Old supposed leak being discussed on Reset and Gaf. Seems fairly accurate in regards to XBSX specs but I'm not sure about the studio acquisitions as I don't really follow that. It was from Jan. 2019.
There's a slight difference in the fact that you are in control of your own vacation, he is not in control of what Sony does, nor does he know exactly what Sony has planned. His information can be dated, yours is always current.
I find his comments contradictory:
o'dium - verified developer at resetera said:
Man, the 9tf dream is years old. Sony ditched that a long time ago. That was their original plan, to release a 9tf machine sometime in 2019. But those plans changed in 2018, so they could release a faster unit in 2020. They have had a while to work on things, it’s not a last minute change. People need to let go of the 9tf stuff, it’s pure fantasy.
- The way he ends with 'pure fantasy' raised a red flag for me because he had just wrote that 9TF system was real the sentence before.
- So if I'm following along, it's not fantasy at all, at one point in time it was pure reality according to him.
- That's unfortunate because we can prove that because Github leaks indicate exactly that and this is information coming from him after the fact
- This leads me to ask: so why deep into 2019 is Sony still testing this expired 2018 chip?
- I'm going to ignore my bias against the fact that I don't believe many studios would be peddling this type of information around, strategic insights into Sony would usually be 1 degree from Cerny.
- But it's also clear his contact to the PS5 kit is not direct. Nor does he have direct 1 degree access from Cerny; so how would a studio go about knowing so much about what Sony is trying to do with it's hardware
- Where i worked they just gave us a kit, not a story to go along with it.
2) But those plans changed in 2018, so they could release a faster unit in 2020
It's based upon how you read this sentence.
1) if you use could as the past tense of can, then this is a statement, the statement being they changed the plans in 2018 so that they can release a newer more powerful model in 2020.
2) if you read it another way: the plans changed in 2018, so they _could_ (as in possibility) release a faster unit in 2020 - and since he's not exactly sure about the concrete values, he doesn't know how the values have changed. That doesn't mean they built a ground up new chip. It could just mean clocking or thermal changes, which he talks about, which is talked about by others before his post.
I'm not saying his information is bad though;I'm not going to judge people who want their box to be the most powerfull etc.
I just wished it was posted on DF or DF made these announcements. I would have way less questions than I do now about it.
you're right I don't have to. I mean I could have sat back and said squat.Or maybe you don't need to over analyse someone talking colloquially on a message board? I've posited that perhaps the earlier chip designs were found to be useful for ealy devkits and/or backwards compatibility testing even after being superseded.
If someone wants to buy in to the github leak 100%, who gives a shit? Person A says it's wrong and PS5 is X. Person B says no it's what's in the leak. Neither of these opinions matter. It's a videogame player, and the official specs will come out eventually. If someone ends up being wrong, it doesn't matter, unless they were being an asshole about it. Then we'll all laugh at them after the fact. It's fun to bounce opinions back and forth, but if people are getting defensive about their positions then it's probably time to put their big boy pants on and step away.
If someone wants to buy in to the github leak 100%, who gives a shit? Person A says it's wrong and PS5 is X. Person B says no it's what's in the leak. Neither of these opinions matter. It's a videogame player, and the official specs will come out eventually. If someone ends up being wrong, it doesn't matter, unless they were being an asshole about it. Then we'll all laugh at them after the fact. It's fun to bounce opinions back and forth, but if people are getting defensive about their positions then it's probably time to put their big boy pants on and step away.
There's a slight difference in the fact that you are in control of your own vacation, he is not in control of what Sony does, nor does he know exactly what Sony has planned. His information can be dated, yours is always current.
The github shows only rasterization tests on a GPU. There's no proof that the chip being tested is anything other than a Navi 10 with just as much enabled units as you need for PS4 Pro backwards compatibility.This leads me to ask: so why deep into 2019 is Sony still testing this expired 2018 chip?
right; fair enough.I see what you’re saying, but largely the internet is full of bad English, so picking on this to disprove his info seems a bit harsh.
I’m not saying he’s legit, but....
it’s not that they believe it, it’s just for every fecking insider who says something different to github the same defence line of ‘but github proves otherwise’
it’s not that they believe it, it’s just for every fecking insider who says something different to github the same defence line of ‘but github proves otherwise’
I really hope there's some hardware ML capabilities in their on both consoles. It's the one last unknown for a feature that would be great in consoles and it would suck if we have ~7yrs without it as a part of the 'generational baseline'.
Having RT/VRS/ML + SSDs in consoles and PC will really get the ball rolling on all those technologies; and make them standard much quicker.
When you're all talking about ML, is it just making use of FP16?
For deep learning; yes. A large majority of it is FP16.When you're all talking about ML, is it just making use of FP16?
Try making a simple search on how many hundreds of times @PSman1700 wrote the PS5 is 9TF / 9.2TF in different ways. It's an account with 11 months and >1300 posts, the massive majority of them saying practically the same thing, in the same sub-forum and in the same thread.
We are almost a year form the launch!
If I wanted to find leakers I'd give every studio with a dev kit slightly different versions of the target specs.
Cover that in the article by limiting info to ballpark figures and generalisations. At this point, the forecasting of PS5's hardware just needs to be moved off the largely accepted 9.2 TF target if that's the reality. So OsirisBlack can go to DF, present the evidence, and have DF tell the world it isn't 9.2 TFs and PS5 is faster than that.
I agree. I've considered two different threads, with one policed better for open-ended discussion, but it'd be a lot of work. At the end of the day, no opinions expressed here matter, and if some like to keep repeating their mantra ad nauseum, feel free to ignore them. Proper discussion about next gen can be conducted in the Tech forum around particular theories without recourse to lean on internet rumour-mongers.
For those for whom these rumours matter and who are very vocal in their support of a viewport, there will be a reckoning when the truth outs. Much humble pie will be served up somewhere or other.
Great in theory, but certain people just can't help but be aggressively insistent that their opinion is actually verified fact.
you have serious problems with hard data,
NVIDIA doesn't optimize OpenCL anymore, relying instead on CUDA alone. A Vega 64 is 2.5 faster than GTX 1080 in LuxMark, beating the 1080Ti by 50%. LuxMark is in no way indicative of any ML performance.By the way, Radeon VII scored about 1.62 times the GeForce RTX 2080 in "Luxmark" which utilizes an OpenCL-based GPGPU-like ray tracing renderer.