vdopey
Member
Cerny himself stated they were targeting and struggling to hit 2ghz.
where did he say this ? Proof?
Cerny himself stated they were targeting and struggling to hit 2ghz.
where did he say this ? Proof?
To be honest and if mods are honest , all this bad yield stuff in relation to that leak is starting to look like a possibility... but I am very interested what went on ..
No console warrioring at all , and if you have open discussions, my opinion is that this also a topic for a discussion.
But that’s me .
Maybe they did, maybe they didnt its all speculation on your part right now, you have no idea what their plans are / were 11 million units in launch window is no slouch and to be honest if they already have 10 million to ship, they have time to produce more.
We have no idea what the xbox yields are right now, they probably have a yield rate failures but they are repurposing for the xss most likely, ultimately what it will come down to is sales, we will know how either have fared this time next year, thats when the dust will have settled in all of this.
Huh? That makes no sense. Why would they design new one when they could use leftover chips. Were they afraid that there will not be enough of faulty chips to use or what?It's a smaller die with significantly cut down CU counts.
It's not just an XSX die with disabled CUs.
XSX die is 197mm² where as the XSX is something like 360mm².
In that dreary marketing speech named "road to PS5".
I care not to sit through it again to find the exact time stamp.
What a bunch of bullshit.The Series X doesn't have a frequency that high, so I'm sure they are more steady.
Seems that Sony panicked when they discovered that MS was 12 teraflops and boosted from 9 to 10 teraflops in a desperate last effort. And the yields are suffering from it.
It certainly is weird that so many yields are falling and it was certainly weird that the most reliable leaks where claiming a 9 teraflop PS5.
To be honest and if mods are honest , all this bad yield stuff in relation to that leak is starting to look like a possibility... but I am very interested what went on ..
No console warrioring at all , and if you have open discussions, my opinion is that this also a topic for a discussion.
But that’s me .
It also baffles me how MS seems to have the hardware stable as supposed to Sony a hardware company, but that MS dropped the ball on Software, a software company.
Interesting start of a new gen. Would make great book material for sure .
Making shit up should result in a banIn that dreary marketing speech named "road to PS5".
I care not to sit through it again to find the exact time stamp.
Huh? That makes no sense. Why would they design new one when they could use leftover chips. Were they afraid that there will not be enough of faulty chips to use or what?
Anyway I stand corrected.
Like I said, very old test data. There probably was no such thing as RDNA 2 back then.
The number of PS5 units made is huge. I don't see any problem with that. I doubt they will have problems in the future.
As other used has pointed out, there's always this claims of "not enough consoles" so people run to the stores to buy one. Marketing 101.
The leak is from May 2019, and contained miscellaneous test data potentially spanning about a year worth of tests on several different chips.Tests are from May 2019 IIRC. Yes, RDNA 2 exists surely
He said no such thing in that.
Making shit up should result in a ban
Here you go NPCs.
Back to the spec thread circle jerk with you.
What a bunch of bullshit.
Such significant change in frequency would delay production for at least 6 months for testing purposes, since complex chips at 7nm take 4-5 months to manufacture.
If you aren't willing to inform yourself about things please stop spewing nonsense and FUD.
Frequency is high because there is smartshift to allow it. There would be zero need to add smartshift if the goal was to keep the frequency under 2ghz.
Or Smartshift was also a last minute decision in response to MS??
There would be zero need to add smartshift if the goal was to keep the frequency under 2ghz.
Cerny himself stated they were targeting and struggling to hit 2ghz.
It certainly is weird that so many yields are falling and it was certainly weird that the most reliable leaks where claiming a 9 teraflop PS5.
MS already did it this gen with the Xbox One and it was much much more later on than what we are arguing here.
Stop trolling.Every console design it's always designed with room to improve in "the last second". MS already did it this gen with the Xbox One and it was much much more later on than what we are arguing here.
I mean, how many months have passed already since the official PS5 specs reveal? Oh, 6 months already!
What?
I thought smartshift was the future and the superior choice design.
Bad yields are a historical reality in chip manufacturing. But hey, don't let facts get in the way of some FUD peddling.
Bloomberg surely implies that Sony can't produce PS5 that fast and that much. That's not a problem with console. 11 mil. is above their previous target, though
Sony surely planned variable clocks from beginning.
But they already have faulty XSX chips. They don't need to produce them. They are side product (or actually waste) of XSX chips production. That makes sense only when they expect to sell more way more XSS than XSX consoles.More chips per wafer.
At 360mm you might have 79 chips on a wafer, some of which won't be usable. At 197 you could virtually double potential chips per wafer while having more of them be usable.
This. If you are not flexible, the situation is even worse than in 2013 and if they let's say make it 70:30: Have fun.So the real question is what's the split between disc and discless?
AMDs first foray into 7nm had them getting 70%+ yields
50% is bad. 50% and requiring the downscaling of production is even worse.
Bad yields are a historical reality in chip manufacturing. But hey, don't let facts get in the way of some FUD peddling.
What MS did is last minute. Sony surely planned variable clocks from beginning. Their cooling solution is patented in 2018 for that. So no, this is isn't last minute thing. Cut the crap
Stop trolling.
It's neither better or worse.
Those were pure cpu wafers. APUs are more complex since they incorporate gpu designs and the interconnection buses. Have you ever noticed zen apu's use previous generation zen designs instead of the cutting edge?
On top of that these are the first runs of RDNA2.
I think you answered your own questionBut they already have faulty XSX chips. They don't need to produce them. They are side product (or actually waste) of XSX chips production. That makes sense only when they expect to sell more way more XSS than XSX consoles.
Seems that Sony panicked when they discovered that MS was 12 teraflops and boosted from 9 to 10 teraflops in a desperate last effort. And the yields are suffering from it.
It certainly is weird that so many yields are falling and it was certainly weird that the most reliable leaks where claiming a 9 teraflop PS5.
MS armchair devs have also been shouting "but VRS" non stop as if it was better than chocolate and sex and so far we haven't seen anything even worth mentioning...As low as a 50% is a disaster, no matter how hard you try to damage control it.
I am not the one saying is a last minute thing. It's people claiming that they can't change things at "the last minute".
And I'm just saying: A) things can be changed at the last minute and B) Sony didn't change things at the last minute, but months and months ago.
It's no secret that both companies react to each other. And a 12 teraflop Xbox vs a 9 teraflop PS5 would have been too much of a difference and a PR disaster for Sony.
I can easily see how Sony would boost the PS5 frequency as much as the console design could to shorten the gap once they knew the Xbox Series X was 12TF.
They didn't redesign the entire console.
But that's not what the pro Sony armchair developers of GAF have been saying since it was announced that the PS5 had smartshift.
They say is the superior design choice and that is the future for next hardware. Some even claimed that the next Xbox is going to have it.
Sony's always the best, you see.
There you go buddy, like i said: The number of the devil. See you Sep 29, 2020I know you can't say that the PS5 is *censored*TF on GAF.
But that's not what I'm saying. I'm saying it's 10TF because they boosted the frequency to shorten the disadvantage against MS 12TF. And that boosts seems to have come back to bite them in the ass with the yields falling at an espectacular rate.
I'm actively suppressing that though in my mind.I think you answered your own question
15 or 11 or whatever, those first 20 million units of Xboxes or Playstations will fly as hell. At any price.
"Running a GPU at 2GHz was looking like an unreachable target, with the old fixed frequency strategy..."You are twisting words
Launch, launch window, it's all the same and the point still standsIt's not supposed to be 15 million at launch... You need to read things, bud.
This was quite literally proven wrong by Sony themselves 5 posts above... A post you even quoted...
No. 11m till march is not above the 10m till december (end of 2020, not fiscal year 2020).
Like I also wrote in the past, I wouldn't surprise me if they have production issues.
Small chip (ok that is not really a risk) but with high frequencies and a fixed power-target are many variables
Produced 10 mil. till end of 2020, shipped 10 mil. till end of March. Read it
AMD Ryzen 4000 Mobile APUs: 7nm, 8-core on both 15W and 45W, Coming Q1
www.anandtech.com
AMD APU yields were already starting off at 87%.
Granted this is at a small die size, but even at 360mm² this wouldn't be much worse than 70% yields. Sony getting 50% is just plain bad.
How does this in any shape or form suggest that Sony opted for a smartshift solution at the last minute?"Running a GPU at 2GHz was looking like an unreachable target, with the old fixed frequency strategy..."
Not to mention the fact, they indeed were testing at 2GHz as known by the GitHub leaks.
I could read it a thousand times, doesn't change the fact that Sony clearly wasn't aiming for variable clocks from the start.