Can people please stop antagonizing Klee? It's pretty damn gross already.
Can people please stop antagonizing Klee? It's pretty damn gross already.
Good find, it's VR related from what i can see.Hey guys, saw there was a new patent by Mak Cerny published 2nd January 2020.
The PS3 devkit had 3 different gpus before final siliconlmao...yeah the timing/logic does not seem to fit. It is almost as some believe Sony has GPU's lying around of different performance metrics and are pulling them out of a security vault and changing them on the fly based on "what they heard would be in the next Xbox". These things take a shit ton of time to design, test, validate etc. It is not something that can change on the fly. Clock speeds can depending on the cooling solution...but how some react thinking that Sony or MS can just magically spin up a new GPU on the fly is laughable.
Can people please stop antagonizing Klee? It's pretty damn gross already.
I think his theory is, and frankly it sounds very reasonable to me, that Sony from the start wanted ~ 8TF PS5 and when they found out Anaconda will be ~12TF Sony decided to raise the clocks (cause they didn't have any other option after 36CU design was locked) in order to close that gap a bit, right DrKeo ? :)
Can people please stop antagonizing Klee? It's pretty damn gross already.
I think his theory is, and frankly it sounds very reasonable to me, that Sony from the start wanted ~ 8TF PS5 and when they found out Anaconda will be ~12TF Sony decided to raise the clocks (cause they didn't have any other option after 36CU design was locked) in order to close that gap a bit, right DrKeo ? :)
Hey guys, saw there was a new patent by Mak Cerny published 2nd January 2020.
Good find, it's VR related from what i can see.
"Patent History
Publication number: 20200005431
Type: Application
Filed: Sep 13, 2019
Publication Date: Jan 2, 2020
Inventor: Mark Evan Cerny (Burbank, CA)
Application Number: 16/570,258"
anexanhume you saw this one?
Can people please stop antagonizing Klee? It's pretty damn gross already.
I really don't want to open this argument all over again but my strong impression,considering the info we have right now, is this:That only makes sense if they were aiming for 2019, which doesn't really make sense from a business perspective since (A) PS4 sales have still been very strong, and (B) would indicate that they thought they needed to pre-empty MS, which again makes no sense at all. I mean you are basing it on the contextless Github leaks plus the unsubstantiated rumour that Sony was aiming for 2019 release. I mean if they both turn out to be true, it shows a collosal business fuck up by Sony. Neither 2019 makes sense, nor simply uping the clocks to 2ghz, unless they have an innovative cooling system. If the latter they may have aimed for 2ghz, and they may have budgeted for a lower retail price than XSX. However, if they did change direction in 2017 as Klee indicated, that would give them time re-design the system. It would also line up with the idea that they may be using the initial design up clocked to simulate the final hardware, but final dev was even later than XSX (which already seemed pretty late).
If they wanted to sell PlayStation 5 at 399$/€ at launch, with only a very small loss on each unit (<20$), what was the best option?Because give the options of getting to 9.2tf, why on earth would they pick a 36cu high clock scenario? Given the general assumption on power,/heat, it doesn't make sense.
Remember that PlayStation 4 sales during 2018 were above Sony expectations.That only makes sense if they were aiming for 2019, which doesn't really make sense from a business perspective since (A) PS4 sales have still been very strong
VR related mostly but I guess it can also apply to games with 90degree or larger fields-of-view.Hey guys, saw there was a new patent by Mak Cerny published 2nd January 2020.
If they wanted to sell PlayStation 5 at 399$/€ at launch, with only a very small loss on each unit (<20$), what was the best option?
Cell failed to compete Xenon in almost any spec other than FP. And honestly sky high FP number isn't that necessary for CPU and that's why CELL never successed.Don't think cell was weaker than the 360 cpu .
It was just more of a pain in the ass to program for .
But it has been a long time and i can be wrong .
Well it seem no one want to be the person to leak hard PS5 specs since no one is doing it.
But if 40/36 CU design was locked already what can they do?!? The only option is to raise the clocks.That's the point.
Klee loves the attention.
He's even thinking about starting a youtube channel.
I want to go back to my roots, reviewing games and games hardware on YouTube just like I did 20+ years ago. Only this time from the perspective of an older, lifetime gamer.
As a busy father and grandpa I want to talk about the struggles of balancing your Spouse/kids/grandchildren with gaming.
There are a lot of young couples who still are interested in gaming but important duties like maybe changing diapers or feeding or work or family play puts games on the back burner.
I really don't want to open this argument all over again but my strong impression,considering the info we have right now, is this:
if we conveniently choose to ignore all the raw 36CU testing data we got during entire 2019 from apisak,komachi,that github leak etc. etc. the only other alternative is some new,hidden PS5 APU (with more then 40/36CU) waiting in the shadows to be announced at PS Meeting. I think that's an illusion.It is what it is.
Yeah, that definitely sounds like him loving the attention and trying to make a YouTube channel off of it when it was his wife who suggested it in the first place. 🙄 Reported.
I really don't want to open this argument all over again but my strong impression,considering the info we have right now, is this:
if we conveniently choose to ignore all the raw 36CU testing data we got during entire 2019 from apisak,komachi,that github leak etc. etc. the only other alternative is some new,hidden PS5 APU (with more then 40/36CU) waiting in the shadows to be announced at PS Meeting. I think that's an illusion.It is what it is.
How can i possibly know that? It might be business decision or some technical issue.We are all speculating here. Maybe Zen2/Navi 10 design wasn't fully ready for mass production.Maybe Sony wasn't able to sell PS5 in 2019 at the price they wanted and it is possible to do that in late 2020? Maybe their BC solution wasn't fully ready? Maybe their next gen games weren't ready? Who knows.You still didnt answer the question, why delayed to 2020 when there is already a gpu that performs beter then 8 tf from amd on the market.
But if 40/36 CU design was locked already what can they do?!? The only option is to raise the clocks.That's the point.
I think you are taking this information from GitHub as almost the end all/be all for the PS5 when none of us even know the full story. We know the numbers are real, but your conclusion flies on the face of everything we've heard up until this point. If I had to a gun to my head right now, I would be on the PS5 being stronger than 9.2TF.
How can i possibly know that? It might be business decision or some technical issue.We are all speculating here. Maybe Zen2/Navi 10 design wasn't fully ready for mass production.Maybe Sony wasn't able to sell PS5 in 2019 at the price they wanted and it is possible to do that in late 2020? Maybe their BC solution wasn't fully ready? Maybe their next gen games weren't ready? Who knows.
DrKeo made last week some good posts about that whole topic.Why would it be locked though?. They would have known from the Navi testing what the perf and thermals where likely to be.
Good find, it's VR related from what i can see.
"Patent History
Publication number: 20200005431
Type: Application
Filed: Sep 13, 2019
Publication Date: Jan 2, 2020
Inventor: Mark Evan Cerny (Burbank, CA)
Application Number: 16/570,258"
anexanhume you saw this one?
Wait, holy shit this is a huge find. Doesn't this support that Sony could be going with their own solution instead of AMD's? I could have sworn we talked about this before.
VR related mostly but I guess it can also apply to games with 90degree or larger fields-of-view.
It's how you sub-section the image for rendering tiles and by doing it in a non-uniform way (edges using smaller tiles?) is the most effective way. This is not exactly the same as foveated rendering however I believe (we are not talking lower resolution in the edge tiles here necessarily? I just skimmed).
But if 40/36 CU design was locked already what can they do?!? The only option is to raise the clocks.That's the point.
Again. This requires buying into the idea that Sony didn't have time to change things, which seems unlikely given the delay would have given them another 12 months.
Are you also ignoring Matt's statement on the Github leaks?
Again, why assume they were targeting 9.2TF?Because give the options of getting to 9.2tf, why on earth would they pick a 36cu high clock scenario? Given the general assumption on power,/heat, it doesn't make sense.
Yes, it's a theory. It doesn't mean it's true, but saying that it's impossible is basically ignoring every console launch ever.I think his theory is, and frankly it sounds very reasonable to me, that Sony from the start wanted ~ 8TF PS5 and when they found out Anaconda will be ~12TF Sony decided to raise the clocks (cause they didn't have any other option after 36CU design was locked) in order to close that gap a bit, right DrKeo ? :)
Increasing your clocks by 11% isn't 100$ more expensive and isn't that far fetched. Xbox One increased CPU clocks by 9.3% and GPU clocks by 6.6% just before launch, do you think they've spent extra 100$ in order to do that?That makes no sense at all. You get slight increase in performance for potential $100 increase in price, ultimately failing to come anywhere close to the competitor in power and match them in price.
Again, it comes back to the idea that Sony had no idea what Microsoft would be aiming for or what the road map would look like for a 2020 device. I don't buy it.
Even if we believe the rumours that they delayed, that would have still left time to sit down with AMD and grab something new. Regardless, I think Matt's statement around the Github leaks says enough to assume it's pretty unlikely.
That's not really true. Sony had designed the PS5 around their budget and vision, not according to what 700$ GPUs are available in 2020. Case and point, the PS4. It had a 1.84TF GPU while Sony knew AMD will have a 5.6TF GPU in the market before PS4 will launch. If you compare it to a 9.2TF GPU in 2020, it's the same as if AMD will have a 550$ 27TF GPU in 2020, which it won't.That only makes sense if they were aiming for 2019, which doesn't really make sense from a business perspective since (A) PS4 sales have still been very strong, and (B) would indicate that they thought they needed to pre-empty MS, which again makes no sense at all. I mean you are basing it on the contextless Github leaks plus the unsubstantiated rumour that Sony was aiming for 2019 release. I mean if they both turn out to be true, it shows a collosal business fuck up by Sony. Neither 2019 makes sense, nor simply uping the clocks to 2ghz, unless they have an innovative cooling system. If the latter they may have aimed for 2ghz, and they may have budgeted for a lower retail price than XSX. However, if they did change direction in 2017 as Klee indicated, that would give them time re-design the system. It would also line up with the idea that they may be using the initial design up clocked to simulate the final hardware, but final dev was even later than XSX (which already seemed pretty late).
Matt already said to move past from the GitHub info and people still cling to the 36CU narrative.
Regarding "Navi was ready in 2019", no it wasn't. Even after your chip is ready design-wise, there is a lot of work to be done in order to get it in mass production shape. You can't finalize huge things like CU count just a few months from release. PS5 is going into mass manufacturing in probably 3-4 months, their GPU design was locked months ago if not over a year ago.
Regarding "Navi was ready in 2019", no it wasn't. Even after your chip is ready design-wise, there is a lot of work to be done in order to get it in mass production shape. You can't finalize huge things like CU count just a few months from release. PS5 is going into mass manufacturing in probably 3-4 months, their GPU design was locked months ago if not over a year ago.
Its baffling. People took the GitHub information and ran wild with it as of it was the smoking gun. Then when Matt came in and said he was glad people moved past the info being confirmation for anything, people still wanted to believe PS5 is 9.2TF.
That's not really true. Sony had designed the PS5 around their budget and vision, not according to what 700$ GPUs are available in 2020. Case and point, the PS4. It had a 1.84TF GPU while Sony knew AMD will have a 5.6TF GPU in the market before PS4 will launch. If you compare it to a 9.2TF GPU in 2020, it's the same as if AMD will have a 550$ 27TF GPU in 2020, which it won't.
Regarding "Navi was ready in 2019", no it wasn't. Even after your chip is ready design-wise, there is a lot of work to be done in order to get it in mass production shape. You can't finalize huge things like CU count just a few months from release. PS5 is going into mass manufacturing in probably 3-4 months, their GPU design was locked months ago if not over a year ago.
Wait, holy shit this is a huge find. Doesn't this support that Sony could be going with their own solution instead of AMD's? I could have sworn we talked about this before.
Its baffling. People took the GitHub information and ran wild with it as of it was the smoking gun. Then when Matt came in and said he was glad people moved past the info being confirmation for anything, people still wanted to believe PS5 is 9.2TF.
Its baffling. People took the GitHub information and ran wild with it as of it was the smoking gun. Then when Matt came in and said he was glad people moved past the info being confirmation for anything, people still wanted to believe PS5 is 9.2TF.
But we surely have the 5700XT though...Regarding "Navi was ready in 2019", no it wasn't. Even after your chip is ready design-wise, there is a lot of work to be done in order to get it in mass production shape. You can't finalize huge things like CU count just a few months from release. PS5 is going into mass manufacturing in probably 3-4 months, their GPU design was locked months ago if not over a year ago.
Regarding "Navi was ready in 2019", no it wasn't. Even after your chip is ready design-wise, there is a lot of work to be done in order to get it in mass production shape. You can't finalize huge things like CU count just a few months from release. PS5 is going into mass manufacturing in probably 3-4 months, their GPU design was locked months ago if not over a year ago.