News & Rumors: Xbox One (codename Durango)

[3dilettante]

If U read on to other future posts, U will get the patent link and good discussion.. its the next logical evolution of the APU (APD)... .. it goes on to explain lots of terms found in the VGLeaks posts, and points out some of the intentional changes in them too..

Believe what U will, we'll find out more in 6 days time up till //build/ conference

I'm supposed to go through another forum thread, read your mind, and make your point for you?

Where in the discussion does it cover speculated implementation details, or are we assuming Durango runs on a PCB composed of daydreams?
Can you expound on your case for the need for a discovery process and properties table designed for an open PC platform with discrete PCIe devices, when a console knows all this from the outset?

 

[COPS N RAPPERS]

there's a lot of speculation Microsoft has to address on the 21st. for those thinking about specs and the possibility of them changing, one would have to look at sony's latest strategy. killzone's demo was running at half of the memory and was still shown even when sony promised more space to come in the final model.

the possibility of them not changing would be based on if they already have a bunch of consoles made with a selection of games; that's ready to come out as soon a possible.

There's a solution. Get an "off the shelf" 8XXX, with a few more Tflops, connect to the APU, do a show on 21st with some demos, and pray that it can work and mass produced under 300W until the end of the year.

having to mess around with the circuitry at the last moment would be pretty difficult. however, the idea of using one of AMD's presets would be a faster solution and highly possible. according to the speculation a few games are to be showed off but as tech demos so nothing rules out the idea of their demos using refined hardware later on, if they haven't adjusted by now.

 

[bbot]

I wanted to address this, but the thread got closed. This passage in question is this:

"That's what we have with the likes of DF. Richard Leadbetter has been in this biz quite a while, and he has friends and people in the industry. That has to be taken on faith as, short of personally seeing him chilling all pally with devs in a bar somewhere, it's impossible to prove. But if you can accept that premise, than the source is as good as an engineer friend of your own telling you the facts. "

So Rixhard Leadbetter is reliable? really? If I recall correctly, he wrote an article in Digital Foundry, about DaE last year, that claimed he was in Europe. Yet, we find out this year, that he actually lives in Australia. I have to ask, can Leadbetter be trusted if couldn't even get the location of DaE right?

 

[Gradthrawn]

Good Googly Moogly.

This thread is in 3/10 territory and falling fast.

I think you need to update your estimation.

EDIT

In an attempt to make a more constructive reply, let me at least say this.

So Rixhard Leadbetter is reliable? really? If I recall correct, he wrote an article in Digital Foundry, about DaE last year, that claimed he was in Europe. Yet, we find out this year, that he actually lives in Australia. I have to ask, can Leadbetter be trusted if couldn't even get the location of DaE right?

bbot, if you're going to question or discuss anyone's credibility for something like this, you need to address their technical credibility directly. That, or you at least need to challenge the assertion that he has strong industry connections. Your post goes in the complete opposite direction, and doesn't raise any substantive or relevant issues. You don't even provide a source link to back up you're unrelated claim. That doesn't bolster good, constructive, discussion. I'm not a mod, but these are just my suggestions.

 

[MrFox]

Are we back to the 550mm2 SoC they developed in total secrecy?

 

[James Car]

Another factor to consider:

28nm has been in production for quite a while. 22nm is right around the corner. So going with what may be perceived as a "large" silicon budget would relatively quickly be reduced in size and consequently cost.

This combined with a higher expected MSRP based on the relative price strength of xbox360 7 years after launch and the revenue/profits that the ecosystem currently generates leads me to believe there is a possibility for a surprise spec announcement.

Not that I expect it (it seems every dev is under the impression wysiwyg), but it certainly is possible.


Honestly, prior to the vgleaks doc, I fully expected a much more robust spec for xb720. Factors to consider:

1) Previous silicon budget(s)
2) competitive landscape
3) current profit/revenue
4) projected profit revenue / expansion opportunities

All of the above lead me to believe a Pitcarn class GPU (200-300mm2) + 100-200mm2 cpu (28nm) was not only do-able, but expected.

Then came the vgleaks doc and the subsequent wink/nods from the dev community.

Time will tell, but given the pessimism of the community for anything above and beyond the rumored spec, I'm not holding my breath.

 

[3dilettante]

20nm, or a 20/14nm hybrid, is possibly about year (or a little less) for initial 20nm graphics products, with another year for the hybrid at a minimum.
The likely early adopters are at a different price level, so there's a desire to wait until the new nodes are mature enough to make the hop to the new node a cost savings over the current mature process. That may stretch the definition of "right around the corner".

 

[eastmen]

Dont see what all the short fuses are about. We have what 6 days left ? Mabye we should just leave this thread open for images of the tent and leaks and stop talking specs

 

[MrFox]

Dont see what all the short fuses are about. We have what 6 days left ? Mabye we should just leave this thread open for images of the tent and leaks and stop talking specs

Do you think we'll get full specs on the 21st?

 

[AlphaWolf]

Do you think we'll get full specs on the 21st?

I think you will get as much specs as you are ever going to get. E3 will be way more about spectacle (games) and features imo.

 

[bbot]

Good Googly Moogly.



I think you need to update your estimation.

EDIT

In an attempt to make a more constructive reply, let me at least say this.



bbot, if you're going to question or discuss anyone's credibility for something like this, you need to address their technical credibility directly. That, or you at least need to challenge the assertion that he has strong industry connections. Your post goes in the complete opposite direction, and doesn't raise any substantive or relevant issues. You don't even provide a source link to back up you're unrelated claim. That doesn't bolster good, constructive, discussion. I'm not a mod, but these are just my suggestions.

Okay, correction, Richard Leadbetter wrote this in Eurogamer.net. Here's the link to that article:

http://www.eurogamer.net/articles/digitalfoundry-the-curious-case-of-the-durango-devkit-leak

 

[Xenio]

Do you think we'll get full specs on the 21st?

after the show Larry Hryb will make an online meeting session to answer with details to the gamer's question about the reveal
take you ticket

 

[3dilettante]

It would be nice if there were architectural papers published like they were for IBM's Cell, or a GPU article akin to the one Dave Baumann wrote on Xenos for Beyond3D.

However, the former was for an IBM chip with HPC aspirations, and the latter's situation has changed.
At least in theory, AMD has been more open about its CPUs, historically. There was a significant backslide with Bobcat, and the clients for semicustom APUs may have some control over technical disclosure.

 

[James Car]

20nm, or a 20/14nm hybrid, is possibly about year (or a little less) for initial 20nm graphics products, with another year for the hybrid at a minimum.
The likely early adopters are at a different price level, so there's a desire to wait until the new nodes are mature enough to make the hop to the new node a cost savings over the current mature process. That may stretch the definition of "right around the corner".

Likely worst case scenario:
1 year production @ 28nm = ~10m chips.

After that, the cost reduction should be fairly significant and constant.

Even if they were stuck at 28nm, they still would not be losing money selling 250mm2 gpu + 150mm2 cpu for $500.

They might be envious of Apple hardware profits, but they do not have the brand power nor comparatively weak competition to command anywhere near their margins on hardware...

 

[Xenio]

the ui is probably an evolution of:

http://www.youtube.com/watch?v=MqUaxnhp7rc


this is from the xbox facebook page

 

[COPS N RAPPERS]

I think you will get as much specs as you are ever going to get. E3 will be way more about spectacle (games) and features imo.

the thing is, If the specs are presumably weak, why would they want to show them? why go publicly to expose an Achilles tendon? I don't think they're that dumb enough to shoot them selves in the foot.

 

[Shifty Geezer]

So Rixhard Leadbetter is reliable? really?

Richard's contacts are technical people who work with hardware and software, and not with locating people or Geography experts. plus he says DaE was 'seemingly running rampant across Europe, quaffing significant amounts of absinth,' according to internet messaging conversations had with him. No claim to his location whatsoever.

Hopefully you appreciate the validity of Leadbetter's position here, but if not, that's your prerogative. However, this thread needs to appreciate the value of that information and at this point I'm not willing to tolerate noise that's trying to dismiss keystone information. It's time to move on.

As for what changes could happen in Durango, James Car's idea of completely replacing the hardware seems plausible, although very risky. MS would have to rewrite their API to handle all the special-cases that devs are using (memory movers would be emulated on CPU, SHAPE also, eSRAM,...). Either that or they say to their devs, "you know what, scrap that Durango software you've been writing and just focus on the PC version. We're replacing our console with a PC spec." Then there'd be managing the software, because with pure PC hardware it could be hacked to run PC games, meaning no revenue stream for MS from Steam sales etc.

It's certainly possible, but it would be a sign of desperation. And the basis of all this questioning seems to be Orbis and this theory that MS have seen Orbis's specs and pooped their pants and are now running around panicking. I don't believe that management's strategy wasn't more solidly considered than that and that the whole Durango system and experience wasn't considered value except in competition to a hoped-for weaker PS4. Just going by the 10x performance upgrade per generation, MS must have expected something considerably more than 1.2 TF from Sony, yet despite that, chose their lower spec'd hardware.

I don't think MS are overly concerned about the performance differential, and I don't think they'll be motivated to do something about it in hardware.

 

[3dilettante]

Likely worst case scenario:
1 year production @ 28nm = ~10m chips.

10 million consoles sold in the first year?

Even if they were stuck at 28nm, they still would not be losing money selling 250mm2 gpu + 150mm2 cpu for $500.

They aren't selling a standalone chip for $500. Even in a non-loss scenario, why would they not want to make more money?

 

[Xenio]

As for what changes could happen in Durango, James Car's idea of completely replacing the hardware seems plausible, although very risky. MS would have to rewrite their API to handle all the special-cases that devs are using (memory movers would be emulated on CPU, SHAPE also, eSRAM,...). .

is the emulation still incomplete?
From what remember on the dev kit an additional 8 cores cpu is still not enough powerful to emulate SHAPE;
And why they should rewrite the API?
I thought that they should rewrite the lower software layer, the drivers layer. If the API are high level as are chatted they should be the same even if the hardware, and drivers layer, changes.
Or not?

 

[James Car]

Richard's contacts are technical people who work with hardware and software, and not with locating people or Geography experts. plus he says DaE was 'seemingly running rampant across Europe, quaffing significant amounts of absinth,' according to internet messaging conversations had with him. No claim to his location whatsoever.

Hopefully you appreciate the validity of Leadbetter's position here, but if not, that's your prerogative. However, this thread needs to appreciate the value of that information and at this point I'm not willing to tolerate noise that's trying to dismiss keystone information. It's time to move on.

As for what changes could happen in Durango, James Car's idea of completely replacing the hardware seems plausible, although very risky. MS would have to rewrite their API to handle all the special-cases that devs are using (memory movers would be emulated on CPU, SHAPE also, eSRAM,...). Either that or they say to their devs, "you know what, scrap that Durango software you've been writing and just focus on the PC version. We're replacing our console with a PC spec." Then there'd be managing the software, because with pure PC hardware it could be hacked to run PC games, meaning no revenue stream for MS from Steam sales etc.

It's certainly possible, but it would be a sign of desperation. And the basis of all this questioning seems to be Orbis and this theory that MS have seen Orbis's specs and pooped their pants and are now running around panicking. I don't believe that management's strategy wasn't more solidly considered than that and that the whole Durango system and experience wasn't considered value except in competition to a hoped-for weaker PS4. Just going by the 10x performance upgrade per generation, MS must have expected something considerably more than 1.2 TF from Sony, yet despite that, chose their lower spec'd hardware.

I don't think MS are overly concerned about the performance differential, and I don't think they'll be motivated to do something about it in hardware.

In general I agree, however, the competitive landscape outside of the direct Sony comparison may be what steers MS to a higher spec. If MS were generally leaning on the notion of casual + core gamer, but have recently seen the casual market has shifted strongly into the free/very cheap gaming scene on ios/android, then that leaves only the core gamer to bank on. WiiU has demonstrated rather clearly that regardless of perceived brand strength and prior success with the casual crowd, they (the casual crowd) cannot be counted on for revenue in this current marketplace.

If their research of core gamer response to the durango spec has lead them to believe that they risk losing a percentage of their base which is currently netting them $1b+ annually, then they may decide that a higher spec (and assumed associated costs) is less risk than the rumored durango spec. (and assumed associated losses).

Not that I expect it, but it is certainly plausible.