AncientOrigin
Member
If it were 13 they shouldn’t even release this would be ridiculous.No one cares if it has a bit better Ray Tracing if the TF would be almost the same.Would be an absolute useless console.
-
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.
Rumor: PS5 Pro Codenamed "Trinity" targeting Late 2024 (some alleged specs leaked)
- Thread starter MisterXDTV
- Start date
- Rumor Platform
AncientOrigin
Member
The Navi 31 multi-chip module features 58 billion transistors, a 165% increase in transistor density than the previous generation Navi 2x, across seven dies: one Graphics Compute Die (GCD) and six Memory Cache Dies (MCD). The full Navi 31 die contained 12,288 FP32 cores, equivalent to 6144 Stream Processors.[7]Reportedly, the Navi 31 die has been designed to scale up to 3.0 GHz frequency, though AMD's Radeon RX 7900 XTX reference design can hit a boost frequency of 2.5 GHz.[8] The Navi 31 die is fabricated on TSMC's N5process node.
From the Wikipedia page.AMD wanted the 7000 series to be very high clocked and lots of rumors before the announcement said that even some cards could hit 4GHz.
But something went wrong and they had to lower the speeds like 2,5 or lower.Now after so much time and new nodes and improved RDNA 3-4 i think The AMD card for the PS5 pro will be high clocked.Especially on newer node should be very easy to accomplish.My guess would be at least 2.6 maybe even 2.7-2.8 GHz
From the Wikipedia page.AMD wanted the 7000 series to be very high clocked and lots of rumors before the announcement said that even some cards could hit 4GHz.
But something went wrong and they had to lower the speeds like 2,5 or lower.Now after so much time and new nodes and improved RDNA 3-4 i think The AMD card for the PS5 pro will be high clocked.Especially on newer node should be very easy to accomplish.My guess would be at least 2.6 maybe even 2.7-2.8 GHz
Mr.Phoenix
Member
Doesnt matter if you are on 7nmor 4nm.. if you are using (eg) 200W, you have to cool for 200W. The PS5 is as big as it is, because a 200W+ cooler takes up a ot of space. Veen the Series X is big too, its just like 0.4 liters less in volume than the PS5.
If their new APU, somehow draws over 200W, then they would be using a similarly sized cooler and thus, case, unless they want it sounding like a jet engine like the PS4pro. So expect t tobe at least similar in volume to the PS5.
And yeah, they can redesign it, and about it being aesthetically pleasing or not, that's always subjective.
Mr.Phoenix
Member
wtf is wrong with that guy.. how stupid can someone be?
not only is it so bad that if you heard what he said you would think the PS5pro is replacing the PS5... the guy is actually championing that nonsense MS said at the start of the gen that the Series X (launched along with the Series S mind you), is their mid gen refresh.
DeepEnigma
Gold Member
Caio
Member
Maybe I will say something stupid, but shouldn't the PS5 PRO be based on a Radeon RX 7800 XT ? This PC card will be available in 2023, right ? PS5 PRO is supposed to be released within Nov 2024, so a custom APU based on that PC GPU is very possible. Actually, the specs of the RX 7800 XT seems a bit too high to fit into a Console, but it could be a very custom version of it.
ChiefDada
Member
Not true. If the PS5 Pro had 0% raster uptick but implemented hardware tech discussed below, that alone would be a game changer.
Mr.Phoenix
Member
We have said this in this thread.
Yes, if the rumors are true, then its pretty much a no-brainer that the PS5pro GPU is based on the 7800XT. Albeit, a more accurate depiction, would be to say the PS5pro is is mostly identical to the 7700XT.
And no, it's not too high to fit in a console, especially when you consider that some of what makes that GPU won't necessarily make it into the console. The key thing though is that it has GCD (that part that would make it into a console), that measures around 200mm2.
But for some context, the XSX, on a 7nm process, managed to fit a 56CU GPU along with a CPU into a 350mm2 chip. Do you really think that on a 4nm process, they couldn't fit a 60CU chip with a CPU onto a 300-350mm2 chip?
Its just wild to me that:
They think pro consoles are replacing base consoles.
If its not necessary, thats for Sony to worry about if no one buys one.
The XSX really is the mid gen refresh for Xbox?? Oh boy....who in their right mind would think that would fly when tech is constantly advancing, moving forward.
....the PS5 and Series consoles didnt start development in 2020.....
Good god...
So...if the XSX is the mid gen refresh...the Series S...is the base console?
Last edited:
MisterXDTV
Member
Tqaulity
Member
Maybe I missed it but the one area not really discussed here that is important to look to get an idea of what's coming to future PlayStation consoles is to look at AMD's mobile roadmap. Sony in particular has pretty much taken the top end mobile GPUs of the day and used it as the basis for both the PS4 and PS4 Pro (Xbox GPUs have been more customized and diverged a bit from the PC parts). The PS4 was effectively a slightly downclocked 7970M card and the PS4 Pro was a slightly downclocked RX 580M. This makes sense because A) the mobile cards fit a console form factor much better than any desktop equivalents, in terms of size and TDP B) The top end mobile cards are smaller and less powerful than the top end desktop parts which matches the perf profile of consoles.
So what do we know about AMD's high end 7000 series mobile roadmap. Not much officially but we do have this "leak" which points to a top end 7800XT/7900S mobile sku that both feature 60 CUs (SOURCE). This is important because it points to the fact that AMD is capable of building a 60 CU card in a mobile/console form factor pretty soon. TDP is between 135-180W, much lower than the desktop 7800XT card and much more in line with a console spec. (This also is why any talk of a 72CU part is just silly). When you consider that the current top spec Radeon mobile GPU is the 6800XT/6850XT which is still only a 40CU part, this jump to 60CU is quite significant for the mobile space (50% more CUs + Higher Clocks = ~2x improvement over last gen). Also note that the 7800 mobile is the 54CU derivative which is likely for the PS5 Pro as we've already discussed.
So no, a 7800XT derived card is not too big for a console and AMD already has it scaled down for mobile. Also worth noting that the PRO console yields don't have to be as high as a brand new base console launch. If the 7800XT is coming out this year, then process should be pretty mature by end of 2024. I wouldn't completely rule out seeing the full 60 CU part, although I too suspect it may be 54 CU just due to cost and cooling.
Last edited:
Mr.Phoenix
Member
I have completely resigned myself to accepting DF is just obtuse. That has to be the only explanation. It cannot even be that they have an agenda because if that was it, there are better ways to make a case for not needing a pro. So no, this is just blatant stupidity.
I like their tech stuff, but maybe they should just stick to pixel counting. Any media outlet that has such strong feelings on anything and cat stop itself from pushing those feelings onto their audience, is not a news/info site anymore, They are a propaganda site. Just calling it as it is.
Physiognomonics
Member
This perfectly explains their behavior since years ago and since... XB360 vs PS3. They simply are activists. They are not here to do what they are supposed to do: analyse objectively gaming technologies for their viewers. They don't care about the intelligence of their viewers. They are here to push an agenda to them. Indoctrinate them. Why? Who knows.
Mr.Phoenix
Member
Maybe the joke is on us, to think that for some reason, DF won't have fanboys too.They are super subtle about it, (some of them at least) but they are there.
Caio
Member
Which could be the specs of a RX 7800M ? The RX 7800 XT can perform FP32, 38.71 TFLOPS. Isn't safe to assume that PS5 PRO could be easy a RDNA 3.5, 24 TF machine by the end of 2024 ?Maybe I missed it but the one area not really discussed here that is important to look to get an idea of what's coming to future PlayStation consoles is to look at AMD's mobile roadmap. Sony in particular has pretty much taken the top end mobile GPUs of the day and used it as the basis for both the PS4 and PS4 Pro (Xbox GPUs have been more customized and diverged a bit from the PC parts). The PS4 was effectively a slightly downclocked 7970M card and the PS4 Pro was a slightly downclocked RX 580M. This makes sense because A) the mobile cards fit a console form factor much better than any desktop equivalents, in terms of size and TDP B) The top end mobile cards are smaller and less powerful than the top end desktop parts which matches the perf profile of consoles.
So what do we know about AMD's high end 7000 series mobile roadmap. Not much officially but we do have this "leak" which points to a top end 7800XT/7900S mobile sku that both feature 60 CUs (SOURCE). This is important because it points to the fact that AMD is capable of building a 60 CU card in a mobile/console form factor pretty soon. TDP is between 135-180W, much lower than the desktop 7800XT card and much more in line with a console spec. (This also is why any talk of a 72CU part is just silly). When you consider that the current top spec Radeon mobile GPU is the 6800XT/6850XT which is still only a 40CU part, this jump to 60CU is quite significant for the mobile space (50% more CUs + Higher Clocks = ~2x improvement over last gen). Also note that the 7800 mobile is the 54CU derivative which is likely for the PS5 Pro as we've already discussed.
So no, a 7800XT derived card is not too big for a console and AMD already has it scaled down for mobile. Also worth noting that the PRO console yields don't have to be as high as a brand new base console launch. If the 7800XT is coming out this year, then process should be pretty mature by end of 2024. I wouldn't completely rule out seeing the full 60 CU part, although I too suspect it may be 54 CU just due to cost and cooling.
CanYouFlyBobby?
Banned
A bit offtopic perhaps but will AMD always use 64 shaders per CU going forward?
Loxus
Member
If you mean SIMD32?
Yes, RDNA4 is rumored to continue with 4 SIMD32 per CU.
It's also called Dual Issue Wave32 Units.
CanYouFlyBobby?
Banned
and you see the next generation of amd gpu architecture continue this trend?
Loxus
Member
Yes, because this results in massive die area saving.
Not doubt with RDNA4, AMD will get more out of Dual Issue. At the moment with Dual Issue, two shader instructions can be executed per clock cycle but under certain parallelism conditions.
I believe Sony API should be able to get more out of Dual Issue also. I wouldn't be surprised if Dual Issue and the AI Accelerators are the results of another Sony and AMD collaboration.
draliko
Member
I really do hope we'll get midgen refreshes from both vendors, but last gen I got burned with both of them, the Dev support wasn't really there, they were good when coded for but they weren't really used. Now with fp games being on pcs (sooner or later) having a midgen console is the only way to keep part of your customers hooked on the platform, because tech savvy high spenders have no problems moving to pc (probably already have a good pc gaming pc) and apart from those I really don't see a market for midgen refreshes.
supernova8
Banned
Probably totally unrelated but..... does any of this line up with rumors swirling around that RDNA4 won't have any high-end SKUs?
Maybe AMD is expecting to make some sort of high-end APU (for Sony) that will compete with whatever node capacity they would've used for high-end RDNA4, and then maybe low/mid-range RDNA4 will be on a larger (older) node and thus not compete with PS5 Pro APU supply.
Maybe AMD is expecting to make some sort of high-end APU (for Sony) that will compete with whatever node capacity they would've used for high-end RDNA4, and then maybe low/mid-range RDNA4 will be on a larger (older) node and thus not compete with PS5 Pro APU supply.
Last edited:
TheRedRiders
Member
It's more likely that RDNA 4 is just competing at the low end, I don't see them pushing their flagship through a console, it wouldn't make financial sense. Although I do suspect they will claw back some of their financials through the PS5 Pro.
Not likely AMD would have stood a chance against Nvidia's flagship (5090?) anyway.
supernova8
Banned
Which is funny because the likes of AdoredTV have been banging on and on and on about how AMD has a node advantage and how they will eventually wipe the floor with Nvidia.... and yet it never ever seems to happen.
TheRedRiders
Member
MLiD too, even now he's peddling how Nvidia are doomed in the gaming market.
I remember leading up to the RDNA 3 and 40 Series launch, he kept going on about how 40 Series is going to be "hot and loud", and RDNA 3 is going to be this powerful hyper efficient MCM design... he couldn't have been more wrong, In fact it was the other way around.
supernova8
Banned
It's all very confusing because we seem to bounce between three core arguments for why Radeon can never win:
1) Nvidia simply never drops the ball or comes up with a way to "win" even when they do drop the ball
or
2) AMD is intentionally losing because they're too busy winning in the CPU space
or
3) AMD is in a cartel with Nvidia and is essentially engaged in price fixing by intentionally never beating Nvidia's flagship and pricing everything the lower tiers identical to Nvidia's comparable offerings
TrebleShot
Member
Here’s a thought, could the 8k performance mode be 8k downsampling.
Loxus
Member
I'm thinking something is hidden.
Leaking information that sensitive can get you to loose your job.
I think it's partly true. In that there is no high in chiplets, but multiple chiplets make a high end card.
Like with Mi300. Max CUs per chiplet is 40 - 48 CUs, which might leak as no high-end GPU.
Torrent of Pork
Banned
You can emulate a lot of PS3 games on less powerful hardware than the current PS5, Sony just doesn't give a shit.
Loxus
Member
This is out of thr realm, but this I what I'd like the PS5 Pro specs to be.
Price: USD$599
Fabrication process: TSMC 4 nm FinFET
Die size: 284 mm²
CPU
Architecture: Custom Zen 4c
Cores: 12 Cores / 24 Threads
AVX-512 (256b data path)
L2 Cache: 1 MB per Core
L3 Cache: 24 MB (Shared with GPU)
Clock Speed: Variable frequency, clocked up to 4.65 GHz
FP32: 1.78 TFLOPS
XDNA AI Engine (Ryzen AI)
Clock Speed: Variable frequency, clocked up to 1.3 GHz
Cores: 90 AIE Tiles
FP16: 29.95 TOPS
GPU
Architecture: Custom RDNA 3.5
3 SEs / 30 WGPs / 60 CUs (3 WGPs/6 CUs disabled / 27 WGPs/54 CUs active)
Clock Speed: Variable frequency, clocked up to 2.546 GHz
Stream Processors: 3456
Compute Units: 54 - [FP32: 35.19 TFLOPS]
Texture Mapping Units: 216 - [Texture fill-rate: 549.93 GTexels/sec]
Render Output Units: 96 - [Pixel fill-rate: 244.41 GPixels/sec]
Ray Accelerators: 54 - [Ray-triangle: 137.48 G/sec]
AI Accelerators: 108 - [FP16: 70.39 TFLOPS]
Triangle rasterization: 10.18 GTriangles/sec
L2 Cache: 5 MB
Main Memory
GDDR6 SGRAM
Size: 20 GB
Speed: 18 Gbps
Bus Width: 320-bit
Bandwidth: 720 GB/sec
Price: USD$599
Fabrication process: TSMC 4 nm FinFET
Die size: 284 mm²
CPU
Architecture: Custom Zen 4c
Cores: 12 Cores / 24 Threads
AVX-512 (256b data path)
L2 Cache: 1 MB per Core
L3 Cache: 24 MB (Shared with GPU)
Clock Speed: Variable frequency, clocked up to 4.65 GHz
FP32: 1.78 TFLOPS
XDNA AI Engine (Ryzen AI)
Clock Speed: Variable frequency, clocked up to 1.3 GHz
Cores: 90 AIE Tiles
FP16: 29.95 TOPS
GPU
Architecture: Custom RDNA 3.5
3 SEs / 30 WGPs / 60 CUs (3 WGPs/6 CUs disabled / 27 WGPs/54 CUs active)
Clock Speed: Variable frequency, clocked up to 2.546 GHz
Stream Processors: 3456
Compute Units: 54 - [FP32: 35.19 TFLOPS]
Texture Mapping Units: 216 - [Texture fill-rate: 549.93 GTexels/sec]
Render Output Units: 96 - [Pixel fill-rate: 244.41 GPixels/sec]
Ray Accelerators: 54 - [Ray-triangle: 137.48 G/sec]
AI Accelerators: 108 - [FP16: 70.39 TFLOPS]
Triangle rasterization: 10.18 GTriangles/sec
L2 Cache: 5 MB
Main Memory
GDDR6 SGRAM
Size: 20 GB
Speed: 18 Gbps
Bus Width: 320-bit
Bandwidth: 720 GB/sec
AncientOrigin
Member
Not true. If the PS5 Pro had 0% raster uptick but implemented hardware tech discussed below, that alone would be a game changer.
Good post I don’t know if they will have it for Pro maybe for the PS6.I still think an pro without improved raster performance is not a good idea.Would be very bad news especially if it is priced more than the normal PS5.The only way that would make sense if they would replace the normal PS5 with the pro for same price.
Perrott
Member
That's not the way RDNA 3 TFs are calculated.
The RX 7800XT is actually a 19.4TF card, its just that you can in theory double that number by taking advantage of RDNA 3's double issue computing in FP32 operations.
The PS5 Pro has no way to achieve a real 24TF count, because we'd need a clock speed north of 3.5ghz, which is unthinkable for a console. It will 99% be a 17-18TF machine, with an RDNA 3 TF count of around 34TF.
Of course that, as Mr.Phoenix has already stated earlier in the thread, that double RDNA 3 performance will never be 100% achieved, so the real RDNA 3 performance of the PS5 Pro will most likely fall in the park of 24-29TFs.
Oh god, what a nightmare, I really hate AMD for introducing all this messy PR bullshit into console hardware discussion.
Last edited:
Why 12 cores? Personal,I felt Sony should go with a 10 core CPU with zen 4c as it cuts the difference between a 8 and 12 core.This is out of thr realm, but this I what I'd like the PS5 Pro specs to be.
Price: USD$599
Fabrication process: TSMC 4 nm FinFET
Die size: 284 mm²
CPU
Architecture: Custom Zen 4c
Cores: 12 Cores / 24 Threads
AVX-512 (256b data path)
L2 Cache: 1 MB per Core
L3 Cache: 24 MB (Shared with GPU)
Clock Speed: Variable frequency, clocked up to 4.65 GHz
FP32: 1.78 TFLOPS
XDNA AI Engine (Ryzen AI)
Clock Speed: Variable frequency, clocked up to 1.3 GHz
Cores: 90 AIE Tiles
FP16: 29.95 TOPS
GPU
Architecture: Custom RDNA 3.5
3 SEs / 30 WGPs / 60 CUs (3 WGPs/6 CUs disabled / 27 WGPs/54 CUs active)
Clock Speed: Variable frequency, clocked up to 2.546 GHz
Stream Processors: 3456
Compute Units: 54 - [FP32: 35.19 TFLOPS]
Texture Mapping Units: 216 - [Texture fill-rate: 549.93 GTexels/sec]
Render Output Units: 96 - [Pixel fill-rate: 244.41 GPixels/sec]
Ray Accelerators: 54 - [Ray-triangle: 137.48 G/sec]
AI Accelerators: 108 - [FP16: 70.39 TFLOPS]
Triangle rasterization: 10.18 GTriangles/sec
L2 Cache: 5 MB
Main Memory
GDDR6 SGRAM
Size: 20 GB
Speed: 18 Gbps
Bus Width: 320-bit
Bandwidth: 720 GB/sec
Last edited:
The rx 7800xt may perform like a 19 TFLOPS GPU in unoptimized games but it should perform better with games that better utilize RDNA3 potential.
Caio
Member
Yep, it can be confusing, and not easy to know which is actually the power/TF in real world scenario. But you said PS5 PRO could likely fall in the range of 24-29TFs, considering RDNA3(or even ""RDNA3.5"") Architecture/performance, so that would be great, and what needed to have this 2X the overall performance with heavier Ray Tracing and better effects, etc. I would put a signature on it.
Loxus
Member
Mark Cerny said dev wanted 16 Cores.
12 Cores is a half step up from 8 Cores.
Look at the die I did. 12 Cores work better than 10 Cores.
I got all that to fit in a ~284 mm² area or ~316 mm² on 5nm.
thicc_girls_are_teh_best
Member
This is out of thr realm, but this I what I'd like the PS5 Pro specs to be.
Price: USD$599
Fabrication process: TSMC 4 nm FinFET
Die size: 284 mm²
CPU
Architecture: Custom Zen 4c
Cores: 12 Cores / 24 Threads
AVX-512 (256b data path)
L2 Cache: 1 MB per Core
L3 Cache: 24 MB (Shared with GPU)
Clock Speed: Variable frequency, clocked up to 4.65 GHz
FP32: 1.78 TFLOPS
XDNA AI Engine (Ryzen AI)
Clock Speed: Variable frequency, clocked up to 1.3 GHz
Cores: 90 AIE Tiles
FP16: 29.95 TOPS
GPU
Architecture: Custom RDNA 3.5
3 SEs / 30 WGPs / 60 CUs (3 WGPs/6 CUs disabled / 27 WGPs/54 CUs active)
Clock Speed: Variable frequency, clocked up to 2.546 GHz
Stream Processors: 3456
Compute Units: 54 - [FP32: 35.19 TFLOPS]
Texture Mapping Units: 216 - [Texture fill-rate: 549.93 GTexels/sec]
Render Output Units: 96 - [Pixel fill-rate: 244.41 GPixels/sec]
Ray Accelerators: 54 - [Ray-triangle: 137.48 G/sec]
AI Accelerators: 108 - [FP16: 70.39 TFLOPS]
Triangle rasterization: 10.18 GTriangles/sec
L2 Cache: 5 MB
Main Memory
GDDR6 SGRAM
Size: 20 GB
Speed: 18 Gbps
Bus Width: 320-bit
Bandwidth: 720 GB/sec
IIRC RDNA3 uses 3 Prim Units each capable of 4 culling/2 rasterization of triangles per cycle. Can't see a reason for Sony to cut that down from 3 to 2, so at your clocks we'd see a bump in triangle rasterization to around 15.276 Gpolys/sec.
I'm still unsure if Sony goes with a wider bus; that will increase the size of the chip considerably. Maybe increasing the shared cache between CPU & GPU (maybe to 32 MB or 36 MB) and keeping the bus for memory 256-bit would be a better use of their budget. The memory speed could still stay at 18000 MT/s; I guess on paper that doesn't look much more than Series x's 560 GB/s but PS5 Pro would have that as universal bandwidth instead of segmented into a 10 GB pool. And PS5 Pro would have a much more efficient GPU, more efficient memory controllers and CPU to make better use of that bandwidth (also cache scrubbers).
Also not sure on the CPU increasing the core count; the new Zen archs are BC with the older ones which is great, but it's not like devs are going to use the extra cores for game logic. Can't see the OS suddenly needing those extra cores either; if anything it will need more memory. Maybe that is where a 320-bit bus comes in handy, to fit more capacity and get a boost in bandwidth? Perhaps that works out cheaper than increasing the CPU/GPU shared cache and throwing in another couple of GB of slower DDR4 for OS background tasks?
RDNA3 TF perf for PS5 Pro will probably be in the lower range of that estimate. Though, TF performance doesn't mean very much for gaming performance in general, we've all learned that loud and clear over the past three years
.Can't say 3.5 GHz is impossible for a console though. At one point 2.23 GHz was considered impossible, but that's what the PS5 accomplished. I can easily see the PS6 hitting 3+ GHz, even 3.5 GHz, for the GPU clock. Of course, at that point we're talking about a system on a much smaller node process, likely heavily utilizing chiplets and some form of advanced 3D packaging/die-stacking, and (preferably) technologies that move as much processing close to memory as possible to cut down on bus traffic, which eats away a lot of power usage in systems.
Some of the PS5's customizations like the cache scrubbers, can actually be considered early steps into that type of future design, so PS6 should be a quite interesting system to look at architecture & design-wise.
Last edited:
DenchDeckard
Moderated wildly
so, intelligent folk. We thinking like 17.5 Tf then? that would be realistic?
I can dig it!
I can dig it!
Crispy Gamer
Member
I'm sticking with 20 TF for marketing it's an easier sale plus I think after 4 years they should at least be able to hit that number
Crispy Gamer
Member
Not true. If the PS5 Pro had 0% raster uptick but implemented hardware tech discussed below, that alone would be a game changer.
Can't wait until they use this tech in games Sony already has by far the best processor for their TVs using the PS5 Pro as a test run then full implementation with PS6 would be a smart move
Daniel_Mallorca1985
Member
DeepEnigma
Gold Member
Neither is a new gen to them (RDR2), clearly.
Mr.Phoenix
Member
This one is tricky. Cache/Wider bus, both stand to increase the size of the chip. An alternative would be to just use the same-sized bus but with faster RAM. I expect they increase the cache in the APU vs. the current PS5, but not by anything that defines the architecture of the system. Eg. The PS5 already has a unified cache of 8MB for its CPU and 4MB for the GPU. So I see them making the CPU 16MB and GPU 6 -12MB. I am not sure they can make both the GPU and CPU feed from the same cache.
And then they leave the APU bus at 265bit. Just don't see them spending die real estate on a larger bus, they kinda already showed their hand with the cache scrubbers, which suggests they would always aim for the more efficient getting more with less approach. And More cache improves the system across the board, both CPU and GPU perf boosts. Basically, whatever the PS5pro is, it won't just be a direct 2+2=4 affair, it would be more like a 2+1=4 type thing.
Aso, I think people tend to forget that this is a console, at best, it's designed to hit 30-60fps at quality mode and 60-120fps at performance mode while dropping quality. It doesn't need the kinda bandwidth we would typically see in a PC where the hardware is designed to be able to push upwards of 300fps+ at all sorts of resolutions where applicable.
Agree with you here. The IPC gains alone going from Zen 2 to 4 is already going to be enough. Couple that with an up clock to say 4Ghz and that would make for a meaningful CPU boost over the current PS5 even if the core count is the same.
3.5Ghz? Naaaa.... the good thing is that we get to see the `PS5 GPU` on the market for over a year before the console is released. When we see what clocks they can hit on the PC version of the card, we then know where the PS5 can realistically land. My guess is that the PS5 would be somewhere between the boost and base clock of the 7700xtx if it's on a 5nm process. If on 4nm (which is a more advanced node than what the 7700xt is using) then it would either be at the boost clock of the GPU or a little bit higher.RDNA3 TF perf for PS5 Pro will probably be in the lower range of that estimate. Though, TF performance doesn't mean very much for gaming performance in general, we've all learned that loud and clear over the past three years
Can't say 3.5 GHz is impossible for a console though. At one point 2.23 GHz was considered impossible, but that's what the PS5 accomplished. I can easily see the PS6 hitting 3+ GHz, even 3.5 GHz, for the GPU clock. Of course, at that point we're talking about a system on a much smaller node process, likely heavily utilizing chiplets and some form of advanced 3D packaging/die-stacking, and (preferably) technologies that move as much processing close to memory as possible to cut down on bus traffic, which eats away a lot of power usage in systems.
Some of the PS5's customizations like the cache scrubbers, can actually be considered early steps into that type of future design, so PS6 should be a quite interesting system to look at architecture & design-wise.
Last edited:
Crispy Gamer
Member
I'm seeing a lot of assumptions it'll be zen 4 why is that? Wouldn't that raise the b.o.m significantly? I think it'll be zen 2 with higher clocks (4.5Ghz) so they can go crazy with the gpu for any raytracing customizations I don't think this machine will be a penny over 499.99 they have to cut somewhere plus I just want to see what Cerny and team can do with a bigger gpu budget 
I personally doubt the Pro will break 4+ GHz on the CPU unless it's like the the base PS5 with dynamic power use. I'm sure the Pro will have the As Series X CPU frequency.I'm seeing a lot of assumptions it'll be zen 4 why is that? Wouldn't that raise the b.o.m significantly? I think it'll be zen 2 with higher clocks (4.5Ghz) so they can go crazy with the gpu for any raytracing customizations I don't think this machine will be a penny over 499.99 they have to cut somewhere plus I just want to see what Cerny and team can do with a bigger gpu budget
Crispy Gamer
Member
Yeah I don't see them moving away from boost clocks I think Sony created a smart architecture I would like to see it on a more powerful pro model
Perrott
Member
Zen 4c is very efficient in terms of die space space utilization while simply being better than the PS5's Zen 2. Don't see why they wouldn't want to go for a much modern processor in order to open the gates for an improved raytracing experience on console.I'm seeing a lot of assumptions it'll be zen 4 why is that? Wouldn't that raise the b.o.m significantly? I think it'll be zen 2 with higher clocks (4.5Ghz) so they can go crazy with the gpu for any raytracing customizations I don't think this machine will be a penny over 499.99 they have to cut somewhere plus I just want to see what Cerny and team can do with a bigger gpu budget
Crispy Gamer
Member
Yeah that's a good reason for sure but at what cost? PS5 is already offering performance modes on every game a higher clocked zen 2 would probably be enough for this generation of software the gpu is the problem internal resolutions are already hitting 720p with graphics engines that aren't mind blowing like the UE5 demo. If it's zen 4 with an amazing gpu I'll be all for it but if incorporating a better cpu takes away from the gpu budget I'm struggling to see how that's a good idea
Mr.Phoenix
Member
This isn't like the PS4 where they were basically stuck with Jaguar. And even the PS4 adopted the newer GPU architecture.I'm seeing a lot of assumptions it'll be zen 4 why is that? Wouldn't that raise the b.o.m significantly? I think it'll be zen 2 with higher clocks (4.5Ghz) so they can go crazy with the gpu for any raytracing customizations I don't think this machine will be a penny over 499.99 they have to cut somewhere plus I just want to see what Cerny and team can do with a bigger gpu budget
Furthermore, AMD is already making Zen 4 APUs with integrated RDNA3 graphics. The Ryzen 7000 APU series which in addition to Zen 4 has 12CU of RDNA3. Extrapolation to a PS5pro doesn't get much easier than that. Technically speaking, its probably cheaper for Sony to make a Zen4 PS5 with RDNA3 than to figure out how to get RDNA3 paying nice with Zen 2.
Most importantly though, in addition to the IPC gains, which would mean they would be fine with the same or slightly higher clock, Zen 4 is also vastly more power efficient.
SolidQ
Member
Maybe even Zen5