-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
Albert Penello - Impossible for changes to the die to be made 12 months before release.
- Thread starter Rightisright
- Start date
- News Hardware Platform
MoreJRPG
Suffers from extreme PDS
I mean, you’re the one that seeks out my posts all over the forum so who really is the doofy one? I remember when I needed that spice in my life too. You’ll get there, one day.You're doofy as fuck bro.
My stock portfolio always thanks me.
DeepEnigma
Gold Member
Relax narcissist, you're not that special.
I follow threads that interest me, and enjoy quoting juvenile shit to troll right back.
Last edited:
alabtrosMyster
Banned
Core frequency: the xbox one went through this 4 or 5 months before launch, so it's definitely possible, however it must cost a lot of money to re-validate all these machines (some will fail, it takes time, etc.)
Adding CUs:
I'm not in this situation, but as Penello pointed out, there are two scenarios, enable the redundancy CUs that are already on the DIE for redundancy (many will be defective, this again means that all previously built consoles will have to be re-validated, and this is much more likely that many will fail, to compensate they could consider down-clocking so that more parts pass QA)... my bet is that system architects know exactly how many are likely to pass/fail that effort if they decide to go down that road.
If you go above the 40CUs (assuming a 36CU chip that has 4 CUs for redundancy), new SOC chips would need to be planned and made, a new board layout and cooling solution could be needed as well (or a higher tolerance to high pitch noise on the part of your customers), and all the consoles that are already manufactured would all need to be scrapped, new dev kits could be needed as well. There would be a delay of at least a couple of months.
None of the options are easy, but they are all possible, it just depends how much that extra power is worth in therms of units sold once the product is out, if you think it will make any difference at all to your target audience (Nintendo would argue that it doesn't always make any difference at all).
Dural
Member
8GB wasn't planned from the beginning for PS4 as it wasn't likely to be available for the launch. Sony got lucky in that 4gb chips became available allowing them to double the ram without doing anything. Imagine if the PS4 shipped with 4GB of ram compared to the 8GB of DDR3 in the X1, might have made things a bit more interesting.
Gavon West
Spread's Cheeks for Intrusive Ads
I'll take that bet. What's the wager?
Thaedolus
Gold Member
When I say late I mean almost a year before launch. The Saturn wiki cites Next Gen on that- not sure how reliable that is but it matches up pretty closely with the story I’ve always heard about SEGA designing the Saturn and how reacting to Sony by throwing in extra hardware that it wasn’t necessarily built around made it a difficult system to properly utilize at the time. But hey maybe that’s not correct
Gavon West
Spread's Cheeks for Intrusive Ads
They can make a few smaller tweets before mass production but that's about it -- not without delaying the official release of the PS5 or taking a huge financial hit for hardware changes this late in the game.
I'm not proficient in this tech stuff but I understand that much.
I'm not proficient in this tech stuff but I understand that much.
Nikana
Go Go Neo Rangers!
I want in on this
diffusionx
Gold Member
Designing a new SOC and scrapping all the work you did while delaying the system goes far beyond the bounds of possible IMO. It's just not an option. The fact is it doesn't make any difference. Even if the Xbox One was totally redesigned to be equivalent to PS4, and came out six months later, it still would have gotten its ass kicked, probably in much the same way, at the cost of untold millions.
Even if the PS5 and Xbox leaked specs are right on, the system is in a much better place than the Xbox One was.
It really depends what "late" means. It could mean, "late in the development", as opposed to the system being revamped when it was being sent out to devs and manufacturing was spinning up, which is what "late" would mean in the context we are talking about now.
Last edited:
Hobbygaming
has been asked to post in 'Grounded' mode.
Cool
you and the other guy have to change your avatars to pictures of Ghost of Tsushima's official box art for 1 month if I'm right and if you're right I'll let you choose minePS4 had a 50% GPU advantage but we can lower it to 45%
If PS5 is 45% less powerful than the Series X then you win
LordOfChaos
Member
He's not wrong here. Tapeout would be done. It takes several months to prepare a set of photomasks and build an actual mass production chips after.
What can be changed after tapeout is final clocks, the other possibility is that two chips are prepared simultaneously for tapeout and one is chosen based on market trends/competition.
What can be changed after tapeout is final clocks, the other possibility is that two chips are prepared simultaneously for tapeout and one is chosen based on market trends/competition.
joe_zazen
Member
evidence points toward summer /autumn of 2018, so 2 or 2.5 years.
And who says they scrapped all that work? It looks like under powered junior consoles are on the way.
StreetsofBeige
Gold Member
Penello basically doubling down on PS5 being weak sauce.
Gavon West
Spread's Cheeks for Intrusive Ads
If we're speaking on percentages as in the start of this generation with base Xbox One and Ps4, then we have to change things a bit. The percentage in power between Ps4 and Xbox One were miniscule in comparison to, say, the One X and the PS4 Pro and where 45% more power (using 45% as an example - not the actual power metric) was actually quite substantial in the name of actual teraflops in respect to both machines.Cool
PS4 had a 50% GPU advantage but we can lower it to 45%
If PS5 is 45% less powerful than the Series X then you win
Vis a' vis, if, say, Series X is %15 more powerful but that's communicated in teraflops not in some meager metric like we had between base Xbox One and Ps4 (games being 720p vs 1080p) - I win. But if the difference is akin to comparisons such as the sort I stated - you win!
If that makes sense. Cause I'd bet the farm the power difference is going to be pretty big. Im speaking specifically about power here.
Last edited:
Hobbygaming
has been asked to post in 'Grounded' mode.
Ok so 1.84 vs 1.31 was a 40% compute advantage PS4 vs Xbox OneCool
PS4 had a 50% GPU advantage but we can lower it to 45%
If PS5 is 45% less powerful than the Series X then you win
If Series X has less than a 40% compute advantage (teraflops) then I win
Gavon West
Spread's Cheeks for Intrusive Ads
This is what I mean. There's a huge difference between 1.84 vs 1.31 and 6TF vs 4Tf or 12tf vs 9tf.Ok so 1.84 vs 1.31 was a 40% compute advantage PS4 vs Xbox One
If Series X has less than a 40% compute advantage (teraflops) then I win
1.84 vs 1.31 was small and only displayed in resolution ie 900p vs 1080p etc. Not a big deal in my opinion.
Where as the differences between the One X and the Ps4 Pro was much more pronounced as 6tf > 4tf and the X typically had the graphical/performance advantage. The percentage in power doesnt tell the whole story as 2 whole teraflops is a pretty substantial metric, regardless of how small the percentage number. It just doesnt tell the whole story.
Hobbygaming
has been asked to post in 'Grounded' mode.
We have to use the difference of the overall boxes, the TFLOPs that's the best metric to use as it eliminates our opinions of 900P vs 1080P, 4K CB vs 4K native, etc
Last edited:
Nikana
Go Go Neo Rangers!
Ok so 1.84 vs 1.31 was a 40% compute advantage PS4 vs Xbox One
If Series X has less than a 40% compute advantage (teraflops) then I win
Yeah I thought you meant a 700 gflops lol.
darkinstinct
...lacks reading comprehension.
It's pretty simple: You only revise chips 12 months before launch (Oberon to E0) if you intend to use them. If that magical 12 TF PS5 APU exists, why keep revising a 9.2 TF APU for the devkit? If anything devkits are faster than retail units because of profiling tools.
That leaves two options. 1: Oberon is the base PS5 while a much more powerful PS5 Pro launches with it. Because it doesn't make any sense to launch a 9.2 and a 12 TF SKU. The Pro would be closer to 18 TF, which would be a massive die. You are basically looking at a 399 PS5 with a 799 PS5 Pro. Or 2: Oberon is PS5 and that's it.
Microsoft seems very confident that they are the performance leader next gen. Sony has been quiet about making any such statements since April, instead focussing on SSD speed and the controller. And if you look back, Microsoft has had more GPU power on every single console except Xbox One. They are the ones with a dual SKU strategy. Which makes no sense at all if PS5 would be faster than XSX. One SKU is cheaper, the other more powerful. And PS5 sits in the middle.
Also people that expect Sonys raytracing to make up the difference are delusional. An integrated solution is always superior. Plus Microsoft was the driving force behind RTX. To think that they are now anything but the better raytracing solution is crazy. And for good reason.
It's becoming more and more apparent that Sony did indeed target late 2019 first. Which means they would've designed PS5 before RTX and DXR became a thing. Then Microsoft presented DXR in March of 2018 - and Sony realized that Microsoft was going for raytracing in next gen consoles. At that point they scrambled. And looked for a way to implement raytracing in their already designed APU. This lead to them cancelling their 2019 launch.
Oh and anybody believing the "Oberon is just the devkit" dumbness has to ask themselves why Sony doesn't just do what Microsoft did, give target specs and use existing hardware (like an overclocked 5700XT) to develop until the actual APU is ready. Instead, Sony first sent GCN based devkits, then replaced them with 8TF Navi devkits, then replaced them with 9.2 TF devkits, then iterated that chip four more times, only because they have no intention of using it because they are actually working on a 12 TF chip. And why, just why would you test BC modes on a pure devkit APU? Because devs want to play BC games?
That leaves two options. 1: Oberon is the base PS5 while a much more powerful PS5 Pro launches with it. Because it doesn't make any sense to launch a 9.2 and a 12 TF SKU. The Pro would be closer to 18 TF, which would be a massive die. You are basically looking at a 399 PS5 with a 799 PS5 Pro. Or 2: Oberon is PS5 and that's it.
Microsoft seems very confident that they are the performance leader next gen. Sony has been quiet about making any such statements since April, instead focussing on SSD speed and the controller. And if you look back, Microsoft has had more GPU power on every single console except Xbox One. They are the ones with a dual SKU strategy. Which makes no sense at all if PS5 would be faster than XSX. One SKU is cheaper, the other more powerful. And PS5 sits in the middle.
Also people that expect Sonys raytracing to make up the difference are delusional. An integrated solution is always superior. Plus Microsoft was the driving force behind RTX. To think that they are now anything but the better raytracing solution is crazy. And for good reason.
It's becoming more and more apparent that Sony did indeed target late 2019 first. Which means they would've designed PS5 before RTX and DXR became a thing. Then Microsoft presented DXR in March of 2018 - and Sony realized that Microsoft was going for raytracing in next gen consoles. At that point they scrambled. And looked for a way to implement raytracing in their already designed APU. This lead to them cancelling their 2019 launch.
Oh and anybody believing the "Oberon is just the devkit" dumbness has to ask themselves why Sony doesn't just do what Microsoft did, give target specs and use existing hardware (like an overclocked 5700XT) to develop until the actual APU is ready. Instead, Sony first sent GCN based devkits, then replaced them with 8TF Navi devkits, then replaced them with 9.2 TF devkits, then iterated that chip four more times, only because they have no intention of using it because they are actually working on a 12 TF chip. And why, just why would you test BC modes on a pure devkit APU? Because devs want to play BC games?
I Love Rock 'n' Roll
Member
Is this the same brand that put the 360 on the market? RRD, bricked consoles, video failure, power supply problems, and sooooo on!
TLZ
Banned
Fixed that for you.
DeepEnigma
Gold Member
We know what happened last time he doubled downed.
Mass Shift
Member
You could go with all CUs enabled, the downside is that even if those fewer yields being fab'd are stable, you're still going to have fewer consoles to sell at launch.
Panajev2001a
GAF's Pleasant Genius
For devs to test BC support, it may be enhanced, of their titles?
Trimesh
Banned
Yes. The most obvious reason being that the Sophia Systems dev boxes were being shipped out to developers in '93 and they have hardware that's pretty much identical to the final retail Saturn - including both VDPs and the dual SH2 arrangement.
Heinrich
Banned
Cool
PS4 had a 50% GPU advantage but we can lower it to 45%
If PS5 is 45% less powerful than the Series X then you win
Xbox one had a better CPU. The Xbox One's processor has a faster clock speed of 1.7-GHz, compared to the PS4's 1.6-GHz CPU.
darkinstinct
...lacks reading comprehension.
The whole idea of Sony's BC solution is that you don't have to test it, because it just works in hardware by pretending to be a PS4/PS4 Pro. If you want enhancements you would not need BC modes in a devkit.
Panajev2001a
GAF's Pleasant Genius
There seems a lot of speculation here on your part. Sony has done enhanced BC before in several flavours: look at PS1 emulation on PS2, boost mode on PS4 Pro, system level supersampling support, etc...
Then again, let’s run with your arguments (the bolder one in this post and the previous “why, just why would you test BC modes on a pure devkit APU?” one): why test it at all then?
Last edited:
Hobbygaming
has been asked to post in 'Grounded' mode.
It wasn't much compared to the advantages the PS4 had with its GPU and unified ram
The Alien
Banned
DeepEnigma
Gold Member
Yeah, true.
But they're paying Microsoft for access. Lol.
Sony has been using AWS and their own cloud infrastructure for over a decade. Even NASA's OpenStack.
Hurr, durr, they pay MSFT. And MS pays Sony for a lot of software and tech as well. These businesses are not the fanatics forum dwellers are.
Last edited:
Vawn
Banned
Yeah, true.
But they're paying Microsoft for access. Lol.
Microsoft owns the cloud? Do they own the entire Internet?
Is this similar to how Microsoft "owns" the entire PC platform and gets revenue from Steam for every game sold?
Or how Xbox invented backwards compatibility? Or how Xbox is introducing game streaming even though PS has been doing it successfully for years already? Or how Microsoft's HD DVD format destroyed BluRay and now Sony has to use Microsoft's HD DVD for PS4?
Last edited:
The Alien
Banned
Microsoft has the infrastructure. Sony doesnt. If they didnt, this wouldn't have happened:
/cdn.vox-cdn.com/uploads/chorus_asset/file/16288012/xboxplaystation.jpg)
Microsoft and Sony are teaming up for the future of gaming
It’s all about Xbox and PlayStation beating Amazon and Google
www.theverge.com
Thaedolus
Gold Member
I was actually blown away by how many direct competitors supply components to each other in manufacturing. Samsung makes a killing on the iPhone parts they make, for example. Various companies I’ve worked for do exactly the same thing. Anyone thinking this is a reason to dunk on a company probably has no idea what they’re talking about
Kazekage1981
Member
Well that answers my question if the sony ps5 were to be 13 TFLOPS, would MS have enough time to up the specs of the hardware. I guess not.