• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Let's Design The Mid-Gen Refreshes, Part 2: SONY PLAYSTATION 5 PRO

Do you even think there'll BE a PS5 Pro in 2023? If so, of what nature/type?

  • Yes; 20.6 TF BEAST MODE, electricity bill be damned!

    Votes: 26 33.3%
  • Yes; much more modest like a Slim tho, no spec bumps just smaller and less watt-consuming

    Votes: 18 23.1%
  • Yes; slight/conservative TF bump with upgraded specs in memory and such. Cross btwn a Pro and Slim

    Votes: 16 20.5%
  • No; they'll ride it out with PS5 Disc and PS5 Digital until PS6 is ready

    Votes: 18 23.1%

  • Total voters
    78
  • Poll closed .
So I'm finally ready to post my write-ups on mid-gen and next-next gen console speculations. It's a lot, so I'm breaking it down into parts focused on specific systems. And I also tried focusing on the design philosophy of the products keeping in mind the business models Sony and Microsoft seem to be going for. I did a Part 1 here but that was just in trying to figure out PS5 power consumption amounts, mainly on the GPU side. That's something probably worth keeping in mind as the PlayStation-related stuff is touched up on.

I'm starting with the mid-gen refreshes, one system at a time (originally planned to do all three mid-gen refreshes in one thread, but the write-up grew VERY quickly, so better to give each system its own thread), and then I'll get to the next-next gen stuff. I'm also open to ideas on suggestions with tuning some of these specs, since this needn't be an open-and-shut deal. Seeing that the embargos were finally lifted today, let's start with Sony, and the PlayStation 5 Pro...

----------

In my honest opinion, I think market and technology realities are going to be the driving forces behind what goes into the mid-gen refreshes more than anything. That is to say, I personally think anyone hoping for HBM2, super Big Navi-tier GPU upgrades pushing 20 - 30 TF etc., are in for a massive disappointment in that regard. There are no feasible advances in terms of node shrink reductions, surrounding memory technologies (that would be affordable), or pricing for component R&D and system designs that would enable such mid-gen refreshes being a reality........for the MOST part ;)

Due to this, I'm of the personal opinion that the mid-gen refreshes from Sony and Microsoft will focus on the following main goals:

>Greatly reduce peak power consumption targets​
>Notably reduce system physical sizes (this will be a particular goal for Sony with their mid-gen refresh)​
>Refine specific technological design concepts and techniques as a testing grounds for what 10th-gen systems could bring​
>Improved implementation of Infinity Cache for the L1$ (possibly cut down in terms of full supported cache size​
for the consoles however)​
>AI dedicated silicon geared towards offloading some aspects of game artificial intelligence models.​
>Specific vision processing/sampling silicon baked into the GPU. This could be implemented as a single​
core to each Shader Array. Would help with real-time image processing. This is something that can​
already be done on the regular GPU shader cores, but specialized silicon for the task that integrates​
into the GPU pipeline would do it better, and with lower power consumption.​
>Improved image upscaling (silicon budget dedicated to image scaling and upsampling)​
>Explore a handful of new technological designs and concepts, which could be iterated on with 10th-gen systems​
>Provide modest performance enhancements of 9th-gen game content​
>Increase storage baselines​
>Keep prices no higher than 9th-gen systems, preferably lower...for the most part​
These are the guiding principals I feel are going to drive the mid-gen refreshes from Sony and Microsoft. A bit above I said that there are absolutely some things we SHOULDN'T expect with the mid-gen refreshes. However, I do think it's fair to give a brief mention of at least some of the things that will likely be implemented in them:
>GDDR7, to replace the aging (by 2023) GDDR6​
>RDNA 4-based GPU designs, with some RDNA 5 elements custom-built into them. This is assuming a consistent 14-month​
period for each RDNA generation step, going off the time span between RDNA 1 and RDNA 2 (14 months). So, RDNA 4 spec and​
GPU products would be ready for mass-market by March 2023, and RDNA 5 by May 2024. So you can infer from here timeline​
for mid-gen refreshes would range between 2023 and 2024.​
>Probably some integration of CDNA 4 features into the GPU designs (likely through extensions of the shaders in the RDNA​
silicon, though I can see one of the two (mainly Microsoft) integrate some actual CDNA 4 silicon into their mid-gen refresh)​
>At least Zen 5-based CPU designs (following same logic for RDNA generation timings above, Zen 5 would be ready for market​
by March 2023)​
>At least something chiplet-based (this will most likely be Sony)​
>At least some integration of early-stage persistent memory, such as 3D Xpoint or ReRAM (for reasons explained later, this might​
more likely be something Sony pursues in particular)​

Now then, maybe it's time we move on to giving some system speculations now? Let's go...

7Qv4q73m9Nif9CpxtfVWs6.jpg


[SONY]

Sony's PS5 Pro mid-gen refresh will most likely release in 2023. I see it implementing a chiplet design, not as 2x 36 CU chips, but as 2x 18 CU chiplets, to mimic the base PS5's GPU setup, only without needing four disabled CUs present (on a chiplet design, redundant silicon doesn't need to actually be present on the die). It will be RDNA 4-based, but also take features liberally from the RDNA 5 standard and customize the GPU setup with some of those features. It's very possible Sony would take more from the RDNA 5 spec into their GPU design compared to Microsoft.

Being a chiplet design, we already know the chances are strong that RDNA 3 will be chiplet-based in some capacity. Regardless, I did have my own idea for how a chiplet setup could pan out. On PS5 Pro, each of the two chiplets would feature 18 CUs and 2x 16 ROP/ColorDepth blocks (32 ROPs per chiplet).

To network the two chiplet in tandem, some of the typical GPU logic would need to be split off onto smaller complementary chiplets. One of these would be the "Unifying GPU Chiplet", or UGC for short. The UGC would handle a bevy of things. Chiefly:

>The UGC handles the access routines for the GPUs to main​
system memory by having the DMA built into it. The UGC chiplet​
block then has Infinity Fabric links at a data rate of 640 GB/s​
to the GPU chiplets (links of 320 GB/s to each GPU chiplet), so​
they can then work with the data as required.​
>There could be a Drawcall Management Block (DMB) on the UGC.​
It would be paired as an extension to/of the Command Processor,​
and be responsible for automating and managing the processes​
for issuance of GPU work to the hardware components of the​
chiplets which the Scheduler is mainly responsible for (and​
links directly to both chiplets via the IF links that are​
mentioned above).​
>The Command Processor, and other things such as the Scheduler,​
are in this block. The Shader Input blocks are in the chiplets​
themselves, one to each chiplet, and both have an independent​
connection to this block for data throughput input.​
*The Shader Input blocks may need some extra functionality​
to provide feedback to the UGC, maybe as a means of some​
hardware on the GPU chiplets themselves able to detect​
when individual CUs on eiher GPU chiplet is free for more​
work to be scheduled to it, that can then communicate​
with some silicon on the UGC complimented with the UGC's​
Drawcall Management block, to prioritize new work to those​
CUs to ensure peak saturation of GPU resources are always​
maintained.​
>There is also a small block of cache on this block: 256 KB L0$,​
1 MB L1$, and 4 MB L2$, for any required pre-processing and​
dispatch work, plus any drawcall instructions that can be​
saved for later.​
>The design of the Unifying GPU chiplet block is that it houses​
the traditional GPU components for drawcall instruction sorts​
and issues to CUs in the SAs of the SEs, and has its own block of​
L2$ that it can share data to the CUs with if required. Each​
chiplet's CU has its own L0$, and share an L1$, but there is​
no L2$ for the chiplets in the same way as on PS5 base. To​
compensate, the shared L1$ size of the CUs is enlarged by 20%​
per shared L1$.​

The other big complementary GPU chiplet component in the PS5 Pro would be the Unifying Framebuffer Chiplet block, which can be called the UFC for short. This would house the Display block typically seen in AMD GPUs, among some other things, and likely include some various combine modes for dual framebuffers (for the two GPU chiplets) that could selected as presets by developers depending on what type of rendering pipeline they'd wish to utilize for their game. You can think of these various rendering display preset combinations for the dual framebuffers as a mix of the SEGA Saturn's dual framebuffers and the SNES's various Mode settings (such as Mode 7).

The goal here, though, would be to have programming complexity reduced simply to the developer selecting a combination mode for the two framebuffers as long as they understand how the modes function the hardware itself should do the heavy lifting in combining and sorting the stitched outputs depending on what the game wants, and the combination modes should also be able to be switched between 1-2 cycles, to mix combination modes on the needs of what the game needs for optimal output. This is key for allowing maximized use of the framebuffer capabilities, but it also means the UFC needs to have these different framebuffer combination preset modes readily accessible on some type of private local memory. It's best to picture it, then, as an advanced VDP (Video Display Processor); some chunk of NOR flash and embedded SRAM cache in the UFC would be best for this (the NOR flash could store the presets and even allow for XIP (Execute In Place) if desired, while the SRAM would be for fast memory; some hierarchy of L0$, L1$ and L2$ is probably best).

Aside from the aforementioned GPU talk, a PS5 Pro would likely see improved support for PSVR2, with some bump in the Wifi 6 standard. For wired and perhaps wireless dongle-based PSVR connectivity, there could also be a Thuderbolt port provided via supercharging the USB-C port. The last big technological push I could see for a PS5 Pro is inclusion of persistent memory. Sony actually have some patents for ReRAM, which can potentially be used as both a storage-class and DRAM-like class memory technology. By the time a PS5 Pro would be ready, I think Sony would at least have storage-class ReRAM ready. The goal of it would be similar to the role Optane memory serves on compliant desktop PCs; as a bridge between storage-class memory and system RAM.

A block of 32 GB of ReRAM developed in-house (and likely manufactured/fabbed by Sony via TSMC) would be able to provide a notable performance boost to data I/O on a PS5 Pro while having much lower latency than NAND, support for smaller granularity levels in block data sizes, much higher endurance P/E cycles, and more bandwidth compared to even high-class SSDs on the market. While there is currently no commercial ReRAM on the market, there is at least one company with an IP license for storage-class ReRAM providing 25 GB/s of bandwidth. By the time of a PS5 Pro, especially if the ReRA itself had time to mature in the commercial market from 2021 or even 2022 and onward, Sony could possibly have a 24 GB/s - 25 GB/s bandwidth ReRAM solution that could be implemented in a mid-gen refresh at an affordable rate, serving as a great starting ground for similar technology in a PS6.

Due to this, however, I actually DON'T see them doing too much with an SSD I/O spec bump. While the SSD size will likely double (to 1.536 TB, as 6x 256 GB modules, most likely Toshiba brand as in the PS5 itself), the actual bandwidth performance will very likely remain the same. So, 5.5 GB/s raw bandwidth with compressed typical ranges of 11 GB/s - 12 GB/s, and up to maximum lossy compression range of 17 GB/s - 22 GB/s. This will still be very impressive even at the time of PS5 Pro and provide perfect compatibility with PS5 base, it just wouldn't be the fastest option available anymore. However, considering the investment in ReRAM to make up for this, it's not a bad trade-off.

Regarding main memory, GDDR7 would be the standard. HBM2 would simply be too disruptive as a technological shift to implement in a mid-gen refresh, and still likely carry a price premium compared to GDDR7, while not offering too large a performance benefit (at least in terms of bandwidth; latency would probably be a different conversation) within a price bracket suitable for a mass-market mid-gen console refresh. While it would likely provide lower power consumption, the mid-gen console refreshes would still get more than enough power reduction through other means, to have enough to justify GDDR7 which would, most likely, provide at least SOME power consumption reduction over GDDR6.

For PS5 Pro in particular, Sony would very likely stick with a 256-bit memory bus (they seem to love this bus size ;) ), and they'd want at least some type of increase of GB per TF bandwidth over PS5 base (~ 43 GB/per TF), regardless of how features like Infinity Cache on AMD's RDNA architectures shape out and develop. For those reasons, even if it'd require a slight overclock, it's very possible Sony would go for 20 Gbps GDDR7 modules, as 8x 2 GB modules, for a total of 80 GB/s per module, and a system bandwidth total of 640 GB/s on a 256-bit bus.

Audio would likely be a slight iteration on the Tempest Engine; if possible, it could have some of the SPE-style logic simplified further in order to allow for even easier utilization by developers, and a slight performance increase. Nothing too radical, however; they'd want to ensure it doesn't compete too much in terms of bandwidth with the CPU and especially the GPU. The CPU, as hinted way earlier, would be Zen 5-based; a similar 8 core/16-thread setup as the PS5's Zen 2, with better IPC and not only a unified L3$ cache (which Zen 3 would have already introduced), but some implementation of Infinity Cache on the CPU cache level side as well, this likely being a standard Zen 5 CPU feature however, but nonetheless worth utilizing. The same 3.5 GHz clock of PS5 would be supported, but a clock increase to something like 3.8 GHz or even 4 GHz would not be out of the realm of possibility.

Finally, the GPU. As mentioned before, no 72 CU GPU design here; while a chiplet approach would be supported, we'd see it as 2x 18 CU chiplet blocks. Process-wise, while 3nm (perhaps even 3nm EUV) would be readily available by this time in a general sense, DO keep in mind that costs are NOT scaling down with node shrinks; rather, the opposite is happening i.e prices are INCREASING. With investments already placed in on the ReRAM and (very likely) customizations to any aforementioned features of the GPU chiplet design that don't end up being standard in the RDNA spec by this point, to keep costs down and place investments in other areas Sony would likely go for 5nm EUVL instead, saving 3nm (or 3nm EUV) for a PS6.

It's my personal opinion that the base PS5 is on 7nm EUV. Now, the benefits of 7nm EUV over 7nm DUV (which is what I suspect the Series X is on) are: 17% density gain, and 10% power consumption reduction OR a 10% performance increase, clock-for-clock. Seeing where the PS5 is landing in regards to not just its specs but things that reinforce the perception of certain specs (such as the system's size and cooling solution), I'd say the PS5 may've only gone for half of the possible performance gain benefit of 7nm EUV, so 5%. Some people probably feel differently...some probably would even say it's not 7nm EUV. But I personally feel that to be the case.

With this taken into consideration, a PS5 Pro would see a pure TF performance increase from 10.275 @ 2.23 GHz...to 11.3025 TF @ 2.23 GHz. This, coming with a 30% power consumption reduction thanks to shifting to the 5nm. While 5nm EUVL would provide an additional 10% power consumption reduction, and 5nm itself brings a 30% power consumption reduction, THAT power consumption reduction comes over basic 7nm, and PS5 is already on 7nm EUV and had a 15% power consumption reduction over that. So overall it would come to a 30% power consumption reduction for them on 5nm EUVL instead of 45%.

So picturing all of that, for a 2023 holiday release? Would look rather tempting doesn't it? So let's summarize (semi-TL;DR):

>YEAR: 2023

>NODE: 5nm EUVL

[CPU]

>GEN: Zen 5 (PC release would be ready by July 2023 going with 16-month timing patterns established so far)​
>CLOCK: 3.5 GHz (PS5 Base compat), 4 GHz (default clock for PS5 Enhanced Mode performance)​
>CORES: 8​
>THREADS: 16​
>CACHE:
>L1$: 128 KB​
>L2$: 512 KB​
**Implements a scaled-down form of Infinity Cache​
>L3$: 8 MB​

[GPU]

>GEN: RDNA 4 (+ some RDNA 5 features); assumes a 16-month period between generations. RDNA 2 launch Nov. 2020, RDNA 3​
launch Mar. 2022, RDNA 4 release July. 2023, RDNA 5 Nov 2024 (PS5 Pro would be one of first RDNA 5-based (in some aspects)​
product on market exclusively for 12 months until PC RDNA 5 cards release in May 2024)​
>CLOCK: 2.23 GHz​
>DESIGN: Chiplet (2x chiplets)​
>CUs: 36 (18 per GPU chiplet)​
>ROPs: 64 (2x 16 ROP blocks per chiplet)​
>ALUs: 2,304​
>FEATURES:
>Unifying GPU Chiplet block (UGC)​
>Unifying Framebuffer Chiplet block (UFC)​
>Improved RT (dedicated RT units built into each Dual CU; the​
RT units are linked with adjacent RT units above and below them,​
to accelerate RT calculations. Basically, graphics data on​
each Dual CU would be broken down to have RT calculations done​
for just the shader data that Dual CU is calculating. MOTL)​
>Improved AI ML (silicon-level support for GPT 2.0 data models,​
though some work would still need to be done on the shaders)​
>Improved image upscaling​
>TF: 11.3025 TF (2.23 GHz clock)​
>POWER CONSUMPTION: 138 watts (average)(106 watts from chiplet design​
+ power consumption reduction, + 32 watts for additional GPU hardware​
silicon (including larger cache sizes); this is a general wattage estimate, I​
have no idea what additional aspects of the GPU would generate what specific​
power usage out of this 32 watt figure, the 32 watts would just be the overall​
upper limit regardless of the combinations.​
>DIE AREA: 94 mm^2 (72 mm^2 from basic die area reduction, + 22 mm^2​
from additional GPU hardware; actual die sizes may be larger since​
determining mm^2 by wattage per mm^2 is not a 100% method, but a​
consistent one for limited estimates)​

[MEMORY]

>RAM: 16 GB GDDR7, as 8x 2 GB, 20 Gbps chips @ 640 GB/s (+ 192 GB/s over base PS5)​
>PERSISTENT RAM: 32 GB low-level, storage-class ReRAM, 25 GB/s​
>STORAGE: 1.536 TB NAND, 5.5 GB/s raw, 11 - 12 GB/s lossless compressed, 17 - 22 GB/s maximum lossy compressed​

[PRICE]

>DIGITAL: $299.99​
>DISC: $399.99​
**Both models will replace their respective base PS5 Digital and PS5 Disc Editions through a gradual phase-​
out shift during 2024​

--------------

For Parts 3 and 4, we'll be focusing on Microsoft, who'll have not one, but two big mid-gen refreshes to deal with...

In the meantime, if you have any ideas of what you'd see happening for a PS5 Pro in 2023, share them below and let's talk about it. It's never too early to talk about new hardware ;)
 
Last edited:

Andodalf

Banned
Love the GDDR7 Call, but I feel they up the amount as well. A persistent memory supplement feels like it would go very underutilized on a mid gen refresh. More System Memory would better serve it IMO, especially as it pushes for higher fidelity RT and assumedly better 8k support.
 

FStubbs

Member
Love the GDDR7 Call, but I feel they up the amount as well. A persistent memory supplement feels like it would go very underutilized on a mid gen refresh. More System Memory would better serve it IMO, especially as it pushes for higher fidelity RT and assumedly better 8k support.

You're right. The push will be for 8K. You'll have people in 2023 posting about how 4K makes their eyes bleed.
 

Silver Wattle

Gold Member
Why would they bother doing all that work and then lump in a 11tflop GPU?

Your analysis, like usual, is overthought.

K.I.S.S;

Process: 5nm
CPU: Zen 3, 8 core @ 4.0Ghz
GPU: 72CU(80 CU die) RDNA 3
RAM: 24GB GDDR6X(two densities)
SSD: doubled flash density for double storage.
 
Personally i wont see a 32 cu setup. I would see an 18/19tf of power dual apu smart shift design running at 64cu and the cpu at around 4/ 4.5ghz with infinity caches posible at 256bus speed x2 and possible 18/20 Gddr6/7 but thats probably gonna be ps6

Ps5 pro is going to be the same the die size will eventually get smaller so same thing possible higher cpu clock speed at 3.8 maybe 4ghz possible the tf will go up since ps4 is 1.84 and ps4 pro is 4.2tf so ps5 is around 10.2 so my guess is a possible 13.2 or 14tf gpu same 16gig of ram and the speed of the gpu may rise from 2.2ghz to 2.5ghz the cu from 32 to 36/38
 
Last edited:

-Arcadia-

Banned
I’m not so sure they’ll do it this time, as much as I want them to.

The thing is that this time, 4K and HDR really warranted a refresh. We were able to get a meaningful jump that people beyond Digital Foundry nerds could notice and appreciate (and even then, some people were like, ‘What’s the difference?’).

Will simple, subtle graphical improvements sell people on a mid-generation $400-$500 refresh? I suspect not. Even the level of improvement afforded by Series X and PS5 to cross-gen games, the best point of comparison for what a mid-gen refresh would be, is not impressing people that much.

It ultimately depends on whether Sony and Microsoft can bring a truly compelling jump to the table. Something like full, unrestrained ray-tracing, in the vein of the Minecraft demo, but without the caveats (visually simple game, 1080p) could be it, but will RT on console technology be that far along by then?
 
Last edited:

tvdaXD

Member
How about releasing the console and keeping the refresh just for making it more power efficient & smaller like the PS3. All these different editions are stupid tbh.
 
We're in the land of diminishing returns when it comes to graphics now, unless Microsoft put out a refreshed XSX which increases the gap and means PS5 is getting worse versions of games then I think this is it until PS6.
 

Rikkori

Member
I don't see it; even the PS4 Pro didn't sell gangbusters and the base console there was so much weaker. These things aren't free to design & make, nor to support, then you're also further impacting your own margins on PS5 because you buy fewer etc
 

Rentahamster

Rodent Whores
The amount of power you're getting right at the beginning of this coming generation is relatively larger than what you got last gen with XBO and PS4. That would imply that there was more of a need for a mid-gen refresh last time than now.

A mid-gen size reduction would make sense.

I don't really see anything truly necessitating a mid-gen pro model unless some crazy new hardware raytracing tech comes out in the next couple of years.
 

regawdless

Banned
If they make that thing more powerful and with that, potentially even bigger... I'll buy a PC case and put the PS5 Pro in that case. What's the difference at that point. Of course, given they stay true to their design philosophy of no taste.
 

Madjako

Member
As 20TF GPUs will already be available in 2020, I'm almost certain sony will put such a GPU in the PS5 pro.
Also keep in mind that the PS5 is looking to be a huge success, so sony will have the dynamic and the money to build a monster machine..
 

anothertech

Member
Holiday 2023, 20-30tf, 32gigs ram, probably same size ssd unfortunately, Roughly 1tb

$499-599

They'll be talking about 8k as if it's a thing

PC guys will laugh at the paltry 30tf gpus while singing praise to their 50tf raytracing rigs

That years iteration of assassin's creed will continue to struggle with steady 60fps.
 
Last edited:

yurqqa

Member
You're right. The push will be for 8K. You'll have people in 2023 posting about how 4K makes their eyes bleed.

The issue with 8K is that there are not enough people who could afford the flat or the house that will fit the reasonable 8K TV.

P.S. Of course, some will just buy 50'' ones and insist they see the difference with 4K, cause they sit 50 cm from the screen.
 

geordiemp

Member
It will be able to play 30 fps more stable. Take it or leave it.

I think you posted in the wrong thread, most games have 60 FPS options go read the Ubisoft and other announcements, only 30 FPS only game is watch dogs for now, and 3090 struggled so whatever.

Back on topic, 72 CU effectively doubling would make more sense, but we need to understand more about RDNA2 later today to have some better thoughts.
 

The_Mike

I cry about SonyGaf from my chair in Redmond, WA
I think you posted in the wrong thread, most games have 60 FPS options go read the Ubisoft and other announcements, only 30 FPS only game is watch dogs for now, and 3090 struggled so whatever.

Back on topic, 72 CU effectively doubling would make more sense, but we need to understand more about RDNA2 later today to have some better thoughts.
I remember the same song this gen. Sony said most games would run 60 fps. Look how that turned out.
 
I think there will be a Pro but in 2024 or even 2025. Shrinking the base console is the more immediate concern. The generation is likely to last until 2028 and still go strong for a few years after that so there’s no need to hurry a Pro console. The main benefit of a Pro will likely be better ray tracing and higher frame rates.
 

Aidah

Member
What major reason would there be for a Pro to exist? Things I can think of is much more raytracing grunt (basically a newer/faster GPU), and possibly AI hardware for a super effective DLSS like solution to get much more out of hardware.

As for "slim", that's pretty much a given.
 
Last edited:

kraspkibble

Permabanned.
i have a feeling they are gonna stick with the silly curves for any Slim/Pro model.

just squish it down into a normal looking console.

What major reason would there be for a Pro to exist? Things I can think of is much more raytracing grunt, possibly AI hardware to be utilized alongside it and/or for better performance in general.

improved raytracing performance is the obvious one. if AMD show off their version of DLSS we could see mid gen consoles targeting 8K just like how we had the PS4 Pro/XB1X target 4K.

they won't be native 8K of course but with AI upscaling games could run at 1440p and they'd be upscaled to 8K. both next gen consoles support 8K 30hz thanks to HDMI 2.1.

the thing is...how many people will have 8K TVs?
 
Last edited:

ZywyPL

Banned
Zen5/6, RDNA4/5, GDDR7, for real?... It's like next-gen speculation thread back in 2018 with all the people mentioning 16-core 4GHz CPU, 32GB HBM3 and 20TF+ GPU all over again, while in reality there is only so much you can pack withing a ~500$ budget.

What's the goal/purpose for the theoretical mid-gen refresh you're trying to build here to begin with? It all starts like you're designing a more efficient, smaller and cheaper Slim model, and then suddenly translates to something that looks like a more advanced, more powerful hardware, where you're throwing a bunch tech from the future that surely won't be available by the time such console would need to be in a fully functional dev-kit state, so just 2-3 years from now, and without any justification what that extra power/features would be needed for. And sorry but you can't create a console that's both small&cheap and at the same time advanced&powerful, that's exactly why we got both Slim and Pro models during this gen, each serving just a single specific purpose.

First and foremost, 5nm process node is necessary for either of the goals. With a Slim model it's all pretty straightforward, you make the same chip, just smaller, with less power consumption, therefore with less generated heat, so that allows to wrap everything up in a smaller case thanks to smaller cooler, while all the other components remain unchanged, because there's no need to, and they cannot get any smaller anyway.

Now with a Pro-like model things aren't that simple, because firstly, what such model should do? Same games at 8K? That could be achieved by some clever upscaling hardware/algorithms instead of raw processing power. Same games at 60/120FPS instead of 30/60? That's when the GPU needs to be twice as fast, no other way around it. Both? Now it get's even more complicated.

But anyway, any changes on the CPU side are a no-go, this would disrupt the logic and would make the developers life a real pain in the ass, we saw no less than 6 current-gen consoles, all equipped with the exact same Jaguar CPU with just slight variation in clock speeds, and that's not without a reason, so Zen2 is here to stay, and it's questionable if it would even need any bump in the clock speed considering it has plenty of processing power for even 120FPS and won't be bottlenecking the system like the Jaguar was, but let's say it could get a slight bump to a 3.7-3.9GHZ.

Secondly, the RAM setup - for higher framerates the same 16GB will do, but higher bandwidth would be welcome, so G6X instead of ordinary G6, at whatever speed is enough to not to bottleneck the GPU, keep in mind the consoles always opt for the lowest speed variants of the memory chips. For 8K however 20-24GB would be needed.

Now the cherry on top, the GPU, and seeing how Sony handled it in PS4 Pro I think it would be logical they'd follow the same philosophy of doubling the CUs, so a monolithic 72CU chip, 2x36CU chiplet design, or 2x36CU X3D packaging come to mind, depending which one turns out to be cheaper in the grand scheme of things, including power consumption and heat. And again, no drastic changes in the architecture, so the same RDNA2, with maybe some updated/new features from RDNA3 that should be finalized by the time they're needed. So double the power = double the resolution, or double the framerate, that's a very save, conservative design approach, so something the consoles are about since 2013. As I mentioned before, instead of more processing power there might be some CBR/AI upscaling being used to target 8K resolution, but it'll all depend on the adoption of the said resolution, whether it's even worth it.

So without going into much details, here's how I see the mid-gen refresh models that might appear in the near future:


PS5 Slim - 5nm PS5 basically, duh, maybe with a fixed clocks instead of SmartShift as I thought about it now, that would be neat.
PS5 Pro - 5nm Zen2 @3.7-3.8GHz, RDNA2/3 @20-24TF, 20-24GB GDDR6X @700-850GB/s, that would be my safest bet for not blowing up the budget, while providing a console than can offer either high resolution or high performance modes compared to base PS5.
 

Aidah

Member
i have a feeling they are gonna stick with the silly curves for any Slim/Pro model.

just squish it down into a normal looking console.



improved raytracing performance is the obvious one. if AMD show off their version of DLSS we could see mid gen consoles targeting 8K just like how we had the PS4 Pro/XB1X target 4K.

they won't be native 8K of course but with AI upscaling games could run at 1440p and they'd be upscaled to 8K. both next gen consoles support 8K 30hz thanks to HDMI 2.1.

the thing is...how many people will have 8K TVs?
Oh no man, please don't even mention 8K.

Just much better ray tracing capability (so basically a newer/faster GPU) alongside a very effective hardware accelerated DLSS like solution to get much more out of the hardware in general. Sticking to 4K.
 
Last edited:

McHuj

Member
I'm not yet convinced that they can pull one off and it's mainly due to the process technology not scaling well enough in terms of power consumption. I don't think a process node drop to 5nm will be enough for Sony to double the compute capacity of the design. They'd have to ship an even bigger console than they are shipping now. I can't see that happening.

Although today's Navi announcement will tell us a lot about the feasibility of it. If the 72CU Navi can really hit +2 GHz, in a reasonable power budget, and with only a 256-bit bus, then there is you PS5Pro GPU (Obviously there would be additional enhancements beyond that).

I could see a very slight spec bump for the Slim version. Maybe with an improved process node, they can run the CPU and GPU at locked clocks. But I don't know if that will even make a difference.
 
Why would they bother doing all that work and then lump in a 11tflop GPU?

Your analysis, like usual, is overthought.

K.I.S.S;

Process: 5nm
CPU: Zen 3, 8 core @ 4.0Ghz
GPU: 72CU(80 CU die) RDNA 3
RAM: 24GB GDDR6X(two densities)
SSD: doubled flash density for double storage.

The problem here would be, with 72 CUs @ even same clocks as PS4 you get 20.55168 TF, and GDDR6X is exclusively between Micron and Nvidia (that's why I chose GDDR7 instead). RDNA 3 will be quite old by 2023 and going from the PS4 Pro Sony tend to go with the latest AMD GPU architecture of the time.

Aside from GPU and RAM (maybe node process? 5nm would be cheaper than 5nm EUVL, but then there's a case of priority) I think we're in agreement.

Personally i wont see a 32 cu setup. I would see an 18/19tf of power dual apu smart shift design running at 64cu and the cpu at around 4/ 4.5ghz with infinity caches posible at 256bus speed x2 and possible 18/20 Gddr6/7 but thats probably gonna be ps6

Ps5 pro is going to be the same the die size will eventually get smaller so same thing possible higher cpu clock speed at 3.8 maybe 4ghz possible the tf will go up since ps4 is 1.84 and ps4 pro is 4.2tf so ps5 is around 10.2 so my guess is a possible 13.2 or 14tf gpu same 16gig of ram and the speed of the gpu may rise from 2.2ghz to 2.5ghz the cu from 32 to 36/38

A 64 CU setup would require Sony to change their BC solution, which is probably the biggest challenge. Notice the PS4 Pro went with 36 CUs, to double the PS4 base's 18. The PS5 went with 36 CUs to keep costs down; if they wanted more CUs their options were either 48 (Cerny even gave this as a hypothetical example in Road to PS5) or 72, they needed multiples of 18 or 36. 64 CUs wouldn't fit that particular design requirement, if they still want to do BC the way they've seemingly established.

Dual APU in particular like, two APUs each with a GPU and CPU? That's an interesting take. I wonder how they would handle the CPU logic in such a case, would the system be treated as a dual CPU configuration since you're technically talking two separate CPU dies across two APUs? For scalability purposes I can see favors for that design, but it might also complicate a lot of things for developers, they'd probably need to copy some of the OS functionality to a core on both CPUs of both APUs, etc. I'm certain they could make it work better than SEGA did with, say, the Saturn, but there's still so many challenges with that in a gaming environment and scale of game production's only seemingly getting bigger.

Wait I just realized you're talking about PS6 here xD. I think you'll really enjoy the PS6 ideas I've got, they definitely incorporate a lot of what you're saying (and a lot more TF power to boot). For the PS5 Pro stuff tho yeah, agreed on clocks; GPU is trickier because to hit 13.2 TF they'd have to give up power consumption savings they'd get shifting to a smaller node. Realistically I guess they could do a 2-model approach; something like I've mentioned in the OP as a "PS5 Slim", and then take that same setup but be more liberal with power usage to hit some high-13/14 TF figure going with higher clock (2.8 GHz...granted if RDNA 3 or RDNA 4 can even hit that high) and get a performance boost on 5nm EUVL, that could be your "PS5 Pro".

Give that Pro model 1.536 TB storage, it'd need a much bigger PSU though because total system TDP will probably push a good deal north of 250 watts at that point.

Will simple, subtle graphical improvements sell people on a mid-generation $400-$500 refresh? I suspect not. Even the level of improvement afforded by Series X and PS5 to cross-gen games, the best point of comparison for what a mid-gen refresh would be, is not impressing people that much.

I think we need to give devs more time on this part. The biggest challenge towards more visually impressive games isn't tech anymore, it's the time/resources/budgets required to form big enough/talented enough teams to make the games themselves. That's why console cycles have been getting longer. Just as an example we went an entire gen without a real "new" GTA, that's how crazy it's getting now.

For me personally there's some games coming for next-gen giving that sort of strong impression, though. R&C Rift Apart, Demon's Souls Remake, FS 2020 (once it's on Series X), Horizon: Forbidden West, Exo-Mecha (if they can hit the fluidity of the trailer shown), Bright Memory Infinite (same as Exo-Mecha) etc. But I'm probably easier to impress with a graphical leap than some others here.

We should consider how the casual/mainstream types, the ones who make up the vast majority of those base PS4/XBO install bases and don't have high-end PC rigs, will interpret the next-gen visual jump, though. For them it's going to feel like night-and-day.

As 20TF GPUs will already be available in 2020, I'm almost certain sony will put such a GPU in the PS5 pro.
Also keep in mind that the PS5 is looking to be a huge success, so sony will have the dynamic and the money to build a monster machine..

They could, but that doesn't mean they will. I think there are market logistical reasons to look at to ask why a PS5 Pro with that level of GPU performance would be needed, and how well would it realistically sell in relation to base PS5. Would it even hit PS4 Pro sales ratios in such a case?

My whole idea behind not just mid-gen refreshes but the 10th-gen systems is "smarter over stronger". Not that you won't see power increases; that'll naturally happen and be noticeable. But (and this is probably just my personal opinion), at some point we're probably going to hit a wall where the performance gains ratio relative to the power consumption needed to hit them, by just pushing up the raw power over everything else, is going to be TERRIBLE and lead to stagnation.

That's what almost happened with CPUs in the mid-2000s; seemed like everyone's idea of more power came down to over-reliance on node shrinks and pushing clocks higher and higher (the latter part is something I think AMD need to be wary of in particular when it comes to pushing their GPUs further; higher clocks on their own aren't a full-proofed long-term strategy for product advances IMHO). They realized something had to change in order to get big changes, that's when multi-core processors finally went mainstream. We've only seen clock rates for CPUs return to the highs of what we saw in the early 2000s maybe 2-3 years ago, as multi-core CPU design standards have, well, standardized, and the room has been around to favor higher clocks...but that won't last forever.
 

Audiophile

Gold Member
The best option imo would be a small spec bump for the APU (primarily via clocks on 3nm) but with some sort of AMD-variant of Tensor Cores being the main addition for Upscaling and RT Denoising. Could get games to 8K for either 8K displays or for supersampling on 4K, plus better RT quality. Would provide the most bang for buck.

Alongside that a small bump in RAM quantity and relative bandwidth.

Finally, 2x the SSD capacity and a smaller chassis.


AMD 3nm APU: AMD Zen 2 CPU 8C/16T @ 3.8GHz + RDNA4(?) GPU 36CU @ 2.5GHz + (lots of) Hardware Tensor Cores. 20GB GDDR6 320-Bit @ 640GB/s, 1650GB SSD, Smaller Chassis.

One thing I would really like to see however, is something like a doubling of the RT Intersection Engines in the TMU, so the Pro console can crank out much higher quality RT.


So yeah, basically throw everything at better RT & DLSS-like functionality. Only have a small general compute bump and then use the rest of the node-shrink headroom to provide a smaller, still-quiet and cheaper box with more storage.
 
Last edited:

RaySoft

Member
If anything Sony will just make revision changes to bring down cost and wattage/heat. (like all slim models before) They won't breakoff their platform into two different SKU tiers ala PS4Pro. Unlike with the PS4, ther is no need to this gen. The bump would be for mainstream 8K, and that ain't happening.
 
Last edited:

Marlenus

Member
Don't see the point in having a cpu, 2x gpu and io die based chiplet system. Packaging costs would make it cheaper to stick with a monolithic design.

If PS5 and series X can handle 4k at good framerates I don't think we will see a pro refresh and we will just get slim versions. They might move to zen 3/4/5 and they might update to RDNA 3/4 but outside of that I don't see much.
 
Zen5/6, RDNA4/5, GDDR7, for real?... It's like next-gen speculation thread back in 2018 with all the people mentioning 16-core 4GHz CPU, 32GB HBM3 and 20TF+ GPU all over again, while in reality there is only so much you can pack withing a ~500$ budget.

It's not really too crazy to think about Zen 5/6, RDNA 4/5 or even GDDR7 actually. For the Zen and RDNA timelines I took the time between Zen 2 and Zen 3, and RDNA 1 to RDNA 2, which is seemingly set to be 16 months (so, I have to update the OP a bit). You get RDNA 3 by March 2022 going by that timeline, and RDNA 4 by July 2023. RDNA 5 would be ready by November 2024. Likewise, Zen 3 (like RDNA 2) will be hitting this November, so Zen 4 could be ready March 2022, Zen 5 by July 2023, and Zen 6 by November 2024.

Andodalf Andodalf is the one who convinced me to take a look at GDDR7 specifications; there aren't many but it seems it'll be ready by 2022. GDDR6X is off-limits because that's a Micron/Nvidia exclusive. HBM2E will be too pricey at a premium compared to relative bandwidth gains of GDDR7, and GDDR6 might start falling behind in terms of what bandwidth options it can provide for even very slight bump-ups in spec. Not only that, but it might also increase in terms of production costs as fabs begin to make a mass switch to the GDDR7 spec. That's the main reason I included it here.

What's the goal/purpose for the theoretical mid-gen refresh you're trying to build here to begin with? It all starts like you're designing a more efficient, smaller and cheaper Slim model, and then suddenly translates to something that looks like a more advanced, more powerful hardware, where you're throwing a bunch tech from the future that surely won't be available by the time such console would need to be in a fully functional dev-kit state, so just 2-3 years from now, and without any justification what that extra power/features would be needed for. And sorry but you can't create a console that's both small&cheap and at the same time advanced&powerful, that's exactly why we got both Slim and Pro models during this gen, each serving just a single specific purpose.

Well, there's a few things to keep in mind here. The spec bump in terms of actual TF performance is not all that much to be honest, just about an extra TF. By comparison the PS4 Pro was almost 2.3x the TF boost over the base PS4. So a more "expected" PS5 Pro following that would be closer to 23.45 TF (or 25.795 TF), not what I've listed in the OP. In fact, the spec bump for the PS5 Pro I have in OP only comes from the performance gain moving to 5nm keeping the same clocks, nothing else.

The basis I had in mind for these mid-gen refreshes would be to reduce the size a decent deal, give a boost to storage capacities as NAND prices fall, give a slight boost to memory bandwidth while keeping same BOM on volatile memory as the base systems (when you adjust for inflation), and yes, act as a testing grounds for some technological designs that could be further refined and increased on 10th-gen systems, kind of the way we saw for PS4 Pro and Xbox One X doing similar for the PS5 and Series systems. Sony actually has a few ReRAM designs in testing phases right now; by the time of a PS6 I do see ReRAM being on the market and rather situated at that point. Having a small block of it in a PS5 Pro (as long as they keep the production in-house, which therein saves on costs and means no need to pay royalties when they own rights to the tech) could serve as testing grounds for a PS6 design, and benefit data I/O operations even further than what NAND-based solutions can provide.

There's an argument that it could be a tad too disruptive though, and it really hinges on if Sony can get commercial ReRAM to the market by at least 2021; early 2022 by the latest, otherwise yeah it won't be happening for any mid-gen refresh.

First and foremost, 5nm process node is necessary for either of the goals. With a Slim model it's all pretty straightforward, you make the same chip, just smaller, with less power consumption, therefore with less generated heat, so that allows to wrap everything up in a smaller case thanks to smaller cooler, while all the other components remain unchanged, because there's no need to, and they cannot get any smaller anyway.

Agreed with all of this. 5nm EUVL, in fact, is what I specifically mention for the node processes, for all of the reasons you bring up.

Now with a Pro-like model things aren't that simple, because firstly, what such model should do? Same games at 8K? That could be achieved by some clever upscaling hardware/algorithms instead of raw processing power. Same games at 60/120FPS instead of 30/60? That's when the GPU needs to be twice as fast, no other way around it. Both? Now it get's even more complicated.

But anyway, any changes on the CPU side are a no-go, this would disrupt the logic and would make the developers life a real pain in the ass, we saw no less than 6 current-gen consoles, all equipped with the exact same Jaguar CPU with just slight variation in clock speeds, and that's not without a reason, so Zen2 is here to stay, and it's questionable if it would even need any bump in the clock speed considering it has plenty of processing power for even 120FPS and won't be bottlenecking the system like the Jaguar was, but let's say it could get a slight bump to a 3.7-3.9GHZ.

8K wasn't particularly a focus with what I envision for a PS5 mid-gen refresh, but I absolutely understand the importance of AI image upscaling to achieve it. That's one reason why simply doubling the TF power was not a concern for my take on it.

You might have a point with the Zen CPU upgrade; at the same time, if compatibility is maintained between newer Zen processors and older ones (even if Sony and Microsoft would have to add that compatibility back in), I don't necessarily see the fuss in moving to a newer Zen processor setup. Not talking changes in core or thread counts here, those would stay 8C/16T. But a few newer features in the CPU side for enhanced games that want to use them? I don't necessarily see the harm there especially if those are optional. I agree with the clock range, though it could likely hit 4 GHz but that would be the upper limit.

Secondly, the RAM setup - for higher framerates the same 16GB will do, but higher bandwidth would be welcome, so G6X instead of ordinary G6, at whatever speed is enough to not to bottleneck the GPU, keep in mind the consoles always opt for the lowest speed variants of the memory chips. For 8K however 20-24GB would be needed.

GDDR6X won't be possible; to my knowledge that's something exclusive between Micron and Nvidia. Now, Micron aren't the only ones who make GDDR6; SK Hynix and Samsung are the other out of that Big Three. But the specifications for GDDR6X seem like something Nvidia brought to the table to boost the spec. Even if Micron license out GDDR6X for others, Sony and Microsoft would probably need to pay a royalty to Nvidia for it on top of whatever they're paying Micron.

Probably not worth the costs considering, IMO, GDDR7 would likely be ready by then and see a decent spec bump in pin I/O bandwidth over GDDR6, enough to be competitive with GDDR6X @ 18 Gbps, and beat it @ 20 Gbps. You might be right though in terms of them optioning out for the lower speed variants, so maybe 16 Gbps GDDR7...then again, Sony did design the PS5 with support for faster GDDR6, they simply weren't able to secure it at a price they felt feasible. So it'd seem at least on some level, they are willing to go for faster speed memory if the price is right.

Now the cherry on top, the GPU, and seeing how Sony handled it in PS4 Pro I think it would be logical they'd follow the same philosophy of doubling the CUs, so a monolithic 72CU chip, 2x36CU chiplet design, or 2x36CU X3D packaging come to mind, depending which one turns out to be cheaper in the grand scheme of things, including power consumption and heat. And again, no drastic changes in the architecture, so the same RDNA2, with maybe some updated/new features from RDNA3 that should be finalized by the time they're needed. So double the power = double the resolution, or double the framerate, that's a very save, conservative design approach, so something the consoles are about since 2013. As I mentioned before, instead of more processing power there might be some CBR/AI upscaling being used to target 8K resolution, but it'll all depend on the adoption of the said resolution, whether it's even worth it.

So there's things here I can see working, and other things maybe not. On the basic GPU side, I honestly don't think they'll do 72 CUs. That's part of the reason I went with sticking to a 36 CU design but implementing it as a chiplet. Depending on how RDNA 3 specs out to even higher clocks, you could still hit north of 14 TF on a 36 CU design on 5nm EUVL and staying in a decent system TDP probably a bit north of 250. Then simply scale the clocks down to bring the power usage down, bring the TF amounts down, and that could be a lower-end version to replace the base PS5 at the entry model end. Going by MS's notes on node costs from Hot Chips, the prices are increasing, not decreasing, so even on 5nm if you take that and now you're doubling the CUs again, does that work out for Sony on the pricing end? It doesn't seem like they'd be able to hit even $499 with such a setup, let alone $399.

I also don't know about them sticking to RDNA 2; both PS4 Pro and Xbox One X shifted to GCN 4.0-based architectures, quite a ways removed (in terms of overall capability and such) from their base models. Compatibility was still perfectly maintained. I don't see a reason for either to stick to RDNA 2 or even RDNA 3, but my own timeline for future RDNA architecture releases might differ from your own. I agree on CBR/AI upscaling though; if it's able to be implemented in a way that is mostly "automatic" as a system process instead of needing to be explicitly programmed against on a game-by-game basis ala current DLSS, then it becomes a big focal point for sure.

So without going into much details, here's how I see the mid-gen refresh models that might appear in the near future:


PS5 Slim - 5nm PS5 basically, duh, maybe with a fixed clocks instead of SmartShift as I thought about it now, that would be neat.
PS5 Pro - 5nm Zen2 @3.7-3.8GHz, RDNA2/3 @20-24TF, 20-24GB GDDR6X @700-850GB/s, that would be my safest bet for not blowing up the budget, while providing a console than can offer either high resolution or high performance modes compared to base PS5.

I think we're a lot in agreement if you think about it, in terms of some general targets. We both see 5nm (I personally see 5nm EUVL though, further power consumption savings), we roughly see the same CPU clocks, roughly same on CBR/AI image upscaling and the such.

However, I don't see them using GDDR6X for reasons mentioned above, or increasing the physical memory by any measure (PS4 Pro also kept same physical memory amount as base PS4), sticking with Zen 2 (future Zen archs should be BC with Zen 2, and the new features being optional without interfering with current game logic models), RDNA 2 (same reasons, plus past precedent with the PS4 Pro and Xbox One X), or system memory bandwidth amounts (but not terribly far off from your lowest end estimate).

But this might all come down to some differences we have in Sony's vision for a mid-gen refresh. I think they'd try to kill two birds with one stone, so you'd get a mid-gen refresh somewhere as an inspiration of the PS4 Pro but also something like the PS One from way back which, IIRC, did have a very slight spec bump over the base PS1. Whereas PS4 Pro's boosts came from more raw power, I think a PS5 Pro would get its benefits more from complementary emerging technologies with some presence in the commercial market at that point, though my thoughts on things like ReRAM could be seen as overly optimistic.

In a way you can still get a solid PS5 Slim and even PS5 Pro (in the spirit of how some of you are envisioning it) with what I've laid out previously. Maybe cut out the ReRAM, keep it a 36 CU chiplet design, 5nm EUVL, same base GPU clocks, MAYBE keep CPU explicitly Zen 2 and GPU explicitly RDNA 2, but take some features from future Zen/RDNA designs. 16 GB GDDR7, 640 GB/s, 1.536 TB SSD at base PS5 bandwidths, and push that for $299 in a much smaller form factor, no disc drive. Maybe it could even hit $249.

Then, take that same design, but spec up the Zen and RDNA components to more contemporary archs that would be available at that time (so, at least Zen 4, and something like RDNA 4), but still being fully compatible with base PS5 code. Same 36 CU chiplet design, 5nm EUVL, but now you can bump the clock up to something more like 2.8 GHz or maybe even higher, depending on how far AMD can push GPU clocks going forward (I personally think they'll hit a wall there at some point, but not before this time). Push TF performance up to 14- 15 TF, CPU that can keep up. Beefier Infinity Cache implementation, but amount and bandwidth of main memory stays the same (16 GB, 640 GB/s).

Storage speed increased to something around 12 GB/s raw bandwidth, so the SSD I/O decompression hardware has to be increased to support up to 48 GB/s maximum lossy compression range. 1.536 TB SSD, this one would need to be larger form factor, price probably between $399 to $499. Personally don't think the BOM of such a setup would warrant $499 out of necessity, but it could be a a price providing profit range Sony'd want, and something to consider if Microsoft are pushing their own Series X mid-gen refresh a year later (that'd give Sony room to price this PS5 Pro higher).
 

LucidFlux

Member
Both camps are already in the initial stages of planning future hardware. When it releases will largely depend on both the market and their hardware partner's roadmaps (AMD).

I don't see the market demanding 8k level performance in just a few years so that won't be the driving factor for a refresh, plus the boxes already tick that 8k marketing box.

That said I do expect a shorter generation of 5-6 years. What the specs will look like is up to the engineers but with the fires AMD has lit under Intel and Nvidia's asses, I expect huge advancements across the board by the following generation.
 

Genx3

Member
If Sony decides to make a PS5 Pro it will likely be the same HW on a smaller node just twice as Powerful.
I would expect a PS5 Pro to be around 20 TF's of GPU on 5 nm design with an upclocked CPU. This could sell for $600 and easily be worth every penny in 2023. At $500 its a no brainer.
 
Finally, the GPU. As mentioned before, no 72 CU GPU design here; while a chiplet approach would be supported, we'd see it as 2x 18 CU chiplet blocks. Process-wise, while 3nm (perhaps even 3nm EUV) would be readily available by this time in a general sense, DO keep in mind that costs are NOT scaling down with node shrinks; rather, the opposite is happening i.e prices are INCREASING. With investments already placed in on the ReRAM and (very likely) customizations to any aforementioned features of the GPU chiplet design that don't end up being standard in the RDNA spec by this point, to keep costs down and place investments in other areas Sony would likely go for 5nm EUVL instead, saving 3nm (or 3nm EUV) for a PS6.

It's my personal opinion that the base PS5 is on 7nm EUV. Now, the benefits of 7nm EUV over 7nm DUV (which is what I suspect the Series X is on) are: 17% density gain, and 10% power consumption reduction OR a 10% performance increase, clock-for-clock. Seeing where the PS5 is landing in regards to not just its specs but things that reinforce the perception of certain specs (such as the system's size and cooling solution), I'd say the PS5 may've only gone for half of the possible performance gain benefit of 7nm EUV, so 5%. Some people probably feel differently...some probably would even say it's not 7nm EUV. But I personally feel that to be the case.

With this taken into consideration, a PS5 Pro would see a pure TF performance increase from 10.275 @ 2.23 GHz...to 11.3025 TF @ 2.23 GHz. This, coming with a 30% power consumption reduction thanks to shifting to the 5nm. While 5nm EUVL would provide an additional 10% power consumption reduction, and 5nm itself brings a 30% power consumption reduction, THAT power consumption reduction comes over basic 7nm, and PS5 is already on 7nm EUV and had a 15% power consumption reduction over that. So overall it would come to a 30% power consumption reduction for them on 5nm EUVL instead of 45%.

Not sure where you're getting 11.3TF @2.23GHz from - thats not how FLOPs are calculated. Its fairly simple mathematics.
36CU = 2304 shaders (64 shaders per Compute unit) = 4608 fused multiply-add floating point operations per clock (2 operations per shader).
Multiply 4608 by 2.23GHz, and you get 10,275 GFLOPS of compute or 10.28 TFLOPS

In order to get 11.3TF, you would 2535 shaders which is not divisible by 64 and therefore not a viable configuration. The only alternative would be 2560 shaders (40CU), which at 2.23GHz would net you 11.4 TFLOPS.

TeraFLOPS is simply a metric to measure how many floating point operations can be performed by a GPU per second.

Also, I doubt Sony would make a PS5 Pro thats barely 10% faster than the base PS5. If they're going to be using chiplets, they could easily configure a set with 3x 20CU chiplets for a total of 60CU. 2 CU disabled per chiplet for a total of 54 CU active.
This nets you 3456 active shaders. 5nm should allow for higher clockspeeds than 7nm, but lets not go too crazy for a power-limited console - say 2.3GHz.
This would net you 15.9 TFLOPS, which is perfectly acceptable for a mid-generation refresh. The great thing about this approach is that 1 chiplet can be kept active for PS4 back compat, 2 chiplets for PS4 Pro and PS5 base back compat, and all 3 chiplets for PS5 Pro boost mode.

Will they go for a chiplet approach for a hypothetical PS5 Pro? I don't know. The benefit of chiplets is that the dies will be smaller and thus more manufacturable, which means the yields will be much higher as you can churn out considerably more of them on a given wafer. On 5nm, 20CU chiplets would be positively tiny. The biggest die will undoubtedly be the I/O complex die. Which they can of course make a nice big one for the PS5 Pro, and then cut it down based on bins for a PS5 Slim refresh with just 2 chiplets as needed. HBMX for PS5 Pro is a resounding no. Probably will keep 16GB of RAM, or will at most add a secondary separate pool of DDR5 RAM along with a co-processor for the OS to free up more of the VRAM for gaming.
 
If Sony decides to make a PS5 Pro it will likely be the same HW on a smaller node just twice as Powerful.
I would expect a PS5 Pro to be around 20 TF's of GPU on 5 nm design with an upclocked CPU. This could sell for $600 and easily be worth every penny in 2023. At $500 its a no brainer.
Watching a few different spots it sounds like people at Sony really DON'T want to make a PS5 Pro, and I can see some understanding for that. I still think 20 TF for such is kind of crazy but seeing there's now leaks for 4080 Ti cards with 66 freaking TFLOPs, maybe it wouldn't be so ridiculous for a 2023 mid-gen refresh.

Though if upscaling technology dramatically improves, I think Sony and/or Microsoft would rather take that extra GPU budget and put it towards other things. If that means even more software or more dev-side advancements (maybe more AI coding/asset creation R&D for training models to dramatically expediate dev time processes and cut down dev costs), I'd be all for it.
 

kyliethicc

Member
PlayStation 5 Pro

$400 (digital) / $500 (disc)
launch November 2023

custom AMD SoC (TSMC 5nm)

8 cores 16 threads @ ~ 4.0 GHz
36 CUs @ ~ 2.5 GHz

16 GB GDDR6 (@ 18 Gb/s) + 1 GB DDR4
576 GB/s bandwidth

1.65 TB SSD, PCIe Gen4x4, 1 GB LPDDR4
5.5 GB/s raw read

400 W internal PSU, ~ 200 W max draw

Front ports
1x USB-C (20 Gb/s), 1x USB-A 2.0

Rear ports
AC inlet, HDMI 2.1 Out, LAN, 1x USB-C (20 Gb/s), 2x USB-A (10 Gb/s)

WiFi 7, Bluetooth 6, M.2 expansion bay, (UHDBD drive)



Just take the PS5, keep it at the same price, size, and power draw, while basically just clocking the same chip faster. (Using a die shrink.) Then add 1 GB of DRAM to give game devs 1 GB more of the 16 GB G6 to use for games. Double the SSD size, add the latest WiFi and Bluetooth, faster USBs etc for quality of life.
 
Last edited:
PlayStation 5 Pro

$400 (digital) / $500 (disc)
launch November 2023

custom AMD SoC (TSMC 5nm)

8 cores 16 threads @ ~ 4.0 GHz
36 CUs @ ~ 2.5 GHz

16 GB GDDR6 (@ 18 Gb/s) + 1 GB DDR4
576 GB/s bandwidth

1.65 TB SSD, PCIe Gen4x4, 1 GB LPDDR4
5.5 GB/s raw read

400 W internal PSU, ~ 200 W max draw

Front ports
1x USB-C (20 Gb/s), 1x USB-A 2.0

Rear ports
AC inlet, HDMI 2.1 Out, LAN, 1x USB-C (20 Gb/s), 2x USB-A (10 Gb/s)

WiFi 7, Bluetooth 6, M.2 expansion bay, (UHDBD drive)



Just take the PS5, keep it at the same price, size, and power draw, while basically just clocking the same chip faster. (Using a die shrink.) Then add 1 GB of DRAM to give game devs 1 GB more of the 16 GB G6 to use for games. Double the SSD size, add the latest WiFi and Bluetooth, faster USBs etc for quality of life.
If they keep it at the same power draw hopefully they'd still be able to shrink the size of the system because that's one of the few legitimate criticisms that can be had with the hardware. I figure you're including 1 GB DDR4 for background OS tasks, right? I dunno if they'd go for DDR4 instead of LPDDR4, because if the latter is already being used for SSD cache then they can save on costs by ordering double that and splitting the use between SSD cache and background OS RAM, due to economies of scale.

Also kinda up in the air if they want to move to yet another Wifi, tho for QoL as you said it would probably make sense as long as it isn't too costly. I think WiFi 7 will be leveraged a lot more in terms of how they look at specializing it for I/O purposes, for PS6, because I think by then VR and AR will finally be pretty much ready for the mainstream and be cost-friendly (and design-friendly, in terms of headset slim factor/weight/power-performance balance, also more inviting social-wise) enough to push as a genuine standard value proposition add for home console gaming. Plus it would give 10th-gen systems something to sell beyond yet higher resolutions or compute performance.
 

kyliethicc

Member
If they keep it at the same power draw hopefully they'd still be able to shrink the size of the system because that's one of the few legitimate criticisms that can be had with the hardware. I figure you're including 1 GB DDR4 for background OS tasks, right? I dunno if they'd go for DDR4 instead of LPDDR4, because if the latter is already being used for SSD cache then they can save on costs by ordering double that and splitting the use between SSD cache and background OS RAM, due to economies of scale.

Also kinda up in the air if they want to move to yet another Wifi, tho for QoL as you said it would probably make sense as long as it isn't too costly. I think WiFi 7 will be leveraged a lot more in terms of how they look at specializing it for I/O purposes, for PS6, because I think by then VR and AR will finally be pretty much ready for the mainstream and be cost-friendly (and design-friendly, in terms of headset slim factor/weight/power-performance balance, also more inviting social-wise) enough to push as a genuine standard value proposition add for home console gaming. Plus it would give 10th-gen systems something to sell beyond yet higher resolutions or compute performance.
Yeah maybe LPDDR4 would make more sense. They do have 512 MB of that for the SSD already. But if the SSD gets doubled to 1.65 TB in size, Sony might want to double that cache as well to 1 GB of LPDDR4. So maybe they'd need to add 1 GB extra for OS, so 2 GB total.

Its just to free up more of the G6 for devs. They did that with PS4 Pro. They went from 256 MB DDR3 on launch PS4 to 1 GB DDR3 on PS4 Pro that was used for the OS and they gave devs more of the G5 for games.

As for the Wifi, I just assume it will progress over time. Sony uses the latest spec most of the time.

PS4 - WiFi 4 (802.11n), Bluetooth 2.1
PS4 Pro - WiFi 5 (802.11ac), Bluetooth 4.0
PS5 - WiFi 6 (802.11ax), Bluetooth 5.1
so
PS5 Pro - WiFi 7 (802.11be), Bluetooth 6
PS6 - WiFi 8 (802.11?), Bluetooth 7

Just seems likely. Unless the WiFi standards aren't ready in time.
 
Top Bottom