Next Generation Hardware Speculation with a Technical Spin [2018]

Status
Not open for further replies.
Now the X significantly lowered the price in 3 months.
So:
- all the projections are based on wrong data
- it's just microsoft trying to sell more console at a loss knowing that the reference target will spend lot of money in contents
- blame Obama
?
Retail price doesn't necessarily represent BOM. ;) If the BOM was $500, it won't have been cost reduced any significant amount in three months. Ergo lower prices now represents retailers selling at a lower price. Stock can be sold at a loss if the retailer thinks it in their best interests, and XBox isn't selling well in Europe - chances are the more expensive XB1X is selling abysmally in some countries. Looking at the price at retail doesn't help determine BOM at all without an idea on what the profit margins are.

If XB1X has a confirmed BOM of $500 at launch, that helps with predictions.

I don't recall many predictions in this thread being based on XB1X's price and BOM. It's mostly been observations as to emerging tech and guesses as to availability and cost, and evaluations of die sizes and transistor counts and stuff. So I really don't think 'everyone' has been using XB1X as the starting point. And if they have, it was a good starting point if, as you say, the BOM was actually confirmed, because we get a clear cost of 6 billion transistors, 360 mm^2, 12 GB GDDR5 etc.
 
No-one's providing a link to show this info. Best I can find is...
"No," Xbox leader Phil Spencer told me in an interview this week, after I asked him if Microsoft makes any money selling the Xbox One X at $500.
"So, you're taking a loss?" I said. "I didn't answer it that way," he responded, intentionally not offering more detail.
http://uk.businessinsider.com/xbox-one-x-price-explanation-phil-spencer-e3-2017-6?r=US&IR=T

That says XBox One X is somewhere between lossy and break-even. If there's a better reference for the profitability of the hardware, can someone please link to it.
 
Next Generation Hardware Speculation: Dreams, Wishes, Cluelessness and a Dash of Reality [2018]
I wish that next generation of Switch will record users playtimes without 1 year time limit. They should boost the CPU/storage to achieve this incredible feat.
 
AMD's GCN successor is due to arrive in 2020, as is Zen 3. 7nm EUV is predicted to be up and running by then. HBM3 might even be affordable.

All that being the case, that seems like a good year for a new line of hardware.

-- PS5 - $350 --

- 8 core Zen 3

- 16GB HBM3 - 1 stack @512GB/s

- 8-10 TFlop GPU

- 48GB NVME drive

- Smallest, cheapest 2.5" HDD

- UHD Blu-ray drive

- A split DualShock/2 x Move 2 motion controller

-- PS5 Pro - $600 --

- 8 core Zen 3, clocked higher

- 32GB HBM3 - 2 stacks @1024GB/s

- 16-20 TFlop GPU

- 48GB NVME drive

- Smallest, cheapest 2.5" HDD, but larger than the base model

- UHD Blu-ray drive

- A split DualShock/2 x Move 2 motion controller

- Two bundled 4K60 remastered first party PS4 collections e.g. God of War and Killzone

---2021---

-- PSP 3 - $300 --

- 10" tablet, that can play every PS4 game

- Somewhat modular, with an upgradeable screen. Ship it with a decent 1080p LCD.

- 4 core Zen 3, clocked low

- 8GB HBM3 - 1 stack clocked low for ~200GB/s

- 2 TFlop GPU

- 2 x MicroSD slots. 1 for media, 1 for games

- Proprietary slot for game carts

- SIM card slot

- A split DualShock/2 x Move 2 motion controller
 
Someone needs to explain to me why so many people wish for Sony to go back into the handheld market. PSP did quite well but Vita has been disappointing - IMO neither of the big 2 have the IP/history/inventiveness to compete in that market.

I really doubt we will see a PRO console at launch (or at all if I am honest) the performance difference between a PRO console and a STANDARD console would be far too small and no one would notice the uplift.

Personally I think your expectations for the STANDARD console are where we will end up - an 8 core zen (L2/L3 probably halved) and a 4096 core GCN SoC + RAM + SSD Cache
 
Someone needs to explain to me why so many people wish for Sony to go back into the handheld market. PSP did quite well but Vita has been disappointing - IMO neither of the big 2 have the IP/history/inventiveness to compete in that market.

I agree on the matter of IP/history/inventiveness to an extent. I think it precludes them from competing with Nintendo in the realm of creative methods of interaction, but not from the market altogether.

Going by my own experience with the Vita, and the general mood that I've detected around the Internet from Vita owners, I reckon people are so eager for another Sony portable because the PSP was solid, and the Vita so very nearly got it just right.

That bloody rear touch pad was only ever a pain in the arse, and belies a Sony trying to compete with Nintendo, rather than focusing on what would make a portable PlayStation great: your PlayStation, but portable!

This was the premise with the PSP, and the Vita should have fully realised it, but it focused on gimmicks, like the rear touch pad and augmented reality, at the cost of features that matter.

The Vita needed to map 1:1 to the DualShock 4, it needed TV-out, it's games needed to be natively compatible with the PS4 (and maybe the PS3,) and it needed to use MicroSD cards. But they shat the bed and we ended up with the Vita and PSTV...

I really doubt we will see a PRO console at launch (or at all if I am honest) the performance difference between a PRO console and a STANDARD console would be far too small and no one would notice the uplift.

Personally I think your expectations for the STANDARD console are where we will end up - an 8 core zen (L2/L3 probably halved) and a 4096 core GCN SoC + RAM + SSD Cache

You're probably right (except for no one noticing the uplift.) I just can't escape the idea that a single tier launch means a balancing act that needn't exist in a two tier launch.

They have to target the mass market, and therefore a launch price of no more than $400, but this precludes the market that's willing to spend more money for more pretty, and I think it would be wise to target these people at the start of the generation, rather than a few years in.

The next generation should aim to last at least as long the PS360 generation, because there's no guarantee that we're getting 3nm fabrication any time soon, and I don't know what, if anything, is currently proposed after that.
 
PS5 and Xbox Two will be from 2020 to 2027/28 -

I'd like to think by the late 2020s, systems including consumer ones like consoles could be built with not only 3nm with EUV
but also make use of full 3D chip stacking, not merely 2.5D with stacked DRAM.

cfRR9vI.png


coFGsVL.png


37NRoK7.png


That's the main path I see for having a PS6 and Xbox 'Three' by 2028 that are a meaningful leap beyond a PS5 / Xbox Two.
 
I think a two tier console launch makes perfect sense to cater for all the segments in the pie. It works absolutely flawlessly for Apple and I don't see any harm here either. Base PS5 at 10 tf does mostly native 4k but 30 fps game with next gen visuals selling at $450 while PS5 Pro at 20 tf does native 4k either 60fps or higher settings selling for $650. Sony could then have all the bragging rights while simultaneously making everyone happy. Only question is just how far a theoretical PS5 Pro could be pushed if both launching at 2020?
 


I'd think the "Full 3D" will initially be more interesting for low-power solutions where each sub-system doesn't give out a lot of heat, i.e. mobile solutions.
Though I'd think a more realistic approach for a (portable?) game console SoC would be the larger GPU+SoC on the bottom, and then stacks for CPU cores, DRAM and NVRAM.

That's my 7nm PS4 Portable right there :)

And we shall call it Project Trinity.
 
consoles could be built with not only 3nm with EUV
but also make use of full 3D chip stacking
I'm sceptical. How would you cool a SoC drawing upwards of maybe 100W through several layers of DRAM? (Oh, and not also cook the DRAM. :p) And what if your SoC is larger than the DRAM stack, how do you reasonably cool the edges...?
 
I'm sceptical. How would you cool a SoC drawing upwards of maybe 100W through several layers of DRAM? (Oh, and not also cook the DRAM. :p) And what if your SoC is larger than the DRAM stack, how do you reasonably cool the edges...?
Are there new cooling technologies being developed for these 3D stacks?
 
They magically forgot to put gddr in that chart. Because gddr6 is going to be 864GB/s and 5 pj/bit this year.
 
Status
Not open for further replies.
Back
Top