Support NeoGAF

[Bojji]

Path Tracing is just Ray tracing with the Monte Carlo simulation.
It does send a lot more rays. And that is why performance drops, on all systems.

Yeah but it punishes AMD GPUs much more.

RT tiers seem to look like this:

Light RT - like RT shadows/AO, performs good on AMD

Medium RT - like RTGI or reflections, perform decent on AMD hardware

High RT - combination of RT shadows, RTGI, RTAO and reflections. Can punish AMD hardware much more compared to Nvidia (and Intel usually)

Ultra high - PT, just terrible on AMD

 

[winjer]

Yeah but it punishes AMD GPUs much more.

RT tiers seem to look like this:

Light RT - like RT shadows/AO, performs good on AMD

Medium RT - like RTGI or reflections, perform decent on AMD hardware

High RT - combination of RT shadows, RTGI, RTAO and reflections. Can punish AMD hardware much more compared to Nvidia (and Intel usually)

Ultra high - PT, just terrible on AMD

No doubt that AMD's RT hardware is significantly weaker. But there is something wrong here with CP2077, that goes beyond that.

 

[Mr.Phoenix]

My take on the WG number.

Being that this leak is coming from dev documents, then that number is the final usable number. So if Tom says, 30WGs, then that means the PS5pro has 30WGs available to the devs. Sony isn't sending dev kits with devs and saying, you have a total of 30WGs but can only use 28WGs. So this number accounts for the disabled CUs already.

If the leaks are coming from some sort of engineering sample testing or something, then yes, we can say the number shown doesn't account for disabled CUs.

As for the clocks thing, my personal theory is that the reason no one is really talking about it, is because its not finalized yet. Or its possible that what is currently on the dev kits, has already seen increases and they have been informed that the final clocks may be higher.

 

[winjer]

If the PS5 Pro is to be released by this years holydays, then the specs already have to be completed.
Probably, the chips are already in production.

 

[onQ123]

If the PS5 Pro is to be released by this years holydays, then the specs already have to be completed.
Probably, the chips are already in production.

The clocks can change as long as they have been tested at the higher clocks

 

[winjer]

The clocks can change as long as they have been tested at the higher clocks

Only by small margins.

 

[FireFly]

Different places to test probably but still here 6700x is almost 2x slower than 3070 and in TPU test it was almost 4x. Probably both game code patches and AMD drivers improved performance (a lot).

Enabling PT was reducing performance to 7.7% (!) in the TPU benchmarks. In the latest 2.1 benchmarks it's reducing performance to 16.9%, which is more in line with other path traced titles. The 3070 sees its performance reduce to 27%. If the ray tracing part takes half the time with RDNA 4, then that should bring the performance hit to close to Ampere levels.

 

[Mr.Phoenix]

If the PS5 Pro is to be released by this years holydays, then the specs already have to be completed.
Probably, the chips are already in production.

The specs are completed, but the clocks arent necessarily final. We have seen this happen before. You can have a situation where CPUclocks are increased from 3.8Ghz to 4Ghz, or where after enough testing has been done, they realise that instead of clocking the GPU at 2.2Ghz they can clock it at 2.4Ghz.

At the end of the day, your clocks come down to your cooling solution and what power envelope you are aiming to be in.

I don't think its a coincidence that clocks have not been leaked

 

[winjer]

The specs are completed, but the clocks arent necessarily final. We have seen this happen before. You can have a situation where CPUclocks are increased from 3.8Ghz to 4Ghz, or where after enough testing has been done, they realise that instead of clocking the GPU at 2.2Ghz they can clock it at 2.4Ghz.

At the end of the day, your clocks come down to your cooling solution and what power envelope you are aiming to be in.

I don't think its a coincidence that clocks have not been leaked

But the TFLOPs have, and these depend on the shader count and the clock speed.
If the clock speed changes, so do the TFLOPs.

 

[ChiefDada]

It's 7800XT but downclocked and without inifinity cache. Plus of course better RT performance and Ai hardware.



I posted hybrid RT benchmarks before where AMD fares much better but in pure PT?

There is no 6700 on this chart but of course it would have lower FPS than 6700XT, assuming 5FPS it would put 4x better GPU in 20FPS section. Of course Pro will have 2-4x increase so norm would most likely be 2x or 3x, 4x will be rarer.

So... 3070 level.

Thank you for perfectly illustrating my point of why CU/processor count is a very stupid metric by itself in predicting PS5 Pro RT. This entire chart would be eternally confusing to Richard and DF, which is downright pathetic

You slotting PS5 Pro performance between 4060 and 3070 8gb cards is also incorrect for similar reasons.

 

[Bojji]

Thank you for perfectly illustrating my point of why CU/processor count is a very stupid metric by itself in predicting PS5 Pro RT. This entire chart would be eternally confusing to Richard and DF, which is downright pathetic

You slotting PS5 Pro performance between 4060 and 3070 8gb cards is also incorrect for similar reasons.

Both cards are not VRAM limited in 1080p, so VRAM is irrelevant in this particular test.

7900xtx and 7700xt scale very similar with their TF numbers here, almost double performance from 7900xtx.

Of course RT hardware is the biggest mystery of Pro and Richard shouldn't be that certain about some things. Still I will be mighty surprised if pro will end up with better results than 2x larger RDNA3 GPU.

We don't even know what those 2-4x rt factors are for, if it's standard hybrid RT or PT like calculations? PT is not used on consoles so far so I doubt they would give developers metrics about that.

 

[Loxus]

My take on the WG number.

Being that this leak is coming from dev documents, then that number is the final usable number. So if Tom says, 30WGs, then that means the PS5pro has 30WGs available to the devs. Sony isn't sending dev kits with devs and saying, you have a total of 30WGs but can only use 28WGs. So this number accounts for the disabled CUs already.

If the leaks are coming from some sort of engineering sample testing or something, then yes, we can say the number shown doesn't account for disabled CUs.

As for the clocks thing, my personal theory is that the reason no one is really talking about it, is because its not finalized yet. Or its possible that what is currently on the dev kits, has already seen increases and they have been informed that the final clocks may be higher.

I posted this earlier.
Microsoft showed and talked about XBSX full 28WGP in the documents.

Tom Henderson just saying the PS5 Pro has 30WGP isn't enough confirmation that is the active WGP number.

We need more info like ROPs count.

Based on different RDNA4 rumors like this one, ROPs are:
2SE = 64 ROPs
3SE = 96 ROPs
4SE = 128 ROPs

This will help us understand the active/full amount of WGP.

Then we have to understand how certain WGP configurations can fit in on a die without wasting die space by having empty spaces.

Desktop APU Cezanne die shot shows white spaces. Pockets of silicon without transistors.

Can 64-68 fully enable CUs fit on a die without white spaces is what we should be trying to figure out in the mean time until more concrete info comes out.

I can't be the only one here that's trying to apply logic to these leaks. Come on guys.

 

[ChiefDada]

7900xtx and 7700xt scale very similar with their TF numbers here, almost double performance from 7900xtx.

Yes because they are the same architecture and key areas of the chip scale near linearly. Especially in the case of GPU bandwidth, where bandwidth/CU ratio is higher/greater than 1:

1. CUs: 54 v. 96 = 1.77x
2. ROPS: 96 v. 192 = 2x
3. Bandwidth (Most important): 432gb/s v. 960gb/s = 2.2

Again Richard foolishly stated that 7900xtx should perform path-tracing better "even if it is from an older architecture" (so dumb) because 7900xtx has more CUs than PS5 Pro.

 

[TheRedRiders]

I posted this earlier.
Microsoft showed and talked about XBSX full 28WGP in the documents.

Tom Henderson just saying the PS5 Pro has 30WGP isn't enough confirmation that is the active WGP number.

We need more info like ROPs count.

Based on different RDNA4 rumors like this one, ROPs are:
2SE = 64 ROPs
3SE = 96 ROPs
4SE = 128 ROPs

This will help us understand the active/full amount of WGP.

Then we have to understand how certain WGP configurations can fit in on a die without wasting die space by having empty spaces.

Desktop APU Cezanne die shot shows white spaces. Pockets of silicon without transistors.

Can 64-68 fully enable CUs fit on a die without white spaces is what we should be trying to figure out in the mean time until more concrete info comes out.

I can't be the only one here that's trying to apply logic to these leaks. Come on guys.

You really love your slides ay!

I kid of course.

 

[Loxus]

You really love your slides ay!

I kid of course.

Some people get a better understanding visually with pictures then reading with a lot of text.

So, I try to cater to different learning styles.

 

[ResilientBanana]

Same zen 2 cpu. 10% clock speed boost.

Im out.

This happens every time a pro console is announced. They can't use a different version of architecture, it's not PlayStation 6. Anyone expecting Zen 2.5, Zen 3 or Zen 4 were getting hyped up for a mass amount of disappointment.

 

[playsaves3]

I think the vanilla game has 2x bounces and 2x rays. As you mentioned before, 1x bounce and 2x rays could be feasible on the PS5 Pro. It's much more performant than the default 2x2 while not looking much worse. Also, I assume if Sony wanted it, they would aim for 30fps, not 60, so something like 1440p with an internal res of 1080p (or 960p?) with PSSR, and 1x bounce and 2x rays sounds plausible with a 30fps lock. Visually, it would look way better than the RT mode on the regular PS5 minus a softer IQ but if PSSR is decent, the IQ should still be passable.

The thing is, I'm not sure how much interest Sony will have in implementing something of the sort.

Whatever the case, I think it's a fair question and Rich just laughing it off is just bizarre. At that point, best come clean and just admit you were completely wrong about the performance of the machines instead of being befuddled whenever the Series X doesn't perform like you think it should. Didn't Cerny mention that it was much more difficult to keep a higher number of CUs busy, thus explaining in part why he went with 36? I vaguely remember him touching upon this in Road to PS5. Not that he'd need to tell us that anyway because it was obvious beforehand. DF are really looking like a bunch of tools there,

Tbf I don’t think Rich is being biased assuming we are talking about cpu limitations and I immediately call out bias from either Alex or Richard

 

[Not_A-Cop_Maybe]

I posted this earlier.
Microsoft showed and talked about XBSX full 28WGP in the documents.

Tom Henderson just saying the PS5 Pro has 30WGP isn't enough confirmation that is the active WGP number.

We need more info like ROPs count.

Based on different RDNA4 rumors like this one, ROPs are:
2SE = 64 ROPs
3SE = 96 ROPs
4SE = 128 ROPs

This will help us understand the active/full amount of WGP.

Then we have to understand how certain WGP configurations can fit in on a die without wasting die space by having empty spaces.

Desktop APU Cezanne die shot shows white spaces. Pockets of silicon without transistors.

Can 64-68 fully enable CUs fit on a die without white spaces is what we should be trying to figure out in the mean time until more concrete info comes out.

I can't be the only one here that's trying to apply logic to these leaks. Come on guys.

Man, i'm always in awe of you guys that know this kind of stuff. I've built and assembled my own computers before, but reading stuff like this makes me think I'm trying to decipher Klingon.

 

[SlimySnake]

This happens every time a pro console is announced. They can't use a different version of architecture, it's not PlayStation 6. Anyone expecting Zen 2.5, Zen 3 or Zen 4 were getting hyped up for a mass amount of disappointment.

?

PS4 Pro used the Polaris GPU which was a different GCN architecture.

PS5 Pro uses RDNA4 or RDNA 3.5.

They couldve used Zen 3 or Zen 4. They literally have Zen 4 APUs.

 

[ResilientBanana]

?

PS4 Pro used the Polaris GPU which was a different GCN architecture.

PS5 Pro uses RDNA4 or RDNA 3.5.

They couldve used Zen 3 or Zen 4. They literally have Zen 4 APUs.

Maybe I should have phrased that better. They don't change out the CPU for pro refreshes, because they can't. As far as the GPU is concerned weren't both of those GPU's using GCN2.0?

AMD Playstation 4 GPU Specs

AMD Liverpool, 800 MHz, 1152 Cores, 72 TMUs, 32 ROPs, 8192 MB GDDR5, 1375 MHz, 256 bit

www.techpowerup.com

AMD Playstation 4 Pro GPU Specs

AMD Neo, 911 MHz, 2304 Cores, 144 TMUs, 32 ROPs, 8192 MB GDDR5, 1700 MHz, 256 bit

www.techpowerup.com

 

[SlimySnake]

Maybe I should have phrased that better. They don't change out the CPU for pro refreshes, because they can't. As far as the GPU is concerned weren't both of those GPU's using GCN2.0?

AMD Playstation 4 GPU Specs

AMD Liverpool, 800 MHz, 1152 Cores, 72 TMUs, 32 ROPs, 8192 MB GDDR5, 1375 MHz, 256 bit

www.techpowerup.com

AMD Playstation 4 Pro GPU Specs

AMD Neo, 911 MHz, 2304 Cores, 144 TMUs, 32 ROPs, 8192 MB GDDR5, 1700 MHz, 256 bit

www.techpowerup.com

PS4 used Tahiti or GCN 2.

PS4 Pro used Polaris or GCN 4.

 

[Perrott]

PS4 used Tahiti or GCN 2.

 

[ResilientBanana]

PS4 used Tahiti or GCN 2.

PS4 Pro used Polaris or GCN 4.

You sure?

Analyzing Sony’s PlayStation 4 Pro Hardware Reveal: What Lies Beneath

www.anandtech.com

Polaris is a great example of this. Compared to Fiji (GCN 1.2), AMD revised the command processor, the display controller, the multimedia engines and the geometry processor. Meanwhile the all-important shader cores were technically revised, but in practice other than increased internal memory, they are ISA-identical to GCN 1.2. Finally, AMD did not touch the render backends, the rasterizer, or the work scheduler at all.

This means that Sony can equally pick and choose which parts of the GPU they want to update, which is important given the forward compatibility mandate of the PS4 Pro. Adopting features from Polaris does not mean that this is a full Polaris GPU.

Again, it's very important that they stay on similar architecture.

EDIT: as far as I can tell PS4 Pro is GCN2 with some GCN5 features, but not enough to push it out of GCN2.

 

[SlimySnake]

You sure?

Analyzing Sony’s PlayStation 4 Pro Hardware Reveal: What Lies Beneath

www.anandtech.com

Again, it's very important that they stay on similar architecture.

Yes. PS4 pro is polaris. X1x is also polaris.

PS5 Pro isnt going to be RDNA2 with RDNA 4 pieces stuck inside of it. Its more expensive to do that then just pick whatever AMD has cooking and make changes to it.

Regardless, even if Sony could pick and choose what to get for the GPUs, they could do the same for the CPUs. pick and choose zen 4 shit. They arent moving from jaguar to zen here. they already are on zen. The IPC gain from Zen 2 to Zen 4 is 30%. They couldve kept the clocks the same and have it perform like a 4.55 ghz CPU.

 

[Gaiff]

?

PS4 Pro used the Polaris GPU which was a different GCN architecture.

PS5 Pro uses RDNA4 or RDNA 3.5.

They couldve used Zen 3 or Zen 4. They literally have Zen 4 APUs.

Pretty clear he's talking about the CPU, no?

PS4 and PS4 Pro have the same CPU as far as I'm aware with the Pro having a small bump in clocks. Same deal with the PS5 and PS5 Pro.

 

[Mr Moose]

You sure?

Analyzing Sony’s PlayStation 4 Pro Hardware Reveal: What Lies Beneath

www.anandtech.com

Again, it's very important that they stay on similar architecture.

EDIT: as far as I can tell PS4 Pro is GCN2 with some GCN5 features, but not enough to push it out of GCN2.

 

[Loxus]

?

PS4 Pro used the Polaris GPU which was a different GCN architecture.

PS5 Pro uses RDNA4 or RDNA 3.5.

They couldve used Zen 3 or Zen 4. They literally have Zen 4 APUs.

Don't forget we have Zen4C APUs as well.

They know how they got BC working on Zen2. Zen4C is still within the Zen architecture, adding BC shouldn't cost as much R&D as with Zen2.

Sony is also helping share R&D when it comes to chip design, so using Zen3 or Zen4 shouldn't increase the PS5 Pro cost that much over the PS5.

PS4 Pro launched with the same price as the PS4, $400. PS5 launch price is $500.
I would rather them sell me a latest tech console at $550 - $600, then trying to get the PS5 Pro price to $500.

If it's supposed to be a premium console, I'm willing to pay a premium price. $549-$599 is a great price considering they're selling the Edge Controller and Earbuds for $200, and $300 for Inzone Headset.

 

[SlimySnake]

Pretty clear he's talking about the CPU, no?

PS4 and PS4 Pro have the same CPU as far as I'm aware with the Pro having a small bump in clocks. Same deal with the PS5 and PS5 Pro.

no, hes talking about the GPU in that post.

Polaris is a great example of this. Compared to Fiji (GCN 1.2), AMD revised the command processor, the display controller, the multimedia engines and the geometry processor. Meanwhile the all-important shader cores were technically revised, but in practice other than increased internal memory, they are ISA-identical to GCN 1.2. Finally, AMD did not touch the render backends, the rasterizer, or the work scheduler at all.

This means that Sony can equally pick and choose which parts of the GPU they want to update, which is important given the forward compatibility mandate of the PS4 Pro. Adopting features from Polaris does not mean that this is a full Polaris GPU.

EDIT: as far as I can tell PS4 Pro is GCN2 with some GCN5 features, but not enough to push it out of GCN2.

 

[ResilientBanana]

Yes. PS4 pro is polaris. X1x is also polaris.

PS5 Pro isnt going to be RDNA2 with RDNA 4 pieces stuck inside of it. Its more expensive to do that then just pick whatever AMD has cooking and make changes to it.

Regardless, even if Sony could pick and choose what to get for the GPUs, they could do the same for the CPUs. pick and choose zen 4 shit. They arent moving from jaguar to zen here. they already are on zen.

I just posted a quote stating that the GPU is not full Polaris though. The majority of the GPU was GCN2 for compatibility reasons. This will most likely be the case for the PS5 Pro.

Again, elements of Polaris. This does not equate to a Polaris GPU.

On the flipside, I miss my snow white PS4 Pro.

Zen is CPU though.

I know, the conversation got flipped to GPU somewhere. As far as the CPU goes, it's not changing architecture too drastically, most likely only clock speeds.

 

[Mr Moose]

no, hes talking about the GPU in that post.

Zen is CPU though.

 

[MisterXDTV]

Of course he was talking about the CPU...

Zen is the CPU name

 

[Loxus]

as far as I can tell PS4 Pro is GCN2 with some GCN5 features, but not enough to push it out of GCN2.

This is how I understand the PS4 Pro as well, features may make it be considered GCN4 but at it's core still GCN2. Which is the same situation as PS5.

 

[SlimySnake]

Of course he was talking about the CPU...

Zen is the CPU name

Zen is CPU though.

We were. not anymore. conversation shifted to GPU.

As far as the GPU is concerned weren't both of those GPU's using GCN2.0?

AMD Playstation 4 GPU Specs

AMD Liverpool, 800 MHz, 1152 Cores, 72 TMUs, 32 ROPs, 8192 MB GDDR5, 1375 MHz, 256 bit

www.techpowerup.com

AMD Playstation 4 Pro GPU Specs

AMD Neo, 911 MHz, 2304 Cores, 144 TMUs, 32 ROPs, 8192 MB GDDR5, 1700 MHz, 256 bit

www.techpowerup.com

PS4 used Tahiti or GCN 2.

PS4 Pro used Polaris or GCN 4.

You sure?
EDIT: as far as I can tell PS4 Pro is GCN2 with some GCN5 features, but not enough to push it out of GCN2.

Yes. PS4 pro is polaris. X1x is also polaris.

PS5 Pro isnt going to be RDNA2 with RDNA 4 pieces stuck inside of it. Its more expensive to do that then just pick whatever AMD has cooking and make changes to it.

Regardless, even if Sony could pick and choose what to get for the GPUs, they could do the same for the CPUs. pick and choose zen 4 shit. They arent moving from jaguar to zen here. they already are on zen. The IPC gain from Zen 2 to Zen 4 is 30%. They couldve kept the clocks the same and have it perform like a 4.55 ghz CPU.

 

[ResilientBanana]

This is how I understand the PS4 Pro as well, but features may make it be considered GCN4. Which is the same situation as PS5.

I still don't think that's the case. Sites were very clear in stating that a majority of the GPU was GCN2. I don't think the assumption of GCN4 could be made.

 

[Lysandros]

Again PS4 PRO is mainly based on GCN 4/Polaris (ex. delta color compression) with some GCN 5/Vega features like rapid pack math along with some Sony customizations (likely) inherited from PS4 (volatile bit and onion+ bus) and some new ones like hardware ID buffer. Techpowerup is awfully wrong about the system, they are also wrong about the number of ROPs which is 64 not 32.

 

[Loxus]

I still don't think that's the case. Sites were very clear in stating that a majority of the GPU was GCN2. I don't think the assumption of GCN4 could be made.

I edited my post.

This is how I understand the PS4 Pro as well, features may make it be considered GCN4 but at it's core still GCN2. Which is the same situation as PS5.

I think it's GCN2 at it's core, but the features mean it can be considered as GCN4.

 

[SlimySnake]

I just posted a quote stating that the GPU is not full Polaris though. The majority of the GPU was GCN2 for compatibility reasons. This will most likely be the case for the PS5 Pro.

Every single AMD GPU is already Backwards compatible. Polaris was BC with GCN 2.0. GCN 2.0 was BC with GCN 1.1.

The same is true for RDNA.

Again, if they are able to steal features from RDNA4 for the GPU, they should be able to do the same with the Zen 4 CPUs. This is not a PS6, but its a brand new console that is already undergoing R&D which requires a massive upgrade to the console's RT and machine learning capabilities. Sticking with Zen 2 with all its documented bottlenecks is not a good idea. Had they increased the clocks by 30% like they did for the jaguar CPUs in PS4 Pro, it might not have required a zen 4 upgrade because most games are single threaded and having a 30% increase in clocks would essentially be the same as a 30% IPC gain from zen 4.

 

[SlimySnake]

Again PS4 PRO is mainly based on GCN 4/Polaris (ex. delta color compression) with some GCN 5/Vega features like rapid pack math along with some Sony customizations (likely) inherited from PS4 (volatile bit and onion+ bus) and some new ones like hardware ID buffer. Techpowerup is awfully wrong about the system, they are also wrong about the number of ROPs which is 64 not 32.

Yeah, Sony is basically taking whatever AMD comes up with and then make changes to it. The entire reason going with AMD gen after gen is to be able to take advantage of upgrades AMD comes up with. This happened at the start of this gen when people were speculating that Sony had a custom RT tech. not its the same fucking thing AMD put in all of their GPUs including microsofts.

Microsoft does the same with even fewer customizations. mostly number of CUs. The X1X is basically a 580 where as the PS4 Pro was basically a 480. Vega was in R&D around the same time as a the Pro so Cerny took the rapid packed math feature from it and inserted supported for checkerboarding. he's following the same playbook here, not in this time, instead of checkerboarding, its AI upscaling and instead of rapid packed math, its enhanced RT cores.

 

[playsaves3]

Don't forget we have Zen4C APUs as well.

They know how they got BC working on Zen2. Zen4C is still within the Zen architecture, adding BC shouldn't cost as much R&D as with Zen2.

Sony is also helping share R&D when it comes to chip design, so using Zen3 or Zen4 shouldn't increase the PS5 Pro cost that much over the PS5.

PS4 Pro launched with the same price as the PS4, $400. PS5 launch price is $500.
I would rather them sell me a latest tech console at $550 - $600, then trying to get the PS5 Pro price to $500.

If it's supposed to be a premium console, I'm willing to pay a premium price. $549-$599 is a great price considering they're selling the Edge Controller and Earbuds for $200, and $300 for Inzone Headset.

I was willing to pay up to 699 if the specs backed it really still hoping some parts of these leaks are wrong or potentially change somewhat

 

[playsaves3]

Every single AMD GPU is already Backwards compatible. Polaris was BC with GCN 2.0. GCN 2.0 was BC with GCN 1.1.

The same is true for RDNA.

Again, if they are able to steal features from RDNA4 for the GPU, they should be able to do the same with the Zen 4 CPUs. This is not a PS6, but its a brand new console that is already undergoing R&D which requires a massive upgrade to the console's RT and machine learning capabilities. Sticking with Zen 2 with all its documented bottlenecks is not a good idea. Had they increased the clocks by 30% like they did for the jaguar CPUs in PS4 Pro, it might not have required a zen 4 upgrade because most games are single threaded and having a 30% increase in clocks would essentially be the same as a 30% IPC gain from zen 4.

Or zen 5

 

[Loxus]

I was willing to pay up to 699 if the specs backed it really still hoping some parts of these leaks are wrong or potentially change somewhat

Could be early dev-kit specs.

Tom Henderson did say Testkits will be identical to the final product.

This sounds to me like the Devkits could possibly be using older hardware, which the current leaks might be based on.

 

[MisterXDTV]

We are not starting the CPU discussion again, right?

It is what it is... If you can't live without a new CPU, wait for PS6....

It's that simple

 

[MisterXDTV]

This sounds to me like the Devkits could possibly be using older hardware, which the current leaks might be based on.

The basics of the machine were set YEARS ago as they probably started working on it right after the launch of the standard PS5

We are not going to get a CPU upgrade or any other significant change going from the Fall 2023 devkits to the Spring 2024 final testkits

 

[Loxus]

The basics of the machine were set YEARS ago as they probably started working on it right after the launch of the standard PS5

We are not going to get a CPU upgrade or any other significant change going from the Fall 2023 devkits to the Spring 2024 final testkits

Was RDNA3.5/4 available after the PS5 launched?

The chip in the devkits could be dated in 2022 with RDNA3 and revised with newer hardware in 2023 with RDNA3.5/4.

Revising with Zen3 or 4 isn't impossible. It's still the same architecture family and not a case like going from Jaguar to Zen.

It's not like the PS5 chip didn't have a Major revision.
PS5 had a major revision, going from Navi10 Lite to Navi12.

RDNA1 to RDNA2 and probably Zen1 to Zen2.

This by the way is how the PS5 first concrete specs were leaked.

I wonder if any of you remember the Flute and Gonzalo PS5 leaks?

Back then everyone was trying to decode the codenames.

Those were the good times, even Digital Foundry chimed in.
Now we got people blindly believe in leaks without logic.

 

[solidus12]

Can’t believe that even the “experts” were fooled by that.

 

[Mr.Phoenix]

Was RDNA3.5/4 available after the PS5 launched?

The chip in the devkits could be dated in 2022 with RDNA3 and revised with newer hardware in 2023 with RDNA3.5/4.

Revising with Zen3 or 4 isn't impossible. It's still the same architecture family and not a case like going from Jaguar to Zen.

It's not like the PS5 chip didn't have a Major revision.
PS5 had a major revision, going from Navi10 Lite to Navi12.

RDNA1 to RDNA2 and probably Zen1 to Zen2.

This by the way is how the PS5 first concrete specs were leaked.

I wonder if any of you remember the Flute and Gonzalo PS5 leaks?

Back then everyone was trying to decode the codenames.

Those were the good times, even Digital Foundry chimed in.
Now we got people blindly believe in leaks without logic.

this is not a PS6 though. It's a PS5pro.

You need to just ask yourself a simple question, what do they want to do to the base PS5 game, and what kind of hardware is good enough to do that? At no more than $499. And whatever hardware you arrive at, that is what it will likely be.

The second you start throwing in what you wish they would or could do? This all goes to hell in a baby stroller.

People seem to keep forgetting this. This is not a next-gen PlayStation. Its a slightly better PS5. The base, is the PS5, they are building on top of that. Its like taking the PS5, putting in a better GPU, faster RAM, clocking some things a little higher, and then leaving everything else untouched.

Or another way to look at it, its like having $400 PC, then spending $100 on a GPU and RAM upgrade.

 

[Ashamam]

People seem to keep forgetting this. This is not a next-gen PlayStation. Its a slightly better PS5.

Yep obvious to some and literally by design. Why didn't they improve the CPU? Because they didn't want to! You have DF saying its the best they can do whilst at the same time saying it helps compatibility lol. Its not the best they can do at all, its literally what they wanted to do. For compatibility sure, but also because they don't want a divergence between the two SKU's beyond a very specific targeted area. The Pro is very carefully designed with boundaries to control developers and ensure the standard machine isn't left behind.

 

[MisterXDTV]

this is not a PS6 though. It's a PS5pro.

This is the whole point of the discussion...

It's one thing to be disappointed in the specs, but you can't basically REFUSE what the leaks says just because you don't like them or they are not what you THOUGHT they should have been.

And this applies to any part of the leaks: GPU, CPU, memory, ray-tracing etc...

We should believe in leaked OFFICIAL documents at least until those leaks are PROVEN wrong

I will say it again: PS5 Pro will NEVER have exclusive software in its lifetime as it will always be TIED to the base model, that's the point of a mid-gen.

One could say: "Why bother then?"

Because generations now last double compared to the past: PS4 is still getting meaningful games after more than 10 YEARS on the market.

Mid-gen serves as a bridge between the two generations, And again, it's an OPTIONAL bridge: if you are happy with what you have or you don't find mid-gen good enough to warrant an upgrade, then SKIP it and save your money for PS6

 

[Fafalada]

Again, elements of Polaris. This does not equate to a Polaris GPU.

This is the same debate people had about RDNA2 vs X/PS5 for 2 years straight.
Anyway I know internet factions are completely entrenched in their belief systems on this - so it's not really a debate anymore. But I'll just leave this here - the whole idea of 'backporting functional blocks' into an older architecture is definitely the more(most?) complicated way to do anything in engineering.

 

[Mr.Phoenix]

Yep obvious to some and literally by design. Why didn't they improve the CPU? Because they didn't want to! You have DF saying its the best they can do whilst at the same time saying it helps compatibility lol. Its not the best they can do at all, its literally what they wanted to do. For compatibility sure, but also because they don't want a divergence between the two SKU's beyond a very specific targeted area. The Pro is very carefully designed with boundaries to control developers and ensure the standard machine isn't left behind.

Not just ensure its never left behind, but ensure it remains the base platform. They want devs to build a PS5 game as if the Pro doesn't even exist, then make the PS5pro improvements on top of that.