games for PS4Portable, they wouldn't for PSVita2 either. That means two options on the table: a PS4 portable, or no portable and just let Nintendo have that space uncontested.

 

[arandomguy]

In terms of no caveat PS4 backwards compatibility the question I'd still have is surrounding the issue of the memory subsystem and having to do that in both the power and physical space budget for mobile. Node advancement can be applied to the SoC itself but that doesn't apply to the GDDR5 memory system. Dropping in GDDR5 would be quite a significant percentage in terms of the power budget at mobile levels and also physical space/heat. LPDDR5 even at quad channel doesn't match the bandwidth either so it can't be a drop in substitute. Cache might be a too large of a departure to just assume games will run identically without any developer intervention.

 

[Proelite]

In terms of no caveat PS4 backwards compatibility the question I'd still have is surrounding the issue of the memory subsystem and having to do that in both the power and physical space budget for mobile. Node advancement can be applied to the SoC itself but that doesn't apply to the GDDR5 memory system. Dropping in GDDR5 would be quite a significant percentage in terms of the power budget at mobile levels and also physical space/heat. LPDDR5 even at quad channel doesn't match the bandwidth either so it can't be a drop in substitute. Cache might be a too large of a departure to just assume games will run identically without any developer intervention.

Ldppr5x on 256 bit bus is up to 240GB/s. So plenty.

 

[Proelite]

They don't care about people buying 5$-20$ PS4 games. They want people buying the last console + most expensive games. Portable PS4 would be good for us, but it would bring peanuts revenue for them as no sane person would sub for their most expensive service when half of games would be locked as PS5 games.

Very soon very few games will be released on PS4 platform. Next FF7? Spider-Man 2+? Last Horizon DLC ? All PS5 exclusive. etc.

Sony make their lions share of revenue from third party games, especially live service.

 

[AmFreak]

A portable PS5 isn't an option. That gives Sony two choices for a portable -
1) a new portable with no compatibility with any previous machine, or
2) a portable with PS4 compatibility.

Option 1 starts with no games and devs we hope will make games for it. If devs will, they will.
Option 2 has a starting library of thousands and if those devs were willing to make a game for a new design, they'll be happy to make new games for a portable PS4 with the option of releasing even on PS4 and PS5.

Options 1 makes no sense. Option 2 will get new games just as any new handheld would. I mean, if devs won't make new games for PS4Portable, they wouldn't for PSVita2 either. That means two options on the table: a PS4 portable, or no portable and just let Nintendo have that space uncontested.

Option 3 Try to make it another SKU, try to be the Series S of the PS5 in hand held form.
1 and 2 also shouldn't rule each other out, the PS5 was also new tech and at the same time is compatible with the PS4. Make the hand held Zen2 + RDNA2 just like the PS5, only thing that already available hand helds can't replace there is the gpu width, but that should be possible at a smaller node with a custom design that uses Zen2 instead of 4 and RDNA2 instead of 3, maybe they could also reach good compatibility with half the width at double the clock speed. As a last resort they could also make patches just like they have done with some ps4 games on ps5 and Ms did with bc.

 

[arandomguy]

Ldppr5x on 256 bit bus is up to 240GB/s. So plenty.

That in itself is not going to be a trivial implementation given the size and power constraints.

The current handhelds with 128 bit LPDDR5 are already tight fits spacing wise.

You'd also basically still be doubling the memory portion of the power budget relatively speaking.

 

[Proelite]

That in itself is not going to be a trivial implementation given the size and power constraints.

The current handhelds with 128 bit LPDDR5 are already tight fits spacing wise.

You'd also basically still be doubling the memory portion of the power budget relatively speaking.

Hard to believe space constraints since the steam deck is almost as big as my series s in terms of footprint.

 

[arandomguy]

Hard to believe space constraints since the steam deck is almost as big as my series s in terms of footprint.

The tear down an motherboard pictures are there? You can see the board spaces does not have room at all for another 4 memory chips and traces unless you make it significantly larger.

XBS and handheld sizing isn't entirely comparable as Z height is more of a factor handhelds, and they also have to fit in a battery, output (not just display) and input.

 

[Proelite]

XBS and handheld sizing isn't entirely comparable as Z height is more of a factor handhelds, and they also have to fit in a battery, output (not just display) and input.

In that case, what about a steam deck-sized handheld, but the screen is foldable? Something similar to the GDP Win Max.



In seriousness, is having the other half of the chips on the other side of the mobo not an option?

 

[Nesh]

A portable PS5 isn't an option. That gives Sony two choices for a portable -
1) a new portable with no compatibility with any previous machine, or
2) a portable with PS4 compatibility.

Option 1 starts with no games and devs we hope will make games for it. If devs will, they will.
Option 2 has a starting library of thousands and if those devs were willing to make a game for a new design, they'll be happy to make new games for a portable PS4 with the option of releasing even on PS4 and PS5.

Options 1 makes no sense. Option 2 will get new games just as any new handheld would. I mean, if devs won't make new games for PS4Portable, they wouldn't for PSVita2 either. That means two options on the table: a PS4 portable, or no portable and just let Nintendo have that space uncontested.

Yeah no option sounds viable. It is extremely difficult to penetrate the market with a pure portable device now, with pretty much every tablet and phone being a portable gaming device if someone chooses

The Switch is a console/portable hybrid that plays its own kind of modern games either on TV or on a big portable screen, with controllers and proper local multiplayer. It penetrates both markets by being both. Someone can ask "what about Switch Lite". Switch Lite and Switch though are the exact same products. They play the same games and Lite exists because Switch exists.

Unless Sony can release a PS Portable that plays all of the PS5 games, the chances of releasing a successful product are slim.

 

[Tkumpathenurple]

A new portable that plays all of your PS4 games is pretty good marketing in and of itself. But, assuming PS4 BC either way, would it be better as only a "PS4 Portable" or a standalone portable with PS4 BC?

The latter would afford Sony the chance to phase out the HDD paradigm while keeping an affordable entry point (particularly if it's a microconsole.) It would be nice if they could merge the save file system too, so jumping between devices is more seamless than PS4-PS5.

The tear down an motherboard pictures are there? You can see the board spaces does not have room at all for another 4 memory chips and traces unless you make it significantly larger.

XBS and handheld sizing isn't entirely comparable as Z height is more of a factor handhelds, and they also have to fit in a battery, output (not just display) and input.

That prompted me to look at the Steamdeck's teardown. I'm astonished by how small the motherboard actually is.

A good deal of the circuitry would be done away with in a microconsole, though obviously not all of it.

Rather than try to account for each scenario with multi-purpose hardware, place the options in the hands of the player: attach a battery if you want to leave it in your bag and stream to your phone; slot it into a battery powered screen if you want to place it on a surface and play some local multiplayer; slot it into a dock if you want full connectivity at your TV; just plug it in to a TV (one USB-C port for power, and one for display) if you've just brought it around to your friends.

 

[RDGoodla]

In terms of no caveat PS4 backwards compatibility the question I'd still have is surrounding the issue of the memory subsystem and having to do that in both the power and physical space budget for mobile. Node advancement can be applied to the SoC itself but that doesn't apply to the GDDR5 memory system. Dropping in GDDR5 would be quite a significant percentage in terms of the power budget at mobile levels and also physical space/heat. LPDDR5 even at quad channel doesn't match the bandwidth either so it can't be a drop in substitute. Cache might be a too large of a departure to just assume games will run identically without any developer intervention.

LPDDR5T developed by Hynix already has 10.5 Gbps, which is 168 GB/s in 128 bits.

Is it enough bandwidth for PS4 backwards compatibility?

 

[Nesh]

A new portable that plays all of your PS4 games is pretty good marketing in and of itself. But, assuming PS4 BC either way, would it be better as only a "PS4 Portable" or a standalone portable with PS4 BC?

The latter would afford Sony the chance to phase out the HDD paradigm while keeping an affordable entry point (particularly if it's a microconsole.) It would be nice if they could merge the save file system too, so jumping between devices is more seamless than PS4-PS5.



That prompted me to look at the Steamdeck's teardown. I'm astonished by how small the motherboard actually is.

A good deal of the circuitry would be done away with in a microconsole, though obviously not all of it.

Rather than try to account for each scenario with multi-purpose hardware, place the options in the hands of the player: attach a battery if you want to leave it in your bag and stream to your phone; slot it into a battery powered screen if you want to place it on a surface and play some local multiplayer; slot it into a dock if you want full connectivity at your TV; just plug it in to a TV (one USB-C port for power, and one for display) if you've just brought it around to your friends.

Prob with just a PS4 portable device is that people will be left only with old games. Unless new games that people want to play keep being released on it. Essentially making the PS4 portable a Series S version for the PS5.

 

[Tkumpathenurple]

LPDDR5T developed by Hynix already has 10.5 Gbps, which is 168 GB/s in 128 bits.

Is it enough bandwidth for PS4 backwards compatibility?

Not quite - the PS4 has 176GB/s bandwidth. That's tantalisingly close though.

Prob with just a PS4 portable device is that people will be left only with old games. Unless new games that people want to play keep being released on it. Essentially making the PS4 portable a Series S version for the PS5.

I think that'd be fine. It's the mandatory support and feature parity that makes life tricky for some devs IMO.

 

[arandomguy]

That prompted me to look at the Steamdeck's teardown. I'm astonished by how small the motherboard actually is.

A good deal of the circuitry would be done away with in a microconsole, though obviously not all of it.

Rather than try to account for each scenario with multi-purpose hardware, place the options in the hands of the player: attach a battery if you want to leave it in your bag and stream to your phone; slot it into a battery powered screen if you want to place it on a surface and play some local multiplayer; slot it into a dock if you want full connectivity at your TV; just plug it in to a TV (one USB-C port for power, and one for display) if you've just brought it around to your friends.

Yes but modularity in itself has both a size cost and monetary cost component. The hypothetical actual handheld loadout of that with everything attached would likely be significantly bigger or with other compromises (eg. much smaller battery, awkward ergonomics) than if the entire device were one piece.

LPDDR5T developed by Hynix already has 10.5 Gbps, which is 168 GB/s in 128 bits.

Is it enough bandwidth for PS4 backwards compatibility?

Slightly less as mentioned.

But another issue is I'd wonder what the real world limitations on it are in the context of doing over 128 bits over a gaming SoC. Remember just because the memory is capable of those speeds and the if the memory interface is capable of those speeds it doesn't say anything about what happens in between. Signaling and achievable speeds over 64 bits and/or distance (stacked) over the SoC might be more lenient.

It's an extension I believe of LPPDDR5 and not just faster LPDDR5, and so may have different deployment/implementation considerations.

In seriousness, is having the other half of the chips on the other side of the mobo not an option?

Do you mean immediately underneath other memory chips? Now I won't say I'm an expert on this to that level but while I think it could in theory be possible but to maintain the signal integrity required would probably not be simple (to put it lightly) from either a design or cost perspective.

 

[Karamazov]

Please Sony, support your PSVR2 a little better before thinking of releasing a PSVITA 2.

 

[Globalisateur]

They'll never release a new device in 2025+ using decades old technology in order to please a few fans and be 100% BC with PS4. Not happening. If they release a portable Cerny will use the latest available tech (logically with hardware RT units) as he did with PS Vita and they'll prefer make remakes on that thing to be sold full price.

They main reason they made PS4 BC on PS5 is to help transition people to PS5 ecosystem (in order to eventually sell them 70/80€ PC/PS4 ports and a few new games).

 

[Tkumpathenurple]

LPDDR6 news doesn't seem to be all that prevalent, but a rather old article suggests LPDDR6 may go as high as 17,000MT/s vs LPDDR5's 6,400.

That'd be overkill for, essentially, a PS4 emulating portable, but slower, cheaper variants would be a good fit. 11,000MT/s on a 128 bit bus would be an exact match for PS4's bandwidth if I'm not mistaken.

Yes but modularity in itself has both a size cost and monetary cost component. The hypothetical actual handheld loadout of that with everything attached would likely be significantly bigger or with other compromises (eg. much smaller battery, awkward ergonomics) than if the entire device were one piece.

True. It's certainly no silver bullet. But you could make a similar argument for the Switch - if you only want to play it as a home console, it's got a bunch of extra bulk attached at all times.

I'm biased though, as I've wanted Sony to do this ever since the PSVita TV. Which they later rebranded to PSTV before completely dropping it

Please Sony, support your PSVR2 a little better before thinking of releasing a PSVITA 2.

I agree that they need to. But a microconsole could be the perfect device to facilitate this by attaching to the headset like the Apple Vision Pro. An 18CU GPU at 1GHz would match the power of the Quest 3, and Sony would have kinda untethered VR.

Although it gets really messy with developer support and, probably more importantly, trying to convey that to the public.

They'll never release a new device in 2025+ using decades old technology in order to please a few fans and be 100% BC with PS4. Not happening. If they release a portable Cerny will use the latest available tech (logically with hardware RT units) as he did with PS Vita and they'll prefer make remakes on that thing to be sold full price.

They main reason they made PS4 BC on PS5 is to help transition people to PS5 ecosystem (in order to eventually sell them 70/80€ PC/PS4 ports and a few new games).

I agree that it won't be Jaguar and whatever the PS4's GPU architecture was called. But I wouldn't expect the most cutting edge tech in there either. PS4 spec in PS5 architecture is more likely IMO.

 

[RDGoodla]

PS4 Portable is basically doable using Z1 extreme with larger GPU, and LPDDR5T.

Can Sony just release a pure PS4 Portable in 2024 which is just for PS4 games?

 

[Proelite]

Can Sony just release a pure PS4 Portable in 2024 which is just for PS4 games?

They can, but it'll be better to do one that has PS4 BC on a newer architecture.