It's the TOOOOOOOOOLS
Turns out the tools can't save you from the Series S
[MLiD] Interview with veteran Tools Engineer (mentions Sony dev teams)
- Thread starter anonpuffs
- Start date
It's the TOOOOOOOOOLS
Turns out the tools can't save you from the Series S
It think it could be hard to reach double the amount of Tflops.
30 WGPs, 3 SEs. So 10 WGPs per SE, one disabled for yield. That is 27 WGPs or 54 CUs, even with 2.7 Ghz clock that is 18 Tflops of single issues FP32.
But again Tflops is not the good indicator of performance. If anything I think improving RT performance and better CPU to double framerates on PS5 games is good enough.
Just saw this, well the part about VR going mainstream. I like what Apple is doing with their VR WRT dedicated silicon processing for many parts of the "VR pipeline" to streamline everything that way, but the issue with Apple is simple: their prices are just way too high for VR to ever become mainstream through them, at least for the next decade or so.
They arguably have the best VR tech when you consider the integrated experience, and tech at least comparable to Sony when it comes to gaming-specific applications (in theory), but I think Apple's VR being super-mainstream is probably a good ways off. IMO Sony are still in the best position to make it happen in some form sooner, at least mainstream enough to where PlayStation consoles are mainstream, and that could in effect help takes like Apple's gain even bigger mainstream success later on.
But for that to happen, I think Sony have to find a way to include VR by default with their next console, and I think their Remote Play & PS Link technology will be instrumental there. An entry-level VR headset wouldn't need to do much of the processing on the headset itself; just have the PS6 handle all the processing and stream it to the headset. That's the setup in a sense with PS5 & PSVR2, but for PS6 I think Sony could take some ideas from Apple and combine that with the dedicated ASIC approach they took with PS5's SSD I/O subsystem. Because if you want to accelerate heavy amounts of VR processing, that would be the best approach to do it.
An entry-level PSVR headset wouldn't need to aim for ridiculous resolution or framerate, just something that would be generally acceptable for a mainstream audience. Enthusiasts who want more performance in that regard could just get another higher-end PSVR headset. The goal should be to include something entry-level wise that can be included with every PlayStation console. I think customers would be fine with that at a $499 or even $599 price point, depending on certain other specifications.
For example, a $499 PS6 with the entry-level headset but cut-back on storage capacity, but by that point most people should have an SSD they can swap into their system. And from there a $599 model with larger storage. I think you can do this while still providing a solid performance increase in the hardware itself for non-VR games. Outside of Apple and maybe Meta, Sony/PlayStation are the only real option for VR becoming mainstream, but the VR will have to be included with all systems in some way to make that happen.
I think PS6 with VR as default can provide great gaming experience, and pretty good general usability experience, though not at the ultra-chic streamlined level Apple can offer. They could do it at a much cheaper price, though. And their gaming performance is leagues ahead of Meta's offering, I think they can continue this even with just streaming the vast bulk of the processing to a headset via Remote Play & PS Link, as long as the PS6 takes cues from Apple and has dedicated silicon for offloading & streamlining lots of parts of the VR data processing pipeline (similar to what they've done for the SSD in PS5).
An exciting future, for sure.
They arguably have the best VR tech when you consider the integrated experience, and tech at least comparable to Sony when it comes to gaming-specific applications (in theory), but I think Apple's VR being super-mainstream is probably a good ways off. IMO Sony are still in the best position to make it happen in some form sooner, at least mainstream enough to where PlayStation consoles are mainstream, and that could in effect help takes like Apple's gain even bigger mainstream success later on.
But for that to happen, I think Sony have to find a way to include VR by default with their next console, and I think their Remote Play & PS Link technology will be instrumental there. An entry-level VR headset wouldn't need to do much of the processing on the headset itself; just have the PS6 handle all the processing and stream it to the headset. That's the setup in a sense with PS5 & PSVR2, but for PS6 I think Sony could take some ideas from Apple and combine that with the dedicated ASIC approach they took with PS5's SSD I/O subsystem. Because if you want to accelerate heavy amounts of VR processing, that would be the best approach to do it.
An entry-level PSVR headset wouldn't need to aim for ridiculous resolution or framerate, just something that would be generally acceptable for a mainstream audience. Enthusiasts who want more performance in that regard could just get another higher-end PSVR headset. The goal should be to include something entry-level wise that can be included with every PlayStation console. I think customers would be fine with that at a $499 or even $599 price point, depending on certain other specifications.
For example, a $499 PS6 with the entry-level headset but cut-back on storage capacity, but by that point most people should have an SSD they can swap into their system. And from there a $599 model with larger storage. I think you can do this while still providing a solid performance increase in the hardware itself for non-VR games. Outside of Apple and maybe Meta, Sony/PlayStation are the only real option for VR becoming mainstream, but the VR will have to be included with all systems in some way to make that happen.
I think PS6 with VR as default can provide great gaming experience, and pretty good general usability experience, though not at the ultra-chic streamlined level Apple can offer. They could do it at a much cheaper price, though. And their gaming performance is leagues ahead of Meta's offering, I think they can continue this even with just streaming the vast bulk of the processing to a headset via Remote Play & PS Link, as long as the PS6 takes cues from Apple and has dedicated silicon for offloading & streamlining lots of parts of the VR data processing pipeline (similar to what they've done for the SSD in PS5).
An exciting future, for sure.
OP
OP
I think their discussion about how moore's law is actually dead (semiconductor prices aren't going down nearly as fast per transistor as they used to) is a hard limit on how fast VR adoption is. It feels like Apple's device is the first one where it can actually be counted as a really seamless experience which even the high end headsets today aren't capable of (not due to visual fidelity of the lenses and display, but rather due to the software environment), which is what's required for widespread adoption.
I'd be happy with a better CPU and even 15 tflops but as you say tflops has been proven not to be the most important metric. More memory might be nice too.
OP
OP
It all depends on what the pro is meant to do. If you set some design goals for the Pro:
-enhance PSVR2 performance to 90/120fps with no reprojection:
: will need a much stronger CPU, maybe 3D-stacked cache, moderate increase in GPU will help but the eye-tracked foveated rendering is already working remarkably well in most games - will probably require a redesign of the cooling system
-enhance flat games to 60fps quality mode:
: will need a much stronger GPU
-stabilize flat games with 120fps/unlocked performance modes
: will need a stronger CPU and GPU, perhaps dedicated hardware for injecting FSR3-type frame interpolation or generation or upscaling
I don't think they would want to use FSR 3 when they potentiallt got their own solution.
DLSS 3 Frame Gen work by using games frames data, FSR3 doesn't have that, it is just an Async compute to match frames output.
Sony has an chip inside Bravia allow for frame interpolation and it has kinda have few unique features, for once they actually use the Blu-ray discs frames graphics output, color gradings, other stuffs and "interpolate" just like DLSS 3.
If Sony wants they could implement it that way, the problem is input delay, PS5 don't really have a native solution to input delay as far as I known.
OP
OP
DLSS 3 isn't just frame interpolation, it's also image reconstruction using motion vector information.
D
Deleted member 223
Guest
The few things that were interesting out of that interview were confirmation bias for:
#1: the affordability problem of PC and the pitfalls of its architecture.
#2: Consoles drive the industry despite the disaster that was the PS3 generation for consoles vis a vis PC, with PS5 hardware beginning to rectify the damage in a meaningful way.
The most important of all imo, that continues this line of thought...
#3: The more console hardware is differentiated from PC, the better for the long-term health of the market for consoles. When Sony can put some specific silicon on the hardware to do specialized decompression for example.... while on PC you have to brute force elsewhere by having more RAM, a better CPU, or a GPU with more VRAM, the cost-mathematics alone just make it a winner, beyond just shifting PC hardware to become smarter about addressing the shortfall, playing catch-up to become efficient. More often that not Nvidia/AMD just pass the bill to PC whales, which compounds the affordability problem, and the chaos of the PC hardware market. Akin to "set the trend, make or force others to follow" - in this case make "indirect" competitors adjust and follow... which in more ways than one are direct competitors.
The key to to effective hardware differentiation is not just having a great hardware team and focusing solely on hardware but having a balanced view of the need for differentiated hardware as long as it's servicing a specific purpose that improves the job of developers and addresses future development needs. Mark Cerny in this regard has been a godsend to supervise and lead the process. The PS3 failed in that regard in many ways. The PS4 hardware on the other hand was a compromise that settled on the market realities as well as pressure imposed by MS's custom Windows box - aka the trojan horse, aka Xbox and PC hardware at large. After Sony lost significant consumer marketshare, and with it dev clout during the PS3 era it lost massive leverage on the market. You can look at it as the terms of defeat if you will ....to become more like PC else risk fracturing the market and potentially, not assuredly, but potentially losing that clout fight with Xbox/PC. The PS5 is the revival of consoles as king in this aspect (liberty afforded thanks to PS4's lopsided generational win) akin to PS2, PS1 and the many gens before it when consoles hardly shared the premium market with PC. It all trickles down.... eventually to software, and market-share dimensions. Xbox will continue to be the sacrificial lamb that refuses to die... by MS's will to in-part protect the Windows ecosystem.
The stupidity of Sony seeking growth on PC by supporting the platform and PC market players (Steam) on the content/software aspect works against this positive trend of revival. Completely stupid to say the least.
Bottomline, the less consoles are like PC, the better..... and the market will benefit from it. All it needs is careful stewardship of this process. Obviously any stakeholder in the PC market, including gamers invested in their PC libraries won't take that too kindly. It's what it's... irreconcilable.
#1: the affordability problem of PC and the pitfalls of its architecture.
#2: Consoles drive the industry despite the disaster that was the PS3 generation for consoles vis a vis PC, with PS5 hardware beginning to rectify the damage in a meaningful way.
The most important of all imo, that continues this line of thought...
#3: The more console hardware is differentiated from PC, the better for the long-term health of the market for consoles. When Sony can put some specific silicon on the hardware to do specialized decompression for example.... while on PC you have to brute force elsewhere by having more RAM, a better CPU, or a GPU with more VRAM, the cost-mathematics alone just make it a winner, beyond just shifting PC hardware to become smarter about addressing the shortfall, playing catch-up to become efficient. More often that not Nvidia/AMD just pass the bill to PC whales, which compounds the affordability problem, and the chaos of the PC hardware market. Akin to "set the trend, make or force others to follow" - in this case make "indirect" competitors adjust and follow... which in more ways than one are direct competitors.
The key to to effective hardware differentiation is not just having a great hardware team and focusing solely on hardware but having a balanced view of the need for differentiated hardware as long as it's servicing a specific purpose that improves the job of developers and addresses future development needs. Mark Cerny in this regard has been a godsend to supervise and lead the process. The PS3 failed in that regard in many ways. The PS4 hardware on the other hand was a compromise that settled on the market realities as well as pressure imposed by MS's custom Windows box - aka the trojan horse, aka Xbox and PC hardware at large. After Sony lost significant consumer marketshare, and with it dev clout during the PS3 era it lost massive leverage on the market. You can look at it as the terms of defeat if you will ....to become more like PC else risk fracturing the market and potentially, not assuredly, but potentially losing that clout fight with Xbox/PC. The PS5 is the revival of consoles as king in this aspect (liberty afforded thanks to PS4's lopsided generational win) akin to PS2, PS1 and the many gens before it when consoles hardly shared the premium market with PC. It all trickles down.... eventually to software, and market-share dimensions. Xbox will continue to be the sacrificial lamb that refuses to die... by MS's will to in-part protect the Windows ecosystem.
The stupidity of Sony seeking growth on PC by supporting the platform and PC market players (Steam) on the content/software aspect works against this positive trend of revival. Completely stupid to say the least.
Bottomline, the less consoles are like PC, the better..... and the market will benefit from it. All it needs is careful stewardship of this process. Obviously any stakeholder in the PC market, including gamers invested in their PC libraries won't take that too kindly. It's what it's... irreconcilable.
Last edited by a moderator:
And most of that went to rez boost. PS5 generally sharp enough, more power can be deployed for better pixels, not necessarily more.
