PS5 Pro *spawn

[chris1515]

https://twitter.com/i/web/status/1768487031225606236

 

[chris1515]

https://twitter.com/i/web/status/1768492840404021463

 

[chris1515]

https://twitter.com/i/web/status/1768493959095996813

 

[chris1515]

 

[chris1515]

 

[chris1515]

https://twitter.com/i/web/status/1768532680591696378

https://twitter.com/i/web/status/1768528101883912441

 

[Karamazov]

wonder if it could run some path tracing mode of CP77 at 30fps with PSSR.

 

[Nesh]

Nm

 

[chris1515]

NX gamer say 2430 mhz for frequency

 

[Charlietus]

NX gamer say 2430 mhz for frequency

I can't believe I am saying this but MLID leak looks a lot more credible than this nxgamer slide. He either did his research and fabricated something that will be really close to the actual product or that's a real slide from Sony.

The only thing missing is an additional 500 mb or a gigabyte of ddr4 for the OS, if the upscaling takes 200 mb the GPU is going to need a bit more available memory for the game.

 

[Shifty Geezer]

So Sony have broken free of the fixed legacy CU counts. This is also a pretty niche machine. I think it'll largely be about 8k for a tiny fraction of folk, or current PS5 games at a better framerate and IQ. Although how does PSSR fit in with game's current updacling? Presumably they'll all need a patch to swap reconstruction method.

 

[chris1515]

And I learn NX gamer comparison was around a 7800XT not the real specs at all.

 

[Charlietus]

I did not say it is true. I think frequency is obsfucated. I saw the MLID video now and he said he changed some of the thing in the document to protect his source.

I don't think he would obfuscate something like frequency, he mainly talks about it when showing the image comparing fsr2 with pssr.

But maybe that's something he would do

 

[Globalisateur]

I did not say it is true. I think frequency is obsfucated. I saw the MLID video now and he said he changed some of the thing in the document to protect his source.

This could explain why the whole thing doesn't make sense, like the downclocked GPU vs PS5. This could explain why he never discloses FP32 TFLOPs numbers.

 

[chris1515]

I don't think he would obfuscate something like frequency, he mainly talks about it when showing the image comparing fsr2 with pssr.

But maybe that's something he would do

Why some elements suggest a 2450 mhz frequency and other a 2180 mhz frequency? Kepler think the same thing than me.

 

[Globalisateur]

45% better rendering for unoptimized PS5 games is a LOT more. It should be near impossible to do that on the same 6nm node. Except if they account for some dual issue shenanigans even working on unoptimized PS5 games.

45% would mean every 40fps games could run at 60fps out of the box with the same resolution / fidelity, or solid 30fps games running mostly in the VRR window.

 

[chris1515]

I wrote kopitekimi in the previous message but I meant Kepler, sorry :/

Where is the 2450 MHz frequency coming from?
I was thinking 2180 MHz for native PS5 pro games and 2230 MHz for BC PS5 games on pro, while disabling CU's to 36. That would mean no automatic PS5 games enhancements, and I don't know if I can believe that.

https://twitter.com/i/web/status/1768528101883912441

 

[snc]

Where is the 2450 MHz frequency coming from?

from 300 tops 8bit

 

[Globalisateur]

So there are several conflicting data in those specs. It's either bullshit or he modified some data to protect his sources.

 

[chris1515]

So there are several conflicting data in those specs. It's either bullshit or he modified some data to protect his sources.

He gives a disclaimer saying I modified the document before the section talking about PS5 Pro.