Support NeoGAF

[Audiophile]

You know, the Sony building in Ginza?!?!

Typo'd it as "Song" instead of "Sony" in previous post..

A console that’s powerful enough to emulate the Cell and plays PS3 games.

Been thinking about this and given that they don't seem to be inclined to do good software emulation for PS1/PS2 nor any sort of emulation for PS3; I was wondering why they don't just bite the bullet sooner rather than later, circumvent the entire issue and port the designs for the PS1 CPU/GPU+, PS2 CPU/GPU+ & PS3 CPU on to a single little APU chip.

They may have some issues in terms of licensing with IBM/Toshiba on the PS3 but I doubt it'd an insurmountable issue. They could probably skip the Nvidia RSX and simply emulate that on a really basic AMD/ARM GPU element; as well as bespoke functions of older platforms like sound chips etc. on a basic AMD/ARM CPU element.

They could call it the "PS123" chip and sell it on a small, semi-budget legacy console with a CD/DVD/BD drive as well as access to the store for as many titles as they can provide license-permitting; calling it the "PlayStation Legacy", eventually they could shrink the chip enough and put it in a portable device (Portal 2?) alongside a shrunk PSP/Vita chip set; and even include it as a chiplet next to the main die/s on a latter premium/mainline platform such as a PS6 Pro -- as a marketable feature. The "Legacy" console and its disc drive should be multi-region with 50/100 & 60/120Hz output as well as various scaling/display modes; and digital licenses should offer the option to download any regions' release.

PS4/PS5 and forward will likely be covered by hardware BC for the next few (or remaining) generations.

Once it's done, it's done and while the initial work may be difficult (primarily shrinking the ancient designs, having adaptive modes for memory speeds/latencies/CPU-GPU clocks, matching CPU timings and identifying what hardware components need to be directly recreated on die and what can be flawlessly emulated on a standard CPU/GPU element); it just seems -- relatively speaking -- like the most straightforward approach to the issue, getting in there at a base level and solving all the issues in one go. Sony will also be able to finally dump all of those life support OG PS3s and old Cell blade servers they have racked up for streaming; and implement these likely miniscule chips into their streaming servers alongside whatever they're using or will be using for PS4/5/6 etc.

 

[thicc_girls_are_teh_best]

Not if you like Sony or Nintendo games.

Well, I think that one should come with an asterisks.

That's the thing. Sony games as of late don't really do much for me. Ghosts of Tsushima and PSVR2 related items have been about it. Other games are appearing on PC and with the backlog I have, I don't need to do day one purchases. Nintendo, same deal. I'm not big on most of their first party games anymore. I haven't even finished or barely played BOTW and haven't bought TOTK. Mario stuff I might play when the whim hits me, but a lot of the games I own on the switch are third party. Xbox....Gamepass and the cheap deal they had last month is the only reason I have an X.

Any talk of next gen though doesn't excite me. The games aren't going to look much better, 60fps and the stuff people rant and rave about I don't care about either. It's all the same to me. The games certainly aren't going to get cheaper. If the big push keeps being GAAS and other stuff like Fortnite consider me a complete loss at that point. It could be age but it all feels to be drifting away from what interests me.

Or maybe I just need to step back for a while. Come back to it when its fresh.

The bolded is something I've been trying to get at WRT why even the current porting strategy from Sony/SIE for 1P traditional titles still devalues the console to some degree. If they set a cadence of "Okay, all of our 1P traditional titles, you can expect them on PC within two years", then I think that betrays what the original idea behind the ports was to do. Which, at least from when they started, was to give gamers on PC a taste or sampling of a couple 1P titles, in swaying them to pick up a PlayStation to access the rest and maybe do more of their gaming & spending in the ecosystem. Not to outright replace PC, but to position PlayStation as a complementary to their PC.

The way the ports have been increasing in frequency though, it'd seem "PlayStation" in this context is becoming more and more just relating to the 1P software, and not the console platform. I feel that's ultimately going to become a big mistake. The type of people Sony could convince among PC to pick up a PlayStation as a complementary, would likely be hardcore/core enthusiasts anyhow, who would generate a good amount of revenue in the console ecosystem and provide bigger profit margins for Sony/SIE naturally. But if you're giving these PC gamers, who already had enough patience to hold out altogether or for long periods previously in picking up a PlayStation, your 1P games with a window timeframe that for them is very tolerable...they have very little reason to consider picking up the console.

And I don't really buy this idea that PC and console gamers are completely different audiences; even between consoles, I was watching a Puertorock video the other day going over the Circana/NPD data and they pointed out how Hogwarts Legacy was in the Top 10 (maybe Top 5) for Switch software sales last year, despite the late port and being 3P. And their point was that maybe the only reason Nintendo systems have the perception of where only Nintendo games sell there, is because 3P haven't prioritized them enough. That if they did, Switch charts would probably look a good bit more like PlayStation's, when it comes to the distribution of 1P to 3P games in those Top 10s and Top 20s. I think there's validity to that claim because if you go back to the NES/Famicom and SNES/SFC, you can see that actually being the case more often than not, in all the global markets.

I bring that up to say, we know a lot of modern PC gamers were at some point console-only or console-dominant players from either the PS4/XBO gen, or PS3/360 generation. So at some point, they were mainly into consoles. Then something (or some things) happened to push them to PC. But if there was little in the way of audience crossover, why does virtually every 3P game also release on PC? That whole idea of "heavily different audiences" maybe held true in the '90s and early '00s, but IMO it started to end once the 360 released and we saw MANY once PC-centric or PC-exclusive developers begin prioritizing consoles (mainly the 360) for sales. This was back when Steam was still in its infancy, so the PC gaming market was kind of in a rut. Many of those PC devs couldn't rely on PC gaming customers alone to keep them financially afloat, and MS had really helped pave a path for them to get into the console gaming space.

Now, I feel a lot of THOSE PC gamers, eventually went back to PC and are comfortably there with Steam. They are likely the holdouts with little to no interest in console that I see some people refer to, when they say there's little audience crossover. However like I just mentioned, it doesn't account for the PC gamers who have a console-dominant or console-exclusive background, who'd be at least 50% of the addressable PC gaming audience (of expats from console to PC). These are the ones that could be encouraged to buy a PlayStation, but Sony's current porting strategy, IMO, works against achieving that. And, it might convince more hardcore/core enthusiasts with a current system to switch to PC going forward, which can at the very least hurt adoption rates of PS5 Pro and PS6.

 

[Hudo]

Biggest innovation will the the normalization of $80 MSRP.

 

[thicc_girls_are_teh_best]

Had some spare time earlier and got to thinking about chiplets, got a bit carried away with some mockups/concepts for the entire generation. Kind of out there with stacked HBM and all kinds of craziness..

Also, I know there's a strong aversion to the concept of a low-powered, budget system at launch due to the XSS, but I think that's just a matter of poor execution (~40% bandwidth rather than 65-75%, low absolute memory, 1/3 gpu power rather than 1/2-to-2/3 etc.) so I baked the idea of a well-executed lower power system into this, along with "midgen" consoles and a "twilight" console for the transition to the gen afterwards:

Really good work on these graphs, they have a professionalism to them. Also I like a lot of the specs and feel most of it would be doable for a system itself going for $499 (tho, I'm still on the team of wanting some entry-level VR headset in the box as well, so the actual price would be like $100 more but I think that's fine). I'm mainly interested in the PS6 here but the idea of a Life could also work. Some of my earlier speculation included something like a Lite, but I eventually dropped it. That said I'll go over some of the things that stood out to me:

CPU: I'm a bit interested in why you went with an 8C/16T CPU setup. Do you feel that custom silicon for things like RT & AI would be sufficient enough in offloading draw calls from the CPU for those types of tasks? That they could be fleshed out enough to offload calculations for things like physics which may only be partially handled by the CPU itself going forward?

GPU: If I did my calculations correct, for the PS6 you'd give it ~ 60 TF of FP32 compute, 39.6 Gpolys/sec polygon generation (assuming 12 primitives per cycle, I think current RDNA 3 supports 12 while RDNA2 supports 8), and 646.8 Gpixels/sec. That seems it could be realistic, though I could also see Sony scaling it back to just two identical GPU chiplets with 18 WGP each, to save on wafer costs per APU, still have a scalable GPU via chiplet, and increase clocks to get some performance boost or lean into dual-issue shader features to provide some compute boosts when needed.

The RAM for the GPU is probably where it'd get more complicated. 36 GB HBM3P is a lot and I dunno how that could work out for SIE in BOM prices even at large volume, without eating into profit margins on the hardware. I think Sony are generally becoming less adverse to taking big losses on the hardware right out of the gate, but specs with a 54 WGP GPU design, whatever amount of cache that would require AND 36 GB of HBM3-based memory in particular for $499 on 2nm would probably end up eating well more than $50 in the red per unit.

That's why I think if Sony did go with HBM3 (or HBM3P, if that is a specific variant), they'd stick to 32 GB, and probably wouldn't use DDR. Instead they'd probably use something like their own ReRAM to complement the HBM3 IF they could get sufficient speed at smaller capacities (basically, finding some way to decouple most of the potential bandwidth from the per-module bandwidth of each chip). The idea being that in-house ReRAM would be cheaper than any equivalent DDR and still give them the features of DDR when it comes to byte addressability (a must) and at least close enough latency, but can scale much higher with capacity while allowing non-linear scaling of speed/bandwidth.

From what I understand, currently it's not set up like that with ReRAM but then again we haven't heard anything on it since 2019. Barring that possibility, 4 GB of say LPDDR5 as you have it would be a good compromise and probably enough for background OS tasks so most of the 32 GB HMB3P could be freed up for games. Say they reserve 2 GB of HBM3P for the OS, the 4 GB LPDDR5 also for OS, and some partition of the SSD for OS and general applications (100 GB). I think that would be enough for both the PS6 and PS6 Lite. But in this case the PS6 Lite would have to get a RAM change too: instead of 24 GB HBM3P it'd be cut to 16 GB HMB3P, but still keep the 4 GB LPDDR5 and 100 GB SSD OS reserve for virtual memory-like measured quick data access.

So I guess that'd lead to..

RAM: This is one area I'd make changes. For PS6: 32 GB HBM3P & 4 GB LPDDR5. For PS6 Lite: 16 GB HBM3P & 4 GB LPDDR5

STORAGE: Looks good. I do think Sony would (or should) probably consider a more standard m.2 internal drive so they can halve the amount of space on the PCB (and in the system itself) for storage space (NOT capacity!).

So to do that, they have the single m.2 slot already occupied with a 2 TB SSD fitting the specs you gave, but maybe Sony customize theirs to support 12, or maybe 16 or more channels. There are NVMe controllers now that can support 12+ channels, so that'd give Sony (and others, of course) the flexibility to design an internal SSD with many parallel channels while still supporting the m.2 form factor standard.

Would also make repairs a lot easier I'd assume. However, if users can just swap that SSD out for a bigger one (whether or not it has as many channels built into it), Sony'd have to keep a separate NAND chip obscured somewhere on the PCB storing the OS partition and other critical, security-sensitive files. The controller should be able to read from it and the m.2 drive concurrently, but maybe the NAND chip is slower since it's a single chip, and is just storing specific files, encrypted data, keys etc.

That helps with the user being able to simply take out the internal m.2 SSD and upgrade to a larger one, though I guess this also means if they want to transfer data from the internal storage to a new drive, they'd have to temporarily transfer it to a USB-based external SSD drive first. And I guess for security purposes, that data would have to be stored already in a way that only the PS6 can decrypt the files, and it'd only do so when they're installed on an internal drive.

CONNECTIVITY: Looks good.

EXTERNAL BLU-RAY DRIVE ATTACHMENT: Also looks good, but I think Sony'd probably charge something like $70 - $100 for it, not $50. Especially one that can also read DVDs and CDs (IIRC, Blu-Ray drives need different lasers to read DVDs and CDs, and building those in would increase costs).

What I'd really like is an add-on drive that can also work with classic PS3/2/1 peripherals such as controllers and memory cards. If they could price the base one with PS4/5/6 physical game support for say $69, then one supporting PS1/2/3 discs and peripherals as well for $99 would be very reasonable.

VR/AR: Well obviously, more of this. But I think I'd differ in feeling VR/AR next gen is only fully worth it if Sony can scale down the design, to where they can have a cheap entry model included with the PS6, and that an entry model on its own doesn't go for more than $149 MSRP. This is why I think the developments with the PS Portal are interesting, because that tech could help with enabling a PS6 to handle all the processing brunt for VR while the headset has it streamed.

Which is in a lot of ways how it's already done, but I'm talking even more in this case. And especially in terms of facilitating wireless play. It's a reason why I've been less focused on specs but where if I were to focus a lot on technical specs of the console itself for something that could work, a lot of what you've mentioned I feel could be doable while including a cheap VR headset in the box. My suggested changes were with this approach in mind, and also some with the controller itself (which I think would be a bit more expensive than the current Dualsense, still way cheaper than Dualsense Edge, and facilitate both traditional and VR gaming in one package).

I guess to reiterate how I'd like to see it:

$499 PS6 SKU: PS6 system, 512 GB internal storage (upgradable) (maybe slower than 20 GB/s out of the box but still > 12 GB/s), no disc drive, Entry VR headset (1440p per eye, 120 Hz refresh, headphone jack), Dualsense VRT (VR & Traditional) controller

$599 PS6 SKU: PS6 system, 2 TB internal storage (upgradable), no disc drive, Entry VR headset, Dualsense VRT controller

Meanwhile I forgot to talk about PS6 Lite: the way I pictured it, the PS6 Lite would basically be able to play PS5 & PS5 Pro games natively, SOME PS6 games natively, and all PS6 games via streaming (cloud, Remote Play etc.). I could see Sony positioning it as a PS5 refresh of sorts and it could potentially be a "Super Slim" right off the bat. Unlike the PS6 proper, there'd be an SKU with and without an Entry VR headset included, and different storage configurations.

However architecturally speaking, it'd be in line with the PS6 itself.

$249 PS6 LITE SKU: PS6 Lite system, 256 GB internal storage (upgradable) (12 GB/s bandwidth), no disc drive, no Entry VR headset, Dualsense v2 controller

$349 PS6 LITE SKU: PS6 Lite system, 512 GB internal storage (upgradable), no disc drive, Entry VR headset, Dualsense VRT controller.

I think this type of PS6 would need some other things too, like dual wifi modules (one for internet, one for wireless VR), potentially a Thunderbolt port (or replace one of the USB 4.0 ports with a Thunderbolt 5 or 6 port) mainly for extended higher-fidelity wireless VR of multiple headset users (I think this is also something a PS6 would want or need to prioritize, to help make VR a more group-friendly and social experience).

Hence why when it comes to specs, I can see things being roughly in the realm of 45 - 50 TF when talking things like general compute, and that would be perfectly fine. I feel we're getting to a point where the hardware is "good enough" for photorealism; it's the budgets, team sizes and dev times that are making games take so long. Look at what games like TLOU2 and RDR2 were able to achieve on a "measly" 1.84 TF PS4 or 4.2 TF PS4 Pro; those games are very close to stylized photorealism as-is.

With things like advanced upscaler tech, maybe an AI-powered LOD scaler chip like someone else mentioned ITT, plus further adaptability of features like mesh shading, and I think the need for big raw TF jumps becomes redundant. That helps save on shaders, which helps save on chip sizes, which means more chips per wafer (or smaller wafers), ultimately meaning less costs of production.

Typo'd it as "Song" instead of "Sony" in previous post..



Been thinking about this and given that they don't seem to be inclined to do good software emulation for PS1/PS2 nor any sort of emulation for PS3; I was wondering why they don't just bite the bullet sooner rather than later, circumvent the entire issue and port the designs for the PS1 CPU/GPU+, PS2 CPU/GPU+ & PS3 CPU on to a single little APU chip.

They may have some issues in terms of licensing with IBM/Toshiba on the PS3 but I doubt it'd an insurmountable issue. They could probably skip the Nvidia RSX and simply emulate that on a really basic AMD/ARM GPU element; as well as bespoke functions of older platforms like sound chips etc. on a basic AMD/ARM CPU element.

They could call it the "PS123" chip and sell it on a small, semi-budget legacy console with a CD/DVD/BD drive as well as access to the store for as many titles as they can provide license-permitting; calling it the "PlayStation Legacy", eventually they could shrink the chip enough and put it in a portable device (Portal 2?) alongside a shrunk PSP/Vita chip set; and even include it as a chiplet next to the main die/s on a latter premium/mainline platform such as a PS6 Pro -- as a marketable feature. The "Legacy" console and its disc drive should be multi-region with 50/100 & 60/120Hz output as well as various scaling/display modes; and digital licenses should offer the option to download any regions' release.

Could definitely see myself picking up a "PlayStation Retro" with PS1/2/3 hardware BC and disc/peripheral support with some access to the PS Store for $99. There'd be a decent market for it, and you could (or should be able to) use it with a PS6 over USB if one so chooses.

PS4/PS5 and forward will likely be covered by hardware BC for the next few (or remaining) generations.

Once it's done, it's done and while the initial work may be difficult (primarily shrinking the ancient designs, having adaptive modes for memory speeds/latencies/CPU-GPU clocks, matching CPU timings and identifying what hardware components need to be directly recreated on die and what can be flawlessly emulated on a standard CPU/GPU element); it just seems -- relatively speaking -- like the most straightforward approach to the issue, getting in there at a base level and solving all the issues in one go. Sony will also be able to finally dump all of those life support OG PS3s and old Cell blade servers they have racked up for streaming; and implement these likely miniscule chips into their streaming servers alongside whatever they're using or will be using for PS4/5/6 etc.

That'd be a solid plan.

 

[SHA]

I can't put it in words, there's something about games and it's not tied by generations, when you look at fixed areas without any moving objects with fixed pov, some games look dead with no signs of life like air, humidity, fog, lights,etc... some games just made it right even back in 6th gen, some of them don't even do it to this day, these games are literally dead and brings boredom instead of feeling lost in the game, yes, this happens in games by the way and brings chill, we should stop worrying about chasing that trend and game makers shall make it the norm, it's shameful when we stop talking about flops and hertz about latest technology yet still seeing contents still dealing with the same issues as if nothing happened literally, immersion should be the norm, otherwise, these contents are nothing but a cash grab.

 

[thicc_girls_are_teh_best]

I can't put it in words, there's something about games and it's not tied by generations, when you look at fixed areas without any moving objects with fixed pov, some games look dead with no signs of life like air, humidity, fog, lights,etc... some games just made it right even back in 6th gen, some of them don't even do it to this day, these games are literally dead and brings boredom instead of feeling lost in the game, yes, this happens in games by the way and brings chill, we should stop worrying about chasing that trend and game makers shall make it the norm, it's shameful when we stop talking about flops and hertz about latest technology yet still seeing contents still dealing with the same issues as if nothing happened literally, immersion should be the norm, otherwise, these contents are nothing but a cash grab.

Think I know what you're talking about. It's atmospheric ambiance, and yes a lot of games don't do it well. You can notice that when you stand still, as you said. Having light wind blow foliage around, hearing faint sounds of various insects, animals, & people within the atmosphere, seeing moist residue drip off objects after a rainstorm, or character clothes dampened wet after emerging from water and gradually drying up.

All of those things help sell the atmosphere of the game, and many would probably prefer that with not-quite-high resolutions vs. a bunch of static structures present at 4K or 8K resolutions. IMO I think while neither perfectly sell the idea of atmospheric ambiance the way we're talking, you can see the gap in such between a game like GT7 (which does it quite well) and the recent Forza Motorsport (which does it quite poorly overall). The former having an advantage in things like micro details plays into its advantage in such an area, as well as being more accurate to real-world locales and models (and having an overall more unified and inspired UI throughout the game).

 

[Loxus]

So this is the PS5 and it's SSD controller setup.

And this is Mi300 layout.

My question is, what if Sony was to adopt Mi300 layout.

With let's say, 1 AID, 1 CCD and 2 XCD, but incorporates the SSD controller into the AID eliminating the need for the PCIe lanes from the chip to the SSD controller and DDR memory.

Does the SSD controller achieve HBM bandwidth speeds?

 

[thicc_girls_are_teh_best]

So this is the PS5 and it's SSD controller setup.

And this is Mi300 layout.

My question is, what if Sony was to adopt Mi300 layout.

With let's say, 1 AID, 1 CCD and 2 XCD, but incorporates the SSD controller into the AID eliminating the need for the PCIe lanes from the chip to the SSD controller and DDR memory.

Does the SSD controller achieve HBM bandwidth speeds?

Honestly I don't know if the SSD controller could achieve HBM bandwidth speeds, because there's only so many channels it could support and they're limited by the per-channel bandwidth of whatever the PCIe is. There are currently 16-channel SSD Flash memory controllers being made today; there may be higher-channel ones that could be just a bunch of lower-channel ones interfaced through a mux switch or whatever.

Even the lowest HBM bandwidth, for a chip that doesn't have its speed gimped, is 128 GB/s. A 16-channel PCIe 6.0 SSD controller, assuming there's an NVMe spec that could support 16 channels, would be able to provide a drive bandwidth of 128 GB/s, which would match HBM 1.0 bandwidth. The question becomes would it be worth it, and what per-module of NAND speed could NAND Flash makers provide, at what capacities, to make something like that desirable in a console?

Or are you asking if Sony did a setup where the SSD controller is in the AID, would the SSD controller itself reach HBM speeds? Well, it'd probably depend on how the controller is designed, but it's also a question of practicality. Maybe have 128 GB or 256 GB of NAND flash at HBM-like speeds could be worth it but I'm thinking given the likely use-case of repeated fast data access, a more durable non-volatile memory would be suitable, like ReRAM, which Sony have been doing R&D on or at least were back before COVID hit. But ReRAM bandwidth is closer to DDR4 or DDR5, not HBM.

So it's either something like ReRAM or a larger-capacity off-die MRAM that would be most benefited from building the SSD controller into the AID to allow that type of speed. At that point though would it really need to be a SSD controller at all? PCIe latency could end up being the bottleneck, maybe just better to design a custom ReRAM or MRAM memory controller that maybe interfaces through CCIX 3.0 over PCIe lanes, provided capacities for the memory were sufficient enough to warrant it (it'd need to be something like 64 GB, 128 GB or 256 GB altogether).

 

[thicc_girls_are_teh_best]

New PS hardware rumors; posting this here not because of the PS5 Pro stuff, but the "PS Vita 2". If this is happening (and again, that's a giant if), I think it's the right way to go, but also needs clarification.

For one, I 100% wouldn't expect a PS Vita 2 (or PSP3) to be a dedicated handheld with its own unique library. The title might suggest otherwise but even in the video they clarify what it would actually be more like. A new PS portable would basically be to the PlayStation consoles what a Steam Deck is to PC; a companion that can natively play a ton of games already on the consoles, stream others not natively playable, and do all of it in portable format.

It's just not practical to split up 1P dev resources for games specific to a new PS portable, so there's no chance that route would be taken. However, if a dev making a new PS5 or PS6 game, knows their game isn't at all that technically demanding, and can scale down relatively easily, they can choose to make a native build that'd run on PS4-spec hardware. It'd seem this new PS portable is meant to have performance around a PS4, maybe a PS4 Pro (considering when it's supposed to release)? Well, if that dev made such a scaled-down spec of their game, it'd automatically be natively playable on the new PS portable as well.

And, if someone were to buy a copy of that game, they could play it on their PS portable, or PS4, or PS5 or PS6. The choice is theirs.

However, I can see other uses for this type of PS portable, if it comes to fruition. Assuming they're smart with the button layout, it can probably also be used as another controller for a PlayStation console. If Sony make a push for scalable VR headsets by next gen (which I think is a necessity help make VR more mainstream), they could potentially offer bundles with a PS portable and an entry level headset at a good price. And, obviously, it could be considered a new iteration of the PS Portal with those use-cases, as well.

There's too much upside in making a new portable with native capability of at least all PS4 games (maybe even PS4 Pro level support in a docked configuration), native play of some PS5 & PS6 games (up to the developer), plus Remote Play & cloud streaming of all available games. We're not talking a device that has to do PSP numbers, and by its nature on appeal would do better than the PS Vita by a long shot, possibly even better than a Pro model refresh. I also think Sony could design it to be feasibly available at a good MSRP with good profit margin on the hardware from Day 1, basically like how Nintendo does their hardware.

 

[nial]

I thought we were going to discuss Physint being a PS6 launch title.

 

[thicc_girls_are_teh_best]

I thought we were going to discuss Physint being a PS6 launch title.

Oh that can be on the table, too. But actually, I've seen a lot of people mention that but was it specifically said in the presentation, or are people just guessing given the likely launch window?

 

[nial]

Oh that can be on the table, too. But actually, I've seen a lot of people mention that but was it specifically said in the presentation, or are people just guessing given the likely launch window?

The latter, haha.

 

[Audiophile]

Hopefully PHYSINT is 2030; that way it's got a better chance of being next-gen only and being a full-on showcase.

 

[thicc_girls_are_teh_best]

The latter, haha.

Oh okay. I did hear Kojima say "two years from now" so I assumed that's when we'd get the game, but that is too soon considering not just DS2 but also Overdose (which I guess'll release between DS2 and Physint).

Definitely looking forward to more information tho.

Hopefully PHYSINT is 2030; that way it's got a better chance of being next-gen only and being a full-on showcase.

On one hand that'd be desirable. On the other hand, it'd be a bit a letdown if something of it doesn't come for PS5. Maybe a scaled down conceptual demo for PS5/Pro, and then the actual game later as a PS6 exclusive.

Kind of like how Polyphony did the GT Concept games before the actual GT games (although those were for games on the same target platform). PS5-wise we'd probably still get some DS2 expansion betewen DS2 and the magnum opus.

 

[playsaves3]

Let reality sink in. A 4090 (Most likely the RTX 4000 series in general) is gonna beat out anything Sony puts out. They are gonna just keep pushing exclusives until their approach is no longer viable and most likely start to fail. I hate having to own a PS5 for exclusive titles. I much rather play the big games on my desktop. I imagine Sony is doing ALOT of R&D into AI upscaling right now to prolong things, so the next iterations of PS will probably be about that, VRR, and fake 8K maybe.

Probably rt but the 4090 will be a joke to it when it comes to raster

 

[thicc_girls_are_teh_best]

Probably rt but the 4090 will be a joke to it when it comes to raster

It'll probably also be a joke when it comes to RT. Or to put it another way, RT efficiency per FLOP.

Though it'll comfortably retain a raw compute advantage in terms of TF.

 

[Three]

I think the strategy for next gen is very difficult. Selling hardware at a massive loss and ever ballooning high budget games on a new small install base is effectively over I believe. So if I had to guess it would be that they will aim for a visual fidelity jump with PS6 sold at profit or with little to no loss and a mostly shared library. That is until a couple of years in when PS5 sales have slowed enough and margins on PS6 are high. I think they will also lean heavily into expanding user experiences with cloud, PC, and portable/mobile.

 

[playsaves3]

It'll probably also be a joke when it comes to RT. Or to put it another way, RT efficiency per FLOP.

Though it'll comfortably retain a raw compute advantage in terms of TF.

Yeah people really underselling how different the 2028 landscape will be mid range and possibly low range cards will outperform the 4090 pretty comfortably we will be on the 7090 at that point after all the better question is how close the ps6 gets to a 6090

 

[saintjules]

Saying PS6 doesn't exactly roll off the tongue.

 

[Audiophile]

Just mentioned in another thread; fingers-crossed hall effect analogue stick pots are standard on PS6 controllers..

Pair it with a super-fat battery, finer grain haptics, a slightly better speaker/mic, a higher quality dac/amp for plug-in headsets and a slimmer join between the handle and main body and anything else 'll be gravy.

Speaking of the gravy, I'd like to see those ideas about sensors come to fruition, temperature, sweat, heartbeat, stress sensors etc. Could have some cool gameplay implications.

Perhaps all peripherals going forward could have inductive coupling charging and they could offer powermats as another option next to docks.

I also think a single set of back buttons/paddles as standard would be great. I know they'll likely wanna save it for the Edge version (in which case they can offer two lots, with higher quality materials; and adjustability). But, if you wanna standardise extra functionality and provide more control possibilities in all exclusive games then you need it as a baseline feature. They'd have to be very svelte and unimposing, being naturally accessible while not getting in the way or being easy to accidentally hit.

Final thing, a hybrid of the Vita & PS4 D-Pads.

 

[Shin-Ra]

For DualSense 2 they need to split the controller in two halves like the PSVR2 Sense controllers.

Each hand gets fully independent feedback without vibration pollution from the other side.

 

[thicc_girls_are_teh_best]

I think the strategy for next gen is very difficult. Selling hardware at a massive loss and ever ballooning high budget games on a new small install base is effectively over I believe. So if I had to guess it would be that they will aim for a visual fidelity jump with PS6 sold at profit or with little to no loss and a mostly shared library. That is until a couple of years in when PS5 sales have slowed enough and margins on PS6 are high. I think they will also lean heavily into expanding user experiences with cloud, PC, and portable/mobile.

This is partly why I think the PS6 needs some differentiation factor aside just more power. Yea sure, it's going to be notably more powerful than the PS5, that's just a given. But is it worth pursuing 100 TF if it means you won't even get that many next-gen only games using that power until near the end of the generation, at budgets of maybe half a billion dollars if not more? I don't think it is.

Just chasing more power is a dead-end game anyway because no matter what, gamers on PC will have better options for pure power and in that respect PC gaming is a lot more prolific than say high-end arcade machines were in the '90s, where if a person easily wanted a gaming experience more than what the home console could bring, usually they had to go out of their way to an arcade to play certain games there for maybe a few hours before running out of money and heading back home. And people weren't buying those cabinets for home use because some of them costed as much as $20,000.

This is why I've been adamant on Sony finally using VR/AR as that differentiation factor. Some of their patents suggest they are definitely looking at more immersive gaming, and that doesn't have to just mean super top-of-the-line visuals. Again, even the current systems are getting very close to photorealism if you have the time, budget, labor & talent for it. Plus, development tools are getting smarter, better, and cheaper opening the path for smaller teams to do more with less. A PS6 with some satisfactory entry-level VR included in the package would mean devs can really get creative with game design options, and some nifty QOL features can be built in the UI at the user level for things like seamless transfer of content viewing between the headset and a traditional TV, or splitting the content display between both (say if one person is playing a 2-player game via VR and the other is playing it on a traditional TV display in the same room or two areas in the same home).

All they really need to do is get the PS.Link and Remote Play streaming tech as polished as possible, offload as much of the headset processing to the local box machine as possible, and develop some quality, inexpensive sensors and lenses for "good enough" display (prob no higher than 1440p per eye needed) and slim down the size of the headset, plus maybe include a camera for passthrough of surrounding view (and hopefully, a way of some AR where VR content can be framed in a mixed display with passthrough content). Anything that can get such a headset sellable with a good profit margin at an MSRP of $149 would be perfect.

I don't care if that'd mean a PS6 that's $599 instead of $499 and I don't think hardcore and core early adopters would mind, either. Just make sure there are some banger games at launch and games that justify that hardware direction. As for your idea of they'd do for a PS6, I dunno about them going for a for-profit model at launch, although they wouldn't do anything taking on heavy losses per unit, either. Something with maybe a $50/$60 loss-leading ratio could work, because if you're mainly selling to enthusiasts in those first few years, they are going to inevitably buy many games Day 1. So them buying a single game, or even just getting/continuing a basic PS+ sub, offsets the loss on that PS6. And over time, the production costs come down and they reduce the MSRP while growing to have profit margins on the hardware itself.

As for increasing user experience across other devices, I think that all depends on how Sony do it. They can't devalue the console itself, no matter what, so if it were up to me things like ports of non-GaaS titles to platforms like PC, at the very least, would at minimum come with some 4-6 year buffer window (longer in some cases). At least for new (i.e current-gen, as in of that current generation timeline) releases; certain remasters (including collections) and remakes of games well beyond 6 or so years old, could possibly be Day 1 across console, PC and/or mobile, but I'd make sure to always indicate that it loops back to something prioritizing the console, like an upcoming console-exclusive sequel or new IP arriving within 1-2 years tops of that type of release. I think Sony can focus on mobile more too, with some mobile-exclusive games based on known IP the way stuff like Pokemon GO! did it, and also mobile could be a good destination for more traditional AA mid-sized 1P games to find an audience, so a strategy where some 1P AA are timed exclusive to the console (then ported to mobile and PC), and some timed exclusive to mobile (then ported to console and PC) could be viable. And again, in those cases too, making sure there is something new and exclusive coming specifically for the console.

That's a type of strategy that I think could work in expanding audience on non-PS systems, but without crippling the value appeal of PS consoles to the demographics that ultimately matter most.

Yeah people really underselling how different the 2028 landscape will be mid range and possibly low range cards will outperform the 4090 pretty comfortably we will be on the 7090 at that point after all the better question is how close the ps6 gets to a 6090

Honestly it probably won't nor likely shouldn't get too close to it at all, at least when it comes to things like TF and whatnot. Because a 6090's probably going to be some 120+ TF monster...how much of that do you really NEED for video games, though?

I think a PS6 with sensible focus on AI-based technologies (for things like auto-LOD configuration/scaling, certain types of physics calculations, internal & external resolution scaling, draw call generation, various data I/O operations, mesh generation/adjustments etc.), good fixed graphics boosts in raster performance, good increase in RAM capacity & bandwidth, decent increase in caches etc. while offering say a "modest" 45 - 50 TF of performance would be perfectly fine and punch above its weight..

...especially if it brings other innovations beyond just graphics

 

[CLW]

Stop trying to make VR happen…..

 

[StreetsofBeige]

I think the strategy for next gen is very difficult. Selling hardware at a massive loss and ever ballooning high budget games on a new small install base is effectively over I believe. So if I had to guess it would be that they will aim for a visual fidelity jump with PS6 sold at profit or with little to no loss and a mostly shared library. That is until a couple of years in when PS5 sales have slowed enough and margins on PS6 are high. I think they will also lean heavily into expanding user experiences with cloud, PC, and portable/mobile.

This is partly why I think the PS6 needs some differentiation factor aside just more power. Yea sure, it's going to be notably more powerful than the PS5, that's just a given. But is it worth pursuing 100 TF if it means you won't even get that many next-gen only games using that power until near the end of the generation, at budgets of maybe half a billion dollars if not more? I don't think it is.

Just chasing more power is a dead-end game anyway because no matter what, gamers on PC will have better options for pure power and in that respect PC gaming is a lot more prolific than say high-end arcade machines were in the '90s, where if a person easily wanted a gaming experience more than what the home console could bring, usually they had to go out of their way to an arcade to play certain games there for maybe a few hours before running out of money and heading back home. And people weren't buying those cabinets for home use because some of them costed as much as $20,000.

This is why I've been adamant on Sony finally using VR/AR as that differentiation factor. Some of their patents suggest they are definitely looking at more immersive gaming, and that doesn't have to just mean super top-of-the-line visuals. Again, even the current systems are getting very close to photorealism if you have the time, budget, labor & talent for it. Plus, development tools are getting smarter, better, and cheaper opening the path for smaller teams to do more with less. A PS6 with some satisfactory entry-level VR included in the package would mean devs can really get creative with game design options, and some nifty QOL features can be built in the UI at the user level for things like seamless transfer of content viewing between the headset and a traditional TV, or splitting the content display between both (say if one person is playing a 2-player game via VR and the other is playing it on a traditional TV display in the same room or two areas in the same home).

All they really need to do is get the PS.Link and Remote Play streaming tech as polished as possible, offload as much of the headset processing to the local box machine as possible, and develop some quality, inexpensive sensors and lenses for "good enough" display (prob no higher than 1440p per eye needed) and slim down the size of the headset, plus maybe include a camera for passthrough of surrounding view (and hopefully, a way of some AR where VR content can be framed in a mixed display with passthrough content). Anything that can get such a headset sellable with a good profit margin at an MSRP of $149 would be perfect.

I don't care if that'd mean a PS6 that's $599 instead of $499 and I don't think hardcore and core early adopters would mind, either. Just make sure there are some banger games at launch and games that justify that hardware direction. As for your idea of they'd do for a PS6, I dunno about them going for a for-profit model at launch, although they wouldn't do anything taking on heavy losses per unit, either. Something with maybe a $50/$60 loss-leading ratio could work, because if you're mainly selling to enthusiasts in those first few years, they are going to inevitably buy many games Day 1. So them buying a single game, or even just getting/continuing a basic PS+ sub, offsets the loss on that PS6. And over time, the production costs come down and they reduce the MSRP while growing to have profit margins on the hardware itself.

As for increasing user experience across other devices, I think that all depends on how Sony do it. They can't devalue the console itself, no matter what, so if it were up to me things like ports of non-GaaS titles to platforms like PC, at the very least, would at minimum come with some 4-6 year buffer window (longer in some cases). At least for new (i.e current-gen, as in of that current generation timeline) releases; certain remasters (including collections) and remakes of games well beyond 6 or so years old, could possibly be Day 1 across console, PC and/or mobile, but I'd make sure to always indicate that it loops back to something prioritizing the console, like an upcoming console-exclusive sequel or new IP arriving within 1-2 years tops of that type of release. I think Sony can focus on mobile more too, with some mobile-exclusive games based on known IP the way stuff like Pokemon GO! did it, and also mobile could be a good destination for more traditional AA mid-sized 1P games to find an audience, so a strategy where some 1P AA are timed exclusive to the console (then ported to mobile and PC), and some timed exclusive to mobile (then ported to console and PC) could be viable. And again, in those cases too, making sure there is something new and exclusive coming specifically for the console.

That's a type of strategy that I think could work in expanding audience on non-PS systems, but without crippling the value appeal of PS consoles to the demographics that ultimately matter most.



Honestly it probably won't nor likely shouldn't get too close to it at all, at least when it comes to things like TF and whatnot. Because a 6090's probably going to be some 120+ TF monster...how much of that do you really NEED for video games, though?

I think a PS6 with sensible focus on AI-based technologies (for things like auto-LOD configuration/scaling, certain types of physics calculations, internal & external resolution scaling, draw call generation, various data I/O operations, mesh generation/adjustments etc.), good fixed graphics boosts in raster performance, good increase in RAM capacity & bandwidth, decent increase in caches etc. while offering say a "modest" 45 - 50 TF of performance would be perfectly fine and punch above its weight..

...especially if it brings other innovations beyond just graphics

Consoles are always sold cheap for sake of the razor blade model, but with consoles inching up $100-200 the past 10 years it seems to be ok with gamers. Also, with PC gaming rigs being expensive with high priced GPUs, the gap in pricing seems wide enough console makers can probably inch up another $100-200 and be ok. At least for PS and Xbox. Maybe the Nintendo gamer wont budge as the pricing model is always dirt cheap at about $250-300. I dont see Nintendo going $400+ US.

 

[thicc_girls_are_teh_best]

Just mentioned in another thread; fingers-crossed hall effect analogue stick pots are standard on PS6 controllers..

Pair it with a super-fat battery, finer grain haptics, a slightly better speaker/mic, a higher quality dac/amp for plug-in headsets and a slimmer join between the handle and main body and anything else 'll be gravy.

Yeah, this would be great for a new iteration of the DualSense, especially the hall effect analog sticks. That probably needs to become standard in general going forward.

Speaking of the gravy, I'd like to see those ideas about sensors come to fruition, temperature, sweat, heartbeat, stress sensors etc. Could have some cool gameplay implications.

If Sony do go for a heavy VR/AR push next gen and can get a VR headset slimmed down in production costs & form factor enough to make it default with every system, and designed a controller that can complement both traditional and VR play, I could definitely see them adding these types of features to a premium version of that controller.

We're talking something probably closer to $199 (maybe even $249) but for the tech and immersion factor would be worth it. They'd be selling to the same type of people who would buy a premium VR headset anyhow; the price is not a disruptive factor for those types.

Final thing, a hybrid of the Vita & PS4 D-Pads.

I just want D-pads that are comfortable to use . That's one thing I'll give Xbox credit for with the Series controller; some hate the clicky sound but I actually think it's a nice D-pad. Reminds me of the Saturn and Neo-Geo D-pads in a sense.

Though aesthetically, I really like Sony's D-pads as well; if they could add some smaller buttons to fill out a circular shape with the D-pad and cover the mid-cardinals, and shape the buttons so that they form like a concave dip into the center (but still comfortable at the edges), I think that could be a perfect D-pad.

For DualSense 2 they need to split the controller in two halves like the PSVR2 Sense controllers.

Each hand gets fully independent feedback without vibration pollution from the other side.

This is exactly what I'm thinking they should do, too, but for an additional purpose where you can then use the split form as VR controllers with the headset.

And maybe instead of two halves, the controller would be formed out of three parts, where the central part also acts like a hub and relay extender for the wifi & bluetooth signals, and can also help with room scale mapping for VR or AR content.

That said I know such a controller would cost more to buy on its own...probably at least $80 or maybe even $90. But they could also design a "regular" version that doesn't segment and doesn't have the extra tech that'd be priced like the current regular DualSense.

Consoles are always sold cheap for sake of the razor blade model, but with consoles inching up $100-200 the past 10 years it seems to be ok with gamers. Also, with PC gaming rigs being expensive with high priced GPUs, the gap in pricing seems wide enough console makers can probably inch up another $100-200 and be ok. At least for PS and Xbox. Maybe the Nintendo gamer wont budge as the pricing model is always dirt cheap at about $250-300. I dont see Nintendo going $400+ US.

I agree with this; as long as the perception of value is retained, I don't think $599 for a mainstream console carries the same death sentence it did for PS3 during its generation, or especially even higher prices like the 3DO's $699 in its day. Like you said, prices for PC gaming systems just keep increasing; on the GPU front even for mid-range cards they on their own are costing as much as an entire console, let alone the high-end cards which seem to be pushing regularly $1299 - $1499 at launch and could go higher.

But that's a privilege PC GPU cards are able to pursue, not consoles. Consoles just have to make sure they provide a suitable enough performance increase from last gen and have other means (such as convenience) to sell their value on. That said I do think Nintendo's going to try some type of "premium" Switch 2 SKU pushing $499; more curious if the standard Switch 2 will be $299 or $399, and then if they will try something like a Switch Lite near launch to maintain a $299 option if the standard Switch 2 can't price below $399.

 

[StreetsofBeige]

I agree with this; as long as the perception of value is retained, I don't think $599 for a mainstream console carries the same death sentence it did for PS3 during its generation, or especially even higher prices like the 3DO's $699 in its day. Like you said, prices for PC gaming systems just keep increasing; on the GPU front even for mid-range cards they on their own are costing as much as an entire console, let alone the high-end cards which seem to be pushing regularly $1299 - $1499 at launch and could go higher.

But that's a privilege PC GPU cards are able to pursue, not consoles. Consoles just have to make sure they provide a suitable enough performance increase from last gen and have other means (such as convenience) to sell their value on. That said I do think Nintendo's going to try some type of "premium" Switch 2 SKU pushing $499; more curious if the standard Switch 2 will be $299 or $399, and then if they will try something like a Switch Lite near launch to maintain a $299 option if the standard Switch 2 can't price below $399.

To me, consoles inching up in price's biggest roadblock is if it gets too high there's going to be gamers that just say fuck it and will go smartphone or if they got a half decent PC at home, they'll just game there instead of buying a console. For people who got money are are super entrenched in hardware the pricing seems ok to inch up I think. But those people who cant afford it will bail. You can already see it as console hardware sales havent even really increased vs generations ago despite tons more people gaming with money in their pocket. A lot just went mobile or PC as everyone with some money has both of those gadgets already.

It's like homes vs condos.

When prices all go up together similarly, there's no skew. If houses go up too fast, then the cheaper condos will catch up too. If houses drop a lot, so will condos. If prices for any home go up too fast, people who cant afford anything bail and rent.

 

[thicc_girls_are_teh_best]

To me, consoles inching up in price's biggest roadblock is if it gets too high there's going to be gamers that just say fuck it and will go smartphone or if they got a half decent PC at home, they'll just game there instead of buying a console. For people who got money are are super entrenched in hardware the pricing seems ok to inch up I think. But those people who cant afford it will bail. You can already see it as console hardware sales havent even really increased vs generations ago despite tons more people gaming with money in their pocket. A lot just went mobile or PC as everyone with some money has both of those gadgets already.

It's like homes vs condos.

When prices all go up together similarly, there's no skew. If houses go up too fast, then the cheaper condos will catch up too. If houses drop a lot, so will condos. If prices for any home go up too fast, people who cant afford anything bail and rent.

Yeah I understand what you mean here. However, I would also say it only becomes a problem if prices don't reduce quickly enough over time for when the more price-conscious casual and mainstream gaming customers would be ready to jump in. For the hardcore and core enthusiasts, even if there is a price creep they notice, they won't care too much and still buy in as long as value is perceived regardless. They will want the newest tech as soon as it's available, and unless something acting as a traditional console isn't pricing itself like a high-end PC, that value perception will always remain to some level (though the more it's present, the better).

If a next-gen system launches at $599 but then remains that price four years later or only drops down by $100, that's when it could really run into problems with reaching higher volume of sales. Because that's when the causal and mainstream customers start to come in at higher numbers, and they want a more mainstream price to buy in at. $599 or $499 isn't going to cut it. And getting prices reduced down sensibly over time is 100% on the platform holder to accomplish with good planning and working with component providers to ensure the same.

As for how consoles can continue to provide optimal value proposition in an increasingly competitive entertainment market with alternative gaming products like PCs, smartphones etc.? Well, that's also up to the platform holders to determine. I can picture great and sensible stuff in that regard for Sony, Microsoft, even Nintendo, and I and others have been doing so ITT and other threads, feels like. It's just a question if the companies pursue those options or not, or how much they do vs. how much they don't.

 

[thicc_girls_are_teh_best]

Another MLiD podcast; they talk about some console stuff here but for purposes of this thread, I'm mostly focusing on the PlayStation portable.

They claim that it's "definitely" in development, though dev so far is only at the high/concept level, and money's been exchanged between Sony and AMD. While there's the possibility it could get cancelled (as happened with the Series X Elite and Project Keystone), there are a lot of reasons why the handheld could definitely make it to market.

The most interesting part in their discussion, IMO, is the idea that the PS handheld would basically be a PlayStation 6 in handheld form. AKA, Sony takes the same APU design from the home console, disables a portion of it, clocks it lower, and uses that modified version as the silicon for the portable. Assuming they'd be using 3nm for the PS6's APU, and the APU is chiplet-based...provided the PS6 isn't a massive chip, I do think there's merit to the idea.

However, I did some rough number crunching, and in order for Sony to make it work, they'd need to likely both use a chiplet-based GPU (with blocks of shader cores, for example), and have custom logic that can automatically scale down LOD, mesh/geometry, resolution etc. settings to run on a handheld variant (to take that part of the processor out of the developer's hand and free them up to focus on other stuff).

For example, if the PS6 has an APU with 128 CUs, it'd be ideal if the GPU were a chiplet-based design with four blocks of 32 CUs (2,048 shader cores). That way Sony could use the same wafer for shader cores, and only need to physically include 1/4th the amount in a portable that the console would need. Then just match the clocks for whatever performance profile is required to fit in a handheld TDP (if it were to also be a dockable, then the dock could provide passthrough cooling and power to allow the GPU an upclock). On 3nm by 2028 that should be very doable.

 

[Audiophile]

.

 

[cinnamonandgravy]

cpu = 19c/19t
1 big daddy core, 12 normal cores, and 6 SPUs.

 

[Loxus]

I wonder if Mark Cerny would revive this for the PS6 system memory if given the opportunity.

If PS6 is heavily AI based. I have a feeling system memory may need some custom work like the IO complex in thePS5.

Maybe do something like this instead of an interposer.

Another thing I wonder, is if a SSD Controller can be placed as 3rd Party.

I imagine doing so would greatly reduce latency.

 

[hussar16]

Typo'd it as "Song" instead of "Sony" in previous post..



Been thinking about this and given that they don't seem to be inclined to do good software emulation for PS1/PS2 nor any sort of emulation for PS3; I was wondering why they don't just bite the bullet sooner rather than later, circumvent the entire issue and port the designs for the PS1 CPU/GPU+, PS2 CPU/GPU+ & PS3 CPU on to a single little APU chip.

They may have some issues in terms of licensing with IBM/Toshiba on the PS3 but I doubt it'd an insurmountable issue. They could probably skip the Nvidia RSX and simply emulate that on a really basic AMD/ARM GPU element; as well as bespoke functions of older platforms like sound chips etc. on a basic AMD/ARM CPU element.

They could call it the "PS123" chip and sell it on a small, semi-budget legacy console with a CD/DVD/BD drive as well as access to the store for as many titles as they can provide license-permitting; calling it the "PlayStation Legacy", eventually they could shrink the chip enough and put it in a portable device (Portal 2?) alongside a shrunk PSP/Vita chip set; and even include it as a chiplet next to the main die/s on a latter premium/mainline platform such as a PS6 Pro -- as a marketable feature. The "Legacy" console and its disc drive should be multi-region with 50/100 & 60/120Hz output as well as various scaling/display modes; and digital licenses should offer the option to download any regions' release.

PS4/PS5 and forward will likely be covered by hardware BC for the next few (or remaining) generations.

Once it's done, it's done and while the initial work may be difficult (primarily shrinking the ancient designs, having adaptive modes for memory speeds/latencies/CPU-GPU clocks, matching CPU timings and identifying what hardware components need to be directly recreated on die and what can be flawlessly emulated on a standard CPU/GPU element); it just seems -- relatively speaking -- like the most straightforward approach to the issue, getting in there at a base level and solving all the issues in one go. Sony will also be able to finally dump all of those life support OG PS3s and old Cell blade servers they have racked up for streaming; and implement these likely miniscule chips into their streaming servers alongside whatever they're using or will be using for PS4/5/6 etc.

I could honestly see ps6 or a future console that is nothing but a USB plug in and play. It Wil be the future

 

[thicc_girls_are_teh_best]

cpu = 19c/19t
1 big daddy core, 12 normal cores, and 6 SPUs.

They'd still need dual-threaded cores for PS5 BC, and I dunno if they'd bring back SPUs in that way. They could take some AI/ML custom tech and add it to the CPU, with modifications to function like the SPUs (the way they customized a RDNA2 CU to act like one for Tempest Engine).

That said, if they're switching to ARM next gen for the CPU (it's very possible), they could do maybe 4 Big cores, 8 Little cores (BIG.little style) with the 8 Little cores being dual-threaded while the Big cores can be some other configuration. The Little cores would need to be at least as capable as the PS5's CPU cores though (preferably quite better but able to downclock and scale back perf to match PS5 & Pro in BC tasks).

I wonder if Mark Cerny would revive this for the PS6 system memory if given the opportunity.

If PS6 is heavily AI based. I have a feeling system memory may need some custom work like the IO complex in thePS5.

Maybe do something like this instead of an interposer.

Another thing I wonder, is if a SSD Controller can be placed as 3rd Party.

I imagine doing so would greatly reduce latency.

Something else they would maybe want to consider if looking into accelerating AI is HBM-PIM (Processing-In-Memory). It's a type of memory that has AI accelerators built right into the memory modules, and has been shown to increase AI & ML calculations by upwards 70%, while saving on energy consumption (since the data isn't being transferred across the memory bus, a lot of energy usage is not required).

There is a DRAM-PIM variant, but not GDDR-PIM AFAIK. GDDRW (Wide, I think) exists, but that's for capacity and footprint targets, not bandwidth increases. If by some chance SIE can't leverage HBM-PIM for the next system, then they should at least go for some HBM. Barring that, better hope GDDR7 or whatever's after that is up to task. Without PIM though, I think they'd really need to boost processing near memory as much as possible, so perhaps HBM3 or something that can utilize an active interposer would actually be preferable to an HBM getting rid of that.

 

[Audiophile]

THIS CHART I made might come in handy for eyeing up potential memory configs with GDDR; when spec-ulating on PS6.

Layout, speed, amount, number of modules and memory bus real estate will be a key point in optimising the system design; and balancing for perf/cost. Could be useful for Xbox too as well as gaming/general consumer GPUs into the next decade.

It's pretty easy to work out on the fly: bus-width x chip speed, then divide by 8... but it's nice to have a clear table laid out in front of you, I looked everywhere and couldn't find a comprehensive one.

I'll do a full LPDDR4/4X/5/5X/6 chart soon as well..

 

[N1tr0sOx1d3]

One thing is for sure there won’t be a psvr3 maybe for ps7

 

[Deerock71]

8K at 240fps for the total loss.

 

[MrRenegade]

Make the new console without wires. You just put it on the carpet, and it cleans it. While over the air power and internetz is provided. That would be a proper innovation. It cleans your carpet, streams the game to your monitor and also analyzes the clipped nails whether you have some kind of illness unknown to you.

 

[Jesb]

You can bet your last $ that VR isn’t going to be part of it.

 

[Hudo]

The biggest innovation Sony could do for PS6 is trying to solve their shitty production pipeline and expanding their first-party studio lineup.
Making it more powerful than the last console won't do shit if your studios take 7+ years to make one fucking game.

 

[Panajev2001a]

The biggest innovation Sony could do for PS6 is trying to solve their shitty production pipeline and expanding their first-party studio lineup.
Making it more powerful than the last console won't do shit if your studios take 7+ years to make one fucking game.

The HW may be an enabler of that, helping studios make their games faster while offering more potential.

 

[Panajev2001a]

One thing is for sure there won’t be a psvr3 maybe for ps7

I hope there will be, PSVR2 was a great HW update at a contained price. It is worth iterating again as the market is improving.

I think PSVR3 would not make the mistake of losing PSVR2 compatibility and so hope they find a way to sort out PSVR BC too.

 

[Radical_3d]

I was contemplating this thread resurrection and watching all this possible future techs and found myself thinking: what’s the point, if in the end Sony won’t have a fabrication process of next generation because Apple buys it all. Then occurred to me that half the rich fucks of the planet are in GAF actually. How about we stop buying iPhones a couple of years before the PS6 launches so we can have a competitive hardware. I know, I know. Wasting money makes us happy. But, just a little bit so we can have a nice generation? /s

 

[PeteBull]

From ps5pr0 leaks and how slow hardware industry is advancing we can easily predict that as long as ps6 launches in 2028(holidays) it wont be that much stronger from currently top pc gpu 4090, at max it will be 50% stronger so since 4090 is about 3x stronger from ps5, then ps6 will be 4-5x stronger only from base ps5.

 

[Matt_Fox]

From ps5pr0 leaks and how slow hardware industry is advancing we can easily predict that as long as ps6 launches in 2028(holidays) it wont be that much stronger from currently top pc gpu 4090, at max it will be 50% stronger so since 4090 is about 3x stronger from ps5, then ps6 will be 4-5x stronger only from base ps5.

If that turns out true it would be a great result!

50% stronger than a 4090 would massively raise the bar.

 

[PeteBull]

If that turns out true it would be a great result!

50% stronger than a 4090 would massively raise the bar.

For comparision, ps1 to ps2 was around 100x jump, not 4-5x.
When jump between hardware power is tiny then ppl always complain about "next gen games" looking avg/mid, back in ps2 days we got TTT at launch, and jump was crazy big from tekken 3 coz it was much better looking version from even system12- famous arcade one.

Just look at proper power jump, so ps3 to ps5, and how good same game, aka demons souls remaster can look, provided ofc high quality very competetent devteam and big jump in hardware power.



And on top of all that we got stable 60fps instead of dips below 15(in blighttown) that we had to suffer tru on ps3.

 

[Rocco Schiavone]

If that turns out true it would be a great result!

50% stronger than a 4090 would massively raise the bar.

An APU 50% stronger than a 4090???

 

[night13x]

ps6 from sony official standpoint:

EXPERIENCE GAMES UP TO 16K RESOLUTION!!!!!

reality:

devs releasing 4k games with better ray tracing - running at 30 fps.

 

[Sanepar]

Powerwise will be around real 50 tflops. I would say a little bit better than a 4090(42tflops) if it is 2028.

But Sony needs to figure out how to delivery at least 3 aaa per year. They need more studios and focus what they do best = single player only games

 

[yurinka]

If that turns out true it would be a great result!

50% stronger than a 4090 would massively raise the bar.

PS6 will be released late 2027/early 2028, let's say at an aprox. $400-$500 price point. They may sell it at a loss of maximum let's say $100-$200.

Meaning, must be something that can be built at that point in time for aprox. $500-$700. Not just the GPU: the entire console. At that point of release there will be GPUs with that horsepower, but not at the required price.

ps6 from sony official standpoint:

EXPERIENCE GAMES UP TO 16K RESOLUTION!!!!!

reality:

devs releasing 4k games with better ray tracing - running at 30 fps.

I think that by 2027 their own DLSS will be good enough to allow them render games at 1080p 30fps and scale them to 4K and 60fps (maybe 8K, or 120fps too).

But Sony needs to figure out how to delivery at least 3 aaa per year. They need more studios and focus what they do best = single player only games

They already are releasing at least 3 AAA per year. Every year the revenue from selling games, both in PS and in general keeps decreasing and being replaced by revenue from selling addons. Most best selling and most played games are GaaS/MP and their market share keep growing.

They can't be limited to SP non-GaaS games, and even less limited to their console. Because if market continues evolving as did in the last decade or so in the long term to be limited to single console non-GaaS SP AAA games won't be sustainable. Specially considering that every new generation the AAA budgets skyrocket compared to the previous one.

 

[Audiophile]

4 mores years of dev experience, software development and hardware development in ML/AI & RT/PT (just think how far we've come on these in the last 5yrs in the PC space).

No widespread increase in target display resolution (4K) which means half the frame budget will no longer getting thrown away just on pixels, in fact with advanced, mature ML upscaling/frame-gen we'll probably be rendering less pixels and getting better IQ.

Engines which over the next 4yrs will better integrate Mesh Shaders and/or Virtualised Microgeometry, the aforementioned ML/AI & RT/PT; as well as better utilise the SSD-I/O functionality.

Cheap stackable cache and additional cache dies, GDDR7 chips that'll be much more mature than GDDR6 was for PS5, N2/N2P/A16 will be 18-24mths old by mid 2028 and backside power delivery (BPDN) may be a prospect (preliminary tests show 10% higher clocks and 30% more power efficient as a baseline) and finally multi-GCD chiplet SoCs; which will likely be much more common by then and could cut cost by 30% or so. Frankly I'm not sure if it's even possible to make an APU with an adequate boost even taking into account all the prospective efficiency gains available. I think Chiplets may be a necessity.

There's a lot of areas to cut costs too. RAM likely won't need to go up that much and GDDR7 will offer a lot more flexible configuration options to dial in the sweetspot. The SSD could just do a straight doubling to 2TB @ 11-13GB/s via 8Ch on PCIe5 which will be old tech by then and be cheap. I doubt more than 11-13GB/s RAW will be of any use. The optical drive can be offloaded to a readily available $50 add-on for all SKUs and the console could put the extra into better core hardware. They can settle for Wi-Fi6e rather than 7, Bluetooth 5.2 will be perfectly adequate, USB4 will be mature and HDMI2.1b @ 48Gbps will likely suffice.

Infinity Fanout Links mentioned a few times previously could allow for two small GPU GCDs on an advanced node like A16 with BPDN, then savings could be made on a 16C Zen 7C made on N2P with a once-again cut down FPU, SMT-logic deleted, a relatively small L3 cache on-die with more stacked on top, clocks in the 4.2-4.5GHz range; likely providing 2.5-3x the base PS5's Zen2 capability per core. Old node N5/N4 MCDs could be used for the SoC/GPU L3 caches and for offloading memory buses, perhaps better facilitating different memory bus setups alongside G7 memory chips with 2GB, 3GB or 4GB capacities and I'm pretty sure at least 4GB will be dedicated to the OS, so they could use low bin, slow chips for that much.

A lot of things line up pretty well for 2028. Plus I expect as always there'll be a few interesting curveballs we didn't think of.

I think we'll see 38-40TF Single-Issue FP32 but a significant bolstering of RT/PT throughput, very capable ML/AI hardware and the rest is going to be all efficiencies. 24GB dedicated just to GAME will probs be enough and a 12GB/s SSD could fill it out in 2s. But after that it's gonna be all about system bandwidth/throughput. ~1TB/s on the RAM and a well balanced distribution of L0/L1/L2/L3 cache to further reduce the load.

This gen has been a massive half measure with terrible timing; all this cool stuff likely gets to fly next gen.