Next Generation Hardware Speculation with a Technical Spin [2018]

[Deleted member 11852]

Depending on competition. If there's a better product, or cheaper product, or better marketed product, your good, appealing product can go unbought.

Those are factors that contribute to appeal. I'm taking as read that the level of appeal will be swayed by price and competition. All things being equal, most people will buy the product that appeals the most.

 

[HBRU]

looks like its really -mainly- CPU upgraded... VR needs 60 fps !
I think all this hype will force Sony to say something...

 

[MrFox]

Did you take a picture of your screen to make it look like a leak?

It's a post on resetera.

 

[Shifty Geezer]

I don't mind the speculation or even far fetched conjectures.

The reason I say it's a scam is because the preface makes it look like they have insider info, a scoop about specs and release date. Then the article is no more than the usual blog or forum post. Speculation with zero additional info. Not even an additional insight that hasn't been already speculated elsewhere.

Even if the speculation eventually ends up correct, it changes nothing to the tabloid shit he wrote to get easy money.

If me, I'd go through consumer law to get my money back as it was missold.

 

[Shifty Geezer]

Those are factors that contribute to appeal. I'm taking as read that the level of appeal will be swayed by price and competition. All things being equal, most people will buy the product that appeals the most.

Well then 'good product' comes under appeal as well. Don't need anything other than appeal to sell. Getting appeal means making a good product and competing.

 

[HBRU]

yes... Its a post on that forum

 

[Deleted member 13524]

Have prices on 7nm come down so much that they can begin mass manufacturing at a low price point? Usually it takes a at least a year of maturing before the prices start to come down.

The Neo APU AFAIK is 16FF+ and it probably started mass production at around the same time nvidia started their Pascal GPU production, which was Q1 2016.
Snapdragon 855 and Exynos 89xx flagships for 2019 phones will most probably use 7nm DUV from Samsung and TSMC. That means mass production of these SoCs will probably start in late 2018 to early 2019.
The PS5's SoC could start at around the same time.

Yeah. MS/Sony are likely to pull the rug from under Nintendo. Nintendo's handheld market has always been tenuous at best, just waiting for anyone else to enter to steal it all away from them...

I know this is sarcasm, but my post was a counterpoint to the OMG FASTEST-SELLING CONSOLE EVA (in the USA) argument.
I didn't suggest the Switch is doomed the first day it gets competition. I do suggest that part of its success comes from a complete lack of competition from Sony/Microsoft and sheer incompetence from Google+Apple for being unable to create an adequate ecossystem for AA-AAA games in their platforms.

If me, I'd go through consumer law to get my money back as it was missold.

I don't think you would be successful.
Charlie is basically selling the consultancy service of going through a 35-page long thread in the very specific B3D enthusiast forum.
For you and me that's not worth $1000. For some others who have zero knowledge around this area and have e.g. venture capitals to manage, it might be worth it.

 

[Deleted member 11852]

Well then 'good product' comes under appeal as well. Don't need anything other than appeal to sell. Getting appeal means making a good product and competing.

Maybe. A product can be good but not appeal, just look at the Zune. I, reviewers and anybody who used one would say the Zune was a good product (clearly superior to the iPod in many areas) but it clearly didn't appeal to the masses. Ditto the Vita. Ditto Betamax video recorders. Are these any less good because they didn't outsell the competition? True, good/bad can be somewhat subjective but in any product industry, when critics and public overwhelmingly praise a product, I'm willing to accept such consensus as a measure of a 'good product' even if that doesn't equate to commercial success.

 

[Shifty Geezer]

What's the source? The already discussed SemiAccurate scam?

 

[MrFox]

What's the source? The already discussed SemiAccurate scam?

Yeah the mods on resetera removed the screengrabs of paid articles since it's not allowed, and the first post is now the above summary instead.

If this is going to spread around, as diarrhea does, it should be mentionned every time that the source is not a leak, it's as wild speculation as any forum posts.

 

[iroboto]

The Neo APU AFAIK is 16FF+ and it probably started mass production at around the same time nvidia started their Pascal GPU production, which was Q1 2016.
Snapdragon 855 and Exynos 89xx flagships for 2019 phones will most probably use 7nm DUV from Samsung and TSMC. That means mass production of these SoCs will probably start in late 2018 to early 2019.
The PS5's SoC could start at around the same time.

According to this:
http://www.tsmc.com/english/dedicatedFoundry/technology/16nm.htm
16FF+ was in volume production July 2015. So it still took a year for pricing to come down enough for it to be available for consumer grade products.

IIRC, 7nm Vega is not available for consumer products in 2018, mainly sampling and feedback, I suspect full volume product will be available on Navi in 2019.

 

[Metal_Spirit]

Yeah the mods on resetera removed the screengrabs of paid articles since it's not allowed, and the first post is now the above summary instead.

If this is going to spread around, as diarrhea does, it should be mentionned every time that the source is not a leak, it's as wild speculation as any forum posts.

You are wildly speculating

 

[Deleted member 13524]

16FF+ was in volume production July 2015. So it still took a year for pricing to come down enough for it to be available for consumer grade products.

I guess it depends on what you'd call "volume", because the only chip that was being produced at that time in 16FF+ was the apple A9, and even then apple had to dual-source with Samsung in a different process.

Regardless, volume production of 7nm is probably starting Q3 2018 (if not before), but I'm not suggesting the PS5 will release in 2018, as that would be really dumb considering the current sales numbers for the PS4.
Even more considering how the hardware gaming PC market has become rather stale due to a perfect storm of graphics cards, RAM and storage prices.

1.5 years after release and we can hardly buy a $400 graphics card that matches the PS4 Pro in performance.
Back in late 2015 we could buy a 5 TFLOPs R9 390 with 8GB GDDR5 for $330.
Right now, even after a ~1 month-long crypto-crash we're lucky to get a RX570 4GB for that price.

 

[HBRU]

8 core Zen + Navi + 16 gbytes RAM (I guess)... really expensive... 700 US$... That make sense so it kills a 500 US$ OneX but not a 200/300 $ ps4/ps4pro

 

[Deleted member 11852]

Regardless, volume production of 7nm is probably starting Q3 2018 (if not before), but I'm not suggesting the PS5 will release in 2018, as that would be really dumb considering the current sales numbers for the PS4.

TSMC began risk production in 7nm a year ago so chances are, they're already running volume production. What isn't known is what yields of complex non-RAM ICs are. You can improve yields with better processes, better wafers and chip redesigns. Back then TSMC themselves predicted double-digit yields in 2017 which means anywhere from 10% up to 99%. In TSMC terms, I don't think something you may see in PS5 would constitute 'volume' production, they're talking tens of millions of cores per year. This is based on conversations we had with them a few years back on what we felt was a large job but really wasn't for them - even employing more complex processes necessitating expensive materials in specific atmosphere environments.

 

[lefantome]

What's the track record of that webiste? never heard of it.

It's too early, I don't think Sony is capable of have realistic devkit now unless it want to release the PS5 early, a move that would not make any sense from a market point of view.

 

[Shifty Geezer]

First XB360 devkits (the Power Macs) released early 2004. Final console released late 2005. In May 2005, PowerMacs were still being used at E5. So a good 18 months between first devkits and final release in that example and only 6 months from E3's PowerMac devkits to final hardware.

In the case of PS5 being x86 and an AMD GPU, any equivalent PC will be suitable for first development. You could even start on the PS4 devkit no doubt, and improve your engine nearer the release date. Or release a first-year, unoptimised title as is usual. So not too early at all by my reckoning.

 

[Grall]

You're not interested in Switch so the fact it is the fastest selling console in US history is irrelevant. Sure.

Strawman argument. Nice!

Oh but the news are perfectly credible.

The "news" are merely reasonable assumptions, with what I understand to be no attributed sources. Which makes them...not news. It's just speculation.

 

[dten]

For recent console leaks, when one outlet breaks news, it wasn't long before several outlets report multi-sourced corroboration. If this remains single sourced, I'd put much less weight to it.

Whether this article has any merit, it's funny that many people just go nuts and virtually take it as fact without waiting for corroboration. Guess that's why fake news will always be with us.

 

[DieH@rd]

What's the track record of that webiste? never heard of it.

It's a very old website, and Charlie often posts accurate insider info [but he's far from 100%].