Xbox : What should MS do next? *spawn

[function]

MS wanted the lowest price; which is why they went with DDR3+ESRAM; not from a performance standpoint, merely a price standpoint.

MS wanted to manage the cost over the long term, while guaranteeing high (and apparently higher than-PS4) levels of bandwidth.

This was more work than slapping on GDDR5. Add to this their own coherent bus, their own Kinect silicon, their own better-than-anywhere-else audio solution, their support for minimising virtualisation overheads ...

... they took on a a far bigger technological challenge than Sony. And yet you say ...

The current Xbox One design IMO reflects the skills of its architects; it's just a box box with a big fan.

Only a troll would say this, and certainly only an idiot would think it. It's a "more exotic design" by far than Sony offered, yet somehow this demonstrates a lack of skill on behalf of the architects?

What?

 

[Shifty Geezer]

I don't know that much about embedded memory. Just curious what else is out there that's faster.

Nothing much, but if creating a bespoke part, they could have gone with something with a bit more pizazz. XB1's ESRAM is a pretty prosaic cost-reduction feature, rather than a performance piece that really gives the platform some teeth in a specific area. If we look at Cerny's remarks about evaluated options for PS4, 1TB/s was on the cards. So really, anything up to that was possible (at who knows what cost). If XB1 had launched with 32 MB of 400 GB/s ESRAM, it'd have had a landmark specification and something to crow about.

 

[function]

But I still wonder what is the real technical explanation why the ESRAM is so slow. Many GPUs have external memory faster than this.

209 GB/s is super fast for a large CPU cache. Masses faster than AMD or even Intel's L3 caches and it's much bigger. I'd need to check, but it's probably a lot faster than Intel's on package edram too.

I can't think of when anyone has put such a large bank of esram on chip, and that's it's running at L2-ish BW is pretty amazing.

One of the chip folks (3dillattante?) talked about why making huge tracks of esram high BW is tough, but I can't remember the reason, perhaps something to do with how you manage large numbers of simultaneous accesses.

Another factor is probably the way that esram seems to be tied to some kind of ROP-related crossbar.

Effective use of esram, including packing buffers into 64 bpp elements and using it for mixed reads and writes during compute, is probably going to be the key to getting good performance from the Bone.

 

[Rangers]

MS wanted to manage the cost over the long term, while guaranteeing high (and apparently higher than-PS4) levels of bandwidth.

This was more work than slapping on GDDR5. Add to this their own coherent bus, their own Kinect silicon, their own better-than-anywhere-else audio solution, their support for minimising virtualisation overheads ...

... they took on a a far bigger technological challenge than Sony. And yet you say ...

I dont know if this is the thread for it, but it's beginning to look to me like MS botched the Xone architecture.

They gained DDR3 at the expense of significantly less power, apparently.

I guess we dont, nor will ever, have access to the P&L sheets, but by Sony's fiscal's it's certainly not looking like the PS4 hardware is a big money loser, in fact it looks to perhaps already be break even or better, all for $399 out of the gate!

Unless MS can show me, by undercutting PS4 on price by at least 50 dollars, if not $100, over a long term, then I dont know if the DDR3 savings really exist, or are significant enough.

Show me the DDR money, basically.

All this time I've framed it mentally as "MS had to be less powerful because they wanted to include Kinect!". But it's beginning to appear like the PS4 style design would have had no real drawbacks, regardless of Kinect or not! It doesn't appear PS4 is any more expensive. The base PS4 is $399, the same as the base Xbox One.

Too be fair to MS, I'd argue 360 was better engineered (in power terms) than PS3 (coming out a year prior but being basically equal, regardless of the reasons, is significant), and Xbox certainly lambasted PS2 in power, if arguably a bit basic in design. So, I mean you aren't going to win them all I guess. And I'd also wonder whether MS engineering got worse this gen, or Sony's just got that much better with Cerny.

But the bolded part is key, it's not set in stone just yet. Just that a picture is taking shape.

The problem with the bolded sentence is, even if such does come to pass, it then becomes arguable strategically whether $50 or even $100 is a worthwhile trade in the market for (possibly) substantially less power. That answer is at the very least debatable imo. But it's currently a moot point cause MS hasn't even reached that point yet to begin with!

And if you want my reason WHY MS botched it IMO, it's because of their bizarre insistence on ES/EDRAM. Which they seem to be committed to regardless of benefit or lack theoreof.

There's other arguments here, an argument can be made that Xone will turn out in the end to be "good enough" hardware. But for now, yes.

To try to keep it a little on topic, MS did put out a call for an industrial Xbox hardware engineer recently, right? That's what they must do now, try to come out with a significantly cheaper slim revision ASAP. I think they may be giving up DDR3 savings on the big bulky, too-high quality design.

But that was another strategic stupidity by them, prizing silent operation above what actually counts. Basically the dumbness of know-nothing suits deciding everything matters except for the actual gaming part.

 

[function]

Nothing much, but if creating a bespoke part, they could have gone with something with a bit more pizazz. XB1's ESRAM is a pretty prosaic cost-reduction feature, rather than a performance piece that really gives the platform some teeth in a specific area. If we look at Cerny's remarks about evaluated options for PS4, 1TB/s was on the cards.

Was it really?

They evaluated it, but it doesn't mean that it was realistic. For a start, there's no competitive node to make an edram equipped SoC on. So that's a bust.

Secondly, if you look at an on package edram chip, you'd need a horrendous amount of IO. We're talking HBM tier IO, and no-one's done that yet. AMD haven't even done it yet, and neither have Nvidia. Not even on the $2000 parts.

And if you're talking about on-die bandwidth for an off chip, on package daughter die, then you get into 360 territory where the effectiveness of that BW has to be considered against the effectiveness of things like depth compressions and ROP caches.

Where that 1000 GB/s of BW was, and where it could be used are just as important as the headline figure.

What we know about MS's solution is that while the headline figure isn't huge, it's more - and a lot more per ROP/TMU/ALU than any external bus could have provided the, and the latency *should* be lower than for an off-chip solution.

 

[mosen]

I dont know if this is the thread for it, but it's beginning to look to me like MS botched the Xone architecture.

They gained DDR3 at the expense of significantly less power, apparently.

I guess we dont, nor will ever, have access to the P&L sheets, but by Sony's fiscal's it's certainly not looking like the PS4 hardware is a big money loser, in fact it looks to perhaps already be break even or better, all for $399 out of the gate!

Unless MS can show me, by undercutting PS4 on price by at least 50 dollars, if not $100, over a long term, then I dont know if the DDR3 savings really exist, or are significant enough.

Show me the DDR money, basically.

All this time I've framed it mentally as "MS had to be less powerful because they wanted to include Kinect!". But it's beginning to appear like the PS4 style design would have had no real drawbacks, regardless of Kinect or not! It doesn't appear PS4 is any more expensive. The base PS4 is $399, the same as the base Xbox One.

Too be fair to MS, I'd argue 360 was better engineered (in power terms) than PS3 (coming out a year prior but being basically equal, regardless of the reasons, is significant), and Xbox certainly lambasted PS2 in power, if arguably a bit basic in design. So, I mean you aren't going to win them all I guess. And I'd also wonder whether MS engineering got worse this gen, or Sony's just got that much better with Cerny.

But the bolded part is key, it's not set in stone just yet. Just that a picture is taking shape.

The problem with the bolded sentence is, even if such does come to pass, it then becomes arguable strategically whether $50 or even $100 is a worthwhile trade in the market for (possibly) substantially less power. That answer is at the very least debatable imo. But it's currently a moot point cause MS hasn't even reached that point yet to begin with!

And if you want my reason WHY MS botched it IMO, it's because of their bizarre insistence on ES/EDRAM. Which they seem to be committed to regardless of benefit or lack theoreof.

There's other arguments here, an argument can be made that Xone will turn out in the end to be "good enough" hardware. But for now, yes.

Add to all of your thoughts that XB1 was months behind PS4 in production. IIRC, while developers had stable PS4 dev-kits/SDK in their hands Microsoft was dealing with making hardware/software to just work. It's strange to me that they had problems to make an inferior hardware in time and many reports indicating that one next-gen console was going to launch in 2014. At the same time I know that same person (Nick Baker) who was architect of Xbox Original and Xbox 360, is the manager of hardware architecture team of XB1.

My team is really responsible for looking at all the available technologies. We're constantly looking to see where graphics are going - we work a lot with Andrew and the DirectX team in terms of understanding that. We have a good relationship with a lot of other companies in the hardware industry and really the organisation looks to us to formulate the hardware, what technology are going to be appropriate for any given point in time. When we start looking at what's the next console going to look like, we're always on top of the roadmap, understanding where that is and how appropriate to combine with game developers and software technology and get that all together. I manage the team. You may have seen John Sell who presented at Hot Chips, he's one of my organisation. Going back even further I presented at Hot Chips with Jeff Andrews in 2005 on the architecture of the Xbox 360. We've been doing this for a little while - as has Andrew. Andrew said it pretty well: we really wanted to build a high-performance, power-efficient box.

These are bold statements: "working with DirectX team to see where graphics are going", "We have a good relationship with a lot of other companies in the hardware industry", "the organisation looks to us to formulate the hardware", "when we start looking at what technology are going to be appropriate for any given point in time, we're always on top of the roadmap".

Although, there are still some unknown characteristics about XB1 that we know nothing about, but I can't see any sign that indicate XB1 is the next XB360 (in term of level of technology/design that XB360 had at it's time). They had right people and enough resources to make a really good & powerful console (and they think they did it with XB1) but right now no one sees XB1 as a cutting edge technology that it supposed to be and they priced it at 400$ just like PS4 which is the more powerful console.

If you are interested, there is a video about Xbox history which should be kinda related to this thread:

XBox Oral History Panel with Nick Baker, Todd Holmdahl, and Albert Penello
https://www.youtube.com/watch?v=_gOoI57q72M

To try to keep it a little on topic, MS did put out a call for an industrial Xbox hardware engineer recently, right? That's what they must do now, try to come out with a significantly cheaper slim revision ASAP. I think they may be giving up DDR3 savings on the big bulky, too-high quality design.

But that was another strategic stupidity by them, prizing silent operation above what actually counts. Basically the dumbness of know-nothing suits deciding everything matters except for the actual gaming part.

If you search Microsoft career you will see a lot of jobs that are related to a team that created Kinect and now is preparing to make "the next big thing" and make "the technology disappear". They are working on a new hardware/softwares/games around computer vision and they have a lot of job openings.

 

[eastmen]

In this instance it could force sony off a golden goose. IF the ps4 moves that far ahead that ms wants to release a new xbox in 4 years sony will find itself against a system that is most likely on 16nm or less with a lot more power under its hood. It would force sony to move off the ps4 quickly and put out another system. Waiting more than a year after a new xbox would surely bea very hard thing to recover from. Heck they might still even be on the same micron process so the power difference might not be very large

 

[wco81]

In this instance it could force sony off a golden goose. IF the ps4 moves that far ahead that ms wants to release a new xbox in 4 years sony will find itself against a system that is most likely on 16nm or less with a lot more power under its hood. It would force sony to move off the ps4 quickly and put out another system. Waiting more than a year after a new xbox would surely bea very hard thing to recover from. Heck they might still even be on the same micron process so the power difference might not be very large

Hope that happens.

It won't be the worst thing for Sony because they would probably be pushing 4K TVs by then, with 4K Blu Ray.

 

[eastmen]

Hope that happens.

It won't be the worst thing for Sony because they would probably be pushing 4K TVs by then, with 4K Blu Ray.

I think VR will be a big factor for both of tehse companies. Neither system will be great for a VR experience but a 2017 system could be a perfect fit.

 

[wco81]

If VR adds over $100 to the price, it'll fail just as spectacularly as Kinect 2.

At least with 4K there's a chance that it'll appeal to people outside of core gamers.

Think about 4K Netflix, 4K Blu Rays, other 4K media sources (including your own photos, home 4K video shot on Go Pros mounted on drones, etc.).

 

[eastmen]

If VR adds over $100 to the price, it'll fail just as spectacularly as Kinect 2.

At least with 4K there's a chance that it'll appeal to people outside of core gamers.

Think about 4K Netflix, 4K Blu Rays, other 4K media sources (including your own photos, home 4K video shot on Go Pros mounted on drones, etc.).

I know all about 4k but I also know how long it took 1080p to become standard. I bought my first 1080p tv back in 2000 and its now 2014 and still some content isn't in 1080p

Also I believe if they stick with apu's like they did this generation we wont see capable 4k machines till the next decade and by 4k capable machines I mean we have to have the same graphics leap or similar that we have from the xbox 360 to the ps4 along with the resolution bump.

I for one do not want to wait another 6 years for new consoles esp not after what we got this gen. Meanwhile I believe we can get good vr at 1440p

 

[Deleted member 11852]

It won't be the worst thing for Sony because they would probably be pushing 4K TVs by then, with 4K Blu Ray.

O/T Sony are already pushing 4K TVs. It's only the budget 'R' series and mid-range 'W' series that are 1080. The entire 'X' series is 4K. Samsung, Toshiba, Panasonic, LG and everybody else's premium TVs are 4K as well and you don't need to be buying a massive 60" or 80" screen to get it.

Of course regardless of what Microsoft are doing, Nintendo and Sony are pushing forward as well. Well Sony will be pushing forward, Nintendo are probably spinning in circles while shuffling sideways a bit :|

Remember Microsoft do not exist in a vacuum. You could see a complete repeat of this generation in the next Xbox not being the console with the highest performance. Unlikely tho it is, Nintendo could surprise everybody

 

[wco81]

1080p did not take that long. When PS3 was announced, they touted Blu Ray and supposedly 1080p games.

Xbox fanboys said nobody had 1080p TVs, though stores were advertising them already in 2006. Best Buy circulars would point out which TVs were 1080p.

By the end of the decade, 1080p TVs were common.

Yes there is a lot of content which isn't 1080p. That has more to do with the ATSC standard, which was developed in the '90s or maybe even earlier. Now spectrum has become more valuable so it's not likely they'll broadcast in higher resolution over the air than what they're doing now, because TV stations are being asked to give up spectrum so mobile networks can have more of it.

But other countries are looking to broadcast in 4K. BBC tested 4K broadcasts from the World Cup. NHK in Japan are testing 8K.

Meanwhile, the Chinese are buying up 4K TVs like crazy, mostly from Chinese brands, some of which are already under $1000.

I'm not saying 4K will be widespread though. It may be no more successful than 3D. I have a 3D TV but like most people, I've never watched anything on it in 3D.

But 4K has a much better chance of catching on than VR.

 

[Deleted member 11852]

1080p did not take that long. When PS3 was announced, they touted Blu Ray and supposedly 1080p games.

I guess it depends when you bought. I bought a HDTV in February 2007, a month before the PS3 launch in the UK. At the time 1080p was very expensive so I got a 40" 720p Samsung. I didn't go 1080p until June 2010 (my Samsung died, having been pretty unreliable). Tomorrow I go 4K, but that's Graham's fault

Xbox fanboys said nobody had 1080p TVs, though stores were advertising them already in 2006. Best Buy circulars would point out which TVs were 1080p.

Certainly in the UK 1080p were available but expensive, as in £1,200+ in 2007. Kind of like the prices 4K sets are now. I'd getting a modest 49" set and that's £1,400 and it just price dropped from £1,600 three days ago. 55" sets come in at about £2,000 and go up from there.

By the end of the decade, 1080p TVs were common.

Very true but Blu-ray was there with content right at the start. 4K content is rarer than yeti shit. If you want to buy Captain America Winter Soldier in 4K then no amount of money will get you that, although the benefit of a 4K screen for passive 3D vieiwng is good - if you care about 3D.

I'm not saying 4K will be widespread though. It may be no more successful than 3D. I have a 3D TV but like most people, I've never watched anything on it in 3D.

Yeah I think 4K is a tougher sell. 1080p is clearly, except to those with poor eyesight, a damn sight better than standard definition. 4K over 1080p is much more subtle at average viewing distances but I think it's the lack of content that will mean its adoption will be slower than 1080p

But 4K has a much better chance of catching on than VR.

I don't see the connection. Only one person at a time can experience VR on a single device, where as many people as you can comfortably for in a room can experience a TV. These are really not aimed at the same market

 

[wco81]

I'm talking about in the context of what the key selling features of next gen consoles would be.

I don't think better 1080p graphics than X1/PS4 is necessarily a big selling point. Sure core gamers will like the incremental improvement and understand that resolution isn't the key metric for image quality.

4K content is there. With the advent of Blu-Ray (and HD-DVD while it existed) studios built up not only an infrastructure for encoding 1080p, they built up a library of 4K masters. New movies would have 4K masters and they went back to the archives and scanned a lot of them in 4K, often manually scanning individual frames. They've been doing this for almost a decade now.

Whether the studios are ready and willing to release 4K content is another question.

If they see home video sales decline, they may be prepared to, as Blu-Ray releases don't hold prices very long. They come out released between $20-30 and in a year, they're around $10.

Supposedly the 4K Blu Ray standard will be announced at the end of this year.

 

[Deleted member 11852]

I'm talking about in the context of what the key selling features of next gen consoles would be.

I know. I'm explaining why I think 4K will spread at a snails pace compared to 1080p and, even in four years, may not be a target worth aiming for.

To be clear, supporting 4K output yes. Expecting games to run at 4k/60fps with lots of whizzes and bangs, no.

 

[wco81]

Non-gaming 4K content may play a role. If 4K Netflix becomes popular.

Or if good 4K displays become much more affordable (there are a lot of no-name Chinese brands being sold in Chinal already) in the West, there could be interest in 4K Blu Ray. That could impact the all-digital plans that console makers may have.

Or maybe good 1080p to 4K scalers will be cost-effective by then, so they would upscale games output as well.

But if 4K in computers become more affordable and common by then, you would think GPU support for that resolution would become more common.

 

[dobwal]

In this instance it could force sony off a golden goose. IF the ps4 moves that far ahead that ms wants to release a new xbox in 4 years sony will find itself against a system that is most likely on 16nm or less with a lot more power under its hood. It would force sony to move off the ps4 quickly and put out another system. Waiting more than a year after a new xbox would surely bea very hard thing to recover from. Heck they might still even be on the same micron process so the power difference might not be very large

Or you end up with a more powerful console that receives little more than ports with minimal advantages over their PS4's counterparts as Sony stands pat and waits for a more opportune moment to release its next gen console.

Everyone complains about the negatives of cross gen porting but devs and pubs are surely going to be less than enthused about trying to push new hardware to its limits when the current gen isn't in decline but rather a zenith given that the PS4/XB1/WiiU are going to be in the midst of trying to saturate the mass market with low price points.

The 360 did great with an early start but it also benefited from Sony's missteps with a highly overpriced console that took some devs years to fully exploit. What are chances of that happening again? Never mind that regardless of the 360 early entrance, the PS2/XB1/GC had already saturated the markets with cheap consoles and the gen was already in decline. All the while the PS2 pretty much still outsold the 360 on a monthly basis during its first year.

It could also be said that while the 360 launched 4 years after the XBOG, the original XB didn't release duing the start of the gen and depending on where one's opinion lies, the XBOG launched anywhere from 18 months (ps2 japan launch)-36 months (dreamcast japan launch) after the gen had started. So the 360 showed up anywhere from 5.5 years to 7 years after the sixth generation of consoles had begun.

All MS would probably end up with is an early head start but finding itself right back where it started, a console that is considered underpowered to its contemporary competitor from Sony. MS has already looks like it pulled a "PS3" with the XB1 and its overly complicated hardware weighed down pricewise by an ancillary feature. The last thing MS wants is for the XB1 and the next XB to be compared to the Sega Saturn and Dreamcast.

 

[steveOrino]

I dont think hardware is the big issue.

Lets not ignore that multi-track PR train wreck that went on towards its launch (internal sabotage?). MS essentially decimated the Xbox brand in 6 months.

I really got a sense after their world unveiling that they had absolutely no clue whom they were trying to sell this thing to. Thats the only explanation I have as to how they fucked up that badly.

X1 was a poorly planned and thought out system which is probably the direct result of the number guys dictating its design.

Consumers didn't want:
1. Kinect
2. Redundant TV pass-through that set top cable/sat boxes do better.
3. Draconian DRM

They fixed 1&3 pretty quickly.

Now they have to rebuild the tarnished brand. That takes time and smart well thought out messaging + delivering on that messaging. A new console isn't going to help.

So tl;dr, MS should continue positive messaging towards gaming consumers and adding value for consumers to pick up their platform over the competition.

 

[Shifty Geezer]

In this instance it could force sony off a golden goose. IF the ps4 moves that far ahead that ms wants to release a new xbox in 4 years sony will find itself against a system that is most likely on 16nm or less with a lot more power under its hood. It would force sony to move off the ps4 quickly and put out another system. Waiting more than a year after a new xbox would surely bea very hard thing to recover from. Heck they might still even be on the same micron process so the power difference might not be very large

Not necessarily. It depends if gamers are ready to move on or not. There was an easy 15+ million eager to leave PS360 behind when the new console released. In four years time, how many early adopters will be ready for new hardware? If the games are still good on the machines they already own, waiting another year for a more powerful machine from a rival could make better sense. And if devs aren't ready to move on either, the new machine might not get the content that justifies its purchase.

You can move too early just as you can move too late. It's plausible that an early machine only sells 5 million in its first year and then a more potent rival can attract the masses with a bigger splash.