Nextbox infinity anticipation station

[MrHat]

http://www.youtube.com/watch?v=ifa9Q7ATfVA

Here, have an Xbone UI preview.

Gotta love that exclusive NFL contract and the blurred NFL footage.

[Margalis]

Might be actually helpful to link something when you mention that it's everywhere.

Gamasutra alone averages about one negative Greenlight piece a week, not counting unpromoted member blogs.

Why would developers like Greenlight? They system has no transparency, is fickle and impenetrable, is more closed than literally any other platform and rewards pandering and large marketing pushes over game quality. As a developer it's a complete crap-shoot. You put your game up, it doesn't have zombies in it, it fails and that's it. it doesn't help that every week there is a story about how a game that failed Greenlight gets on Steam anyway, a game that has a publisher still having to go through Greenlight, a game that bypasses the process for no discernible reason, etc.

Getting onto Steam is great for developers but the process is awful.

Edit: I don't understand why anyone would think that developers like Greenlight. Just because it's Valve?

[Kageru]

Anyone can make an XBox game with XNA and self-publish it. It's trivially easy to get on Xbox vs on Steam.

My understanding is that XNA is being discontinued and there is no self-publishing on the xbox one.

But even if you can't get your game on steam they do not own the PC platform as microsoft does the xbox one.

[Sky]

Gotta love that exclusive NFL contract and the blurred NFL footage.

The first rule of Madden is we don't talk about the exclusive contract. *sits in the corner weeping*

I'm pretty sure that's part of why I sat the last console generation out. Can we have 2k now?

[Paelos]

Might be actually helpful to link something when you mention that it's everywhere.

Gamasutra alone averages about one negative Greenlight piece a week, not counting unpromoted member blogs.

http://www.escapistmagazine.com/news/view/123973-Indie-Devs-Arent-Happy-With-Steam-Greenlight

See what I did there? I posted a link to something on point. That was really tough to do. 

[Margalis]

I assume people here are intelligent enough to use google. I honestly haven't seen anything positive about Greenlight past the first two weeks or so it was available. The consensus is that it's just not a good system.

Edit: It seemed to me that for anyone who has followed what developers think of Greenlight at all this shouldn't require citation. Perhaps my perspective is skewed because I am a developer, know other developers and read development-centric stuff. The reaction to Greenlight has been overwhelmingly negative past the initial honeymoon period.

[Paelos]

 

[UnSub]

The irony of getting on Greenlight is that you need an existing fan base to vote for your indie game.

Newell's also mentioned dumping Valve-covered curation and just throwing it out to the market to decide. So even Valve isn't a fan of Greenlight.

[Rendakor]

Ingmar's just been on a "Citation Needed" kick lately.

[Kageru]

The irony of getting on Greenlight is that you need an existing fan base to vote for your indie game.

Newell's also mentioned dumping Valve-covered curation and just throwing it out to the market to decide. So even Valve isn't a fan of Greenlight.

Which makes sense because most of the time there's lots of stress but no money in it. Getting gaming communities to decide which niche games are interesting, even if only within their community, would be win-win for valve. And they can look at the titles selling enough volume to be worth hosting on a "main" store.

[Soulflame]

I'm totally sincere, do those kinds of games exist? There's like 500,000 hardcore plat formers and shit, but I can't seem to find an old school Zelda style game. The nearest was Bastion combat wise, but it's still not really a Zelda game.


Binding of Isaac, never heard of it, doesn't look like what I am talking about from the steam video though.

Darksiders probably comes closest, but that's not indie.

[Rendakor]

3d dot Game Heroes.

[rk47]

I assume people here are intelligent enough to use google. I honestly haven't seen anything positive about Greenlight past the first two weeks or so it was available. The consensus is that it's just not a good system.

Edit: It seemed to me that for anyone who has followed what developers think of Greenlight at all this shouldn't require citation. Perhaps my perspective is skewed because I am a developer, know other developers and read development-centric stuff. The reaction to Greenlight has been overwhelmingly negative past the initial honeymoon period.

Greenlight greenlit Agarest.
I can now die knowing I played it on PC. Single-handed.

[Rendakor]

Is that the original Agarest?

[Kageru]

Never heard of it, but interesting... It would be really amusing if the PS4 won this round by being a bit more open to indies and co-operating with the PC market over going the exclusives route. They certainly don't have the money to compete with microsoft if it comes down to having to buy exclusive content.

I note their steam page says the company is now a "steam publisher" which I assume means the rest of their catalogue can skip the greenlight process if they want to port them.

Is that the original Agarest?

It looks like it is.

[Jeff Kelly]

The two biggest gripes with Greenlight were

1. The 'mob-mentality' and intransparent voting process by the community. Voter fraud was prevalent where developers created dummy accounts to promote their owm games over others. Vocal members of the community downvoted games for petty or silly reasons (too anime, too casual etc.) and it's still a big unresolved issue just how to design the community process in a way that doesn't allow developers to hijack it and disincentivises petty voting or mob mentality.

Even Gabe himself conceded that "Greenlight is a bad example of the election process". Valve is pretty slow to fix any of the inherent problems probably because they don't know how yet. This creates huge problems for devs though that want to promote their creation on steam and have to live with a broken voting system for now.

Gabe even said that in it's current form it would be "pretty easy to do away with Greenlight entirely because it creates a bottleneck'

2. The curation and support procerss by Valve. Greenlight is only open to developers without a publisher but how exactly a 'greenlighted' game gets on Steam and how the curation process by Valve works is not clearly communicated. This leads to crap like hundreds of fake submissions and copycats (see the World War Z debacle) that they tried to curb with the introduction of a $100 developer fee and also leads to crap like recently with Paranautical Activity by Code Avarice that was rejected from Steam.

The game dev had created a Greenlight page months before it got picked up by Adult Swim and Valve now blocks him on the grounds that 'indie developers aren't allowed to bypass the Grenlight process by getting a publsher'.

Theer is s recent Reddit post by the developer of 'Vox' that describes in detail the problems and issues with Greenlight and that it is 'pretty much a huge popularity contest' that you only win if you already have a huge fanbase (his words). Greenlight was also a topic on this years GDC.

All of the issues and the fact that they haven't been fixed pretty much since Greenlight's inception have led a lot of indie developers to abandon the platform.

[Merusk]

So the complaint is you have to have a marketing department to get things greenlit?  Seems sensible, though if you're just shoveling games on to the platform hoping to get noticed you'd have better luck on iOS and Android wouldn't you?

The "you can't bypass by getting a publisher" thing is totally bullshit. "Oh your game was good enough someone wants to publish it? Well we're not going to put it up on steam..."   

[Lantyssa]

Ingmar's just been on a "Citation Needed" kick lately.

[Citation Needed]

[Fabricated]

Lookit that responsiveness:



Also I guess they've had to underclock the APU by 100-200mhz compared to the PS4 due to heat issues. I hope the PS4 has some good industrial design because the XBONE was designed to have a big ugly ass case so they could cram a huge HSF into it and they're STILL having issues.

http://www.stfuandplay.com/story/content/microsoft-under-clocking-xbox-one-by-100-200mhz#.Ua626JyfiHt

[Jeff Kelly]

I like the phrasing in the news item. Underclocking: No big deal. At least you won't see a RROD (we assume). Nothing to see here..

Do they have any competent people left at all? Designing a case with good airflow and efficient thermal management ain't witchcraft.

You have to invest a little bit more into your case design than just 'generic injection molded box' though.

[Kageru]

The "you can't bypass by getting a publisher" thing is totally bullshit. "Oh your game was good enough someone wants to publish it? Well we're not going to put it up on steam..."   

To an extent. From listening to an interview with the developer it sounded like the publisher (Adult swim) had basically worked out that offering titles the chance to skip the greenlight process was a good low-effort money-maker. Which of course doesn't excuse greenlight from creating those sort of conflicts.

[Jeff Kelly]

XBox go home.

Doesn't matter how often I read/hear it it's still funny. The fact that during the whole product development process nobody with decision making powers opted to phrase it differently speaks volumes about MS's corporate culture.

That phrase alone will power several years worth of bad jokes and memes

[Miasma]

Lookit that responsiveness:



Also I guess they've had to underclock the APU by 100-200mhz compared to the PS4 due to heat issues. I hope the PS4 has some good industrial design because the XBONE was designed to have a big ugly ass case so they could cram a huge HSF into it and they're STILL having issues.

http://www.stfuandplay.com/story/content/microsoft-under-clocking-xbox-one-by-100-200mhz#.Ua626JyfiHt

For a shill piece it was pretty badly edited and performed but that gif is misleading.  The tv starts at the home page and the time it's doing nothing she's saying "xbox snap live tv".

My favorite line: "You don't have to remember the channel anymore, you can just say xbox channel 13".

[Fabricated]

Also some speculation on NeoGAF is that Microsoft is having yield problems due to needing eSRAM?

But it's NeoGaf. Don't read or post on NeoGAF.

[Jeff Kelly]

That would surprise me. Every microcontroller for embedded systems comes with embedded SRAM and when you get to the ARM Cortex M models it's not that rare to get 4 or more MB of eSRAM.

Having one or several cores with a certain amount of SRAM on die that can then be extended with SDRAM over an external bus interface is pretty much standard in embedded systems so I'd assume that the foundries won't have any issues with yield since they should have a lot of experience with that setup even at the sub 20 nm scale. TSMC and other big foundries pretty much live on embedded because it's orders of magnitude more business than desktop processors.

I'd be curious to see the cost calculation for the XBox One system-on-chip. SRAM is expensive because it uses three transistors for every bit where SDRAM only uses one transistor and one capacitor so the manufacturing process is more expensive and SRAM needs a larger area per bit than SDRAM so you need a larger die which also drives up cost.

They say that using GDDR3 memory and 32 (?) Megs of eSRAM is cheaper than going with GDDR5 memory but knowing how much SRAM costs per Meg I can hardly believe it.

[Phred]

Vocal members of the community downvoted games for petty or silly reasons (too anime, too casual etc.) and it's still a big unresolved issue just how to design the community process in a way that doesn't allow developers to hijack it and disincentivises petty voting or mob mentality.

For such a great game designer it's kind of perplexing how greenlight shipped as is. You complain about people's reasons for voting yet read the question asked on the vote.

If that isn't asking for a value judgement what is?

[Margalis]

So the complaint is you have to have a marketing department to get things greenlit?  Seems sensible, though if you're just shoveling games on to the platform hoping to get noticed you'd have better luck on iOS and Android wouldn't you?

The problem is you need a marketing department before you have a game.

Typically the way greenlighting works is you make a pitch, it gets greenlit, you make the thing, then you market it. The way this works is you market the game first, then make it. You need a big marketing push before you actually do the work, and it also leads to pandering stuff like throwing in voxel graphics, zombies and promises you may or may not deliver on, because that's what the Greenlight audience eats up.

It basically takes all the bad stuff about focus-testing and doubles down on it - snap judgements based on superficial readings. It also is strongly biased towards games people know they want (AKA iterations on successful products) rather than games they don't know they want. (But still want)

I would also point to game like EvoLand - it has a premise that could be interesting but is by all accounts a pretty boring game. I also have trouble believing it has sold well. Why did that game succeed in Greenlight and reach Steam? It's not particularly good nor popular. I would assume because they either over-promised on a neat concept or effectively rallied an existing community.

If you look at the games that have made it through Greenlight and are being sold today it looks like a random collection of games with no regard to quality or popularity. If someone told me the names were picked out of a hat I would believe it.

[koro]

This is going to be interesting.

[Paelos]

[Margalis]

There were numerous rumors that MS was way behind on XB1 development, evidence of a hastily arranged PR plan, a conference that showed almost nothing and was puzzlingly short, this canceling of a Q&A event, the underwhelming video of the system in action for real...either MS is struggling or they are doing their best to make everyone think that. They just don't seem prepared on any level.

[Quinton]

That would surprise me. Every microcontroller for embedded systems comes with embedded SRAM and when you get to the ARM Cortex M models it's not that rare to get 4 or more MB of eSRAM.

It's extremely rare to see Cortex M0/M3/M4 MCUs with more than 128-256KB of SRAM.  More than 512KB is virtually impossible to find.   Embedded memory and flash add significant cost to MCUs.

Even much larger SoCs which use SRAM for low power modes (avoiding the need to take external SDRAM out of self-refresh) or high performance graphics/video memory rarely have more than 1MB available on die.  Larger CPU/GPU caches tend to be a better investment for all that die area.

EDIT:  Just did a quick survey on digikey and it's actually more restrictive than that.  I'd say >128KB is where things really drop off fast in this space.
M0/M0+/M3/M4/M4F MCUs -> 40289 parts listed
limit to 128K+ SRAM -> 541 parts listed
132K+ -> 202 parts
200K+ -> 18 parts
264K ->  7 parts
1.3MB -> 1 part

[Jeff Kelly]

Microsoft has stated for a few years now that they had plans to extend the lifecycle of this gen's console to ten years. The Kinect was explicitely created to give the XBox a second wind and to prolong the life of the 360 for a few years longer.

MS took the success of the Wii and iOS as a sign that graphics no longer was king and they thought they could capitalize on that by introducing an accessory that shifts the focus away from quicker, higher, farther.

They also got signals from EA and other big publishers that they also didn't necessarily want a new conole either. The publishers feared that they wouldn't be able to quickly get to a level of adoption of the new console that would allow them to continue with their buiness model that needed 5 million+ customers. Especially since the new gen likely makes development even more expensive.

Then the Kinect failed, the novelty factor of the Wii had reached its end and the WiiU tanked and Sony actually developed a strategy that worked and started developing the next Playstation.

It wouldn't surprise me if Microsoft needed to scramble and quickly ome up with something to not being left empty handed. Sony's dev kits had been going around for some time when MS approached devs.

[Jeff Kelly]

That would surprise me. Every microcontroller for embedded systems comes with embedded SRAM and when you get to the ARM Cortex M models it's not that rare to get 4 or more MB of eSRAM.

It's extremely rare to see Cortex M0/M3/M4 MCUs with more than 128-256KB of SRAM.  More than 512KB is virtually impossible to find.   Embedded memory and flash add significant cost to MCUs.

Even much larger SoCs which use SRAM for low power modes (avoiding the need to take external SDRAM out of self-refresh) or high performance graphics/video memory rarely have more than 1MB available on die.  Larger CPU/GPU caches tend to be a better investment for all that die area.

L1 and L2 cache memories are usually real (6T)-SRAM, the only difference being that L1 is usually on-die and L2 can be both on or off die. Even if it's designed as an SOC I'd consider it to be on die since it's cut from the same wafer anyway.

For example: Intel's Sandy Bridge CPUs use 6T-SRAM for the L1 and L2 cache and 8T-SRAM for the up to 8 MB of L3 Cache.

If you need ultra low current consumption and low voltage/power you use 8T-SRAM because 6T-SRAM suffers from noise at low voltages and can no longer be reliably read out.

You could use 1T-SRAM which is slower but significantly cheaper due to the smaller area per bit (which reduces die size and therefore cost of manufacturing) and some architectures with really large L3 caches (like IBMs Power7) do that.

If I look at current DSP platforms like Analog Devices' SHARC then it's not that unusual to find 5 MBit or more of on-die 6T- or 8T-SRAM (for less than $20 at 1k units), The Cortex A8 (my mistake I shouldn't have said Cortex M) supports up to 4 MiB of on-die SRAM and for example Renesas's V850 core can have up to 3072 kB of on-die SRAM. Even the newer Cortex M4 cores already go up to 1 MB or 1.5 MB.

Depends on your application though if you're willing to pay the price for it.

You can also buy SRAM chips up to GBIt capacity in 6T- and 8T-variants if you want to extend your capacity with external mem.

So I'd assume that the large foundries have the process down even at 18 micron or 13 micron so yield shouldn't be an issue even if you put 32 MB of SRAM on your die.

As far as cost is concerned I can't really imagine that a solution with 8 GB of GDDR3 SDRAM and 32 MB of 6T-SRAM (which is basically a Cache anyway) is more cost effective than Sony's 8 GB of GDDR5 memory.

The aforementioned V850 with 3072 kB of SRAM will probably be around $10 to $15 at 1M units so I don't even want to know what a three-core X86-64 CPU with 32 megs of 6T- or 8T-SRAM would likely cost. (Look at the prices of server-grade or extreme edition CPUs if you want an impression)

So the 32 MB of embedded SRAM are most likely 1T-SRAM (sometimes incorrectly called eDRAM) which should have even less problems with yield. Foundries already make millions of  CPUs with 32 MB or more of L3 cache in 1T-SRAM technology.

[Jeff Kelly]

I read the linked NEOgaf thread (which might have cost me a few IQ points) and if the rumors turn out to be true it is as follows:

Both Sony and MS use a pretty similar CPU and GPU, ATI/AMD based.

Microsofts vision was always the XBox as set top box and media center. Their approach required a certain performance and memory baseline. They were also very late to the game.

At the time the design of the Xbone was planned MS already knew they needed 8 Gig of RAM to do all of the Kinect and TV! TV! SPORTS! SPORTS! features (3 Gig just for the OS, remember) but GDDR5 RAM was too expensive and not widely available so using 8 GB of GDDR5 was considered to be too expensive but they couldn't make do with 4 or 2, they needed 8. Making Kinect part of the base system didn't help things since the Kinect is also eating up a significant part of the budget.

So they decided to use GDDR3 mem which is significantly slower. The rumors that Sony would use GDDR5 memory caught them on the wrong foot and so they needed something to at least get to closer to the PS4's performance. So they decided to integrate a giant 32 MB SRAM cache on the die which could boost performance quite a bit. Devs would have to put in more work than on the PS4 but not that much that it wasn't feasible (so not a PS3 situation)

They don't use 1T-SRAM though like I thought but 6T-SRAM (real SRAM) 6T SRAM uses 6 transistors per bit and so takes up more space than SDRAM per bit (4 times as much, SDRAM is only 20% the area of SRAM per bit). They also use a 45 micron process instead of 32 or 22 micron one so the SRAM alone is HUGE.

At 45 nm one MBit of 6T-SRAM is .52 mm^2 so the 32 MB/256 MBit of eSRAM on the Xbone alone is 133.12 mm^2. That is an area  with an edge length of 11.5 mm (.5 in * .5 in) Combined with the 3 core AMD CPU and GPU this leads to a huge ass die. If they had used quad-density 1T-SRAM which offers similar if slightly less performance it would have been 20% of the size instead.

So they get fewer dies out of each waver but the number of defects per waver remains the same. So they have on average more dies they can't use. This significantly lowers the yield. They had hoped that by the time the XBox comes out process technology had evolved to allow them to shrink the die to 32 or 22  nm effectively increasing the number of chips per wafer.

They can't scrap the SRAM because then the Box would offer crappy gaming performance not much better than the WiiU and they can't move to GDDR5 memory because they would have to completely scrap the board design and start over from scratch. So they apparently decided to reduce the clock speed of the GPU/SRAM. Reducing the clock speed or deactivating parts of the core improves yield because you might be able to use chips that might not have worked at the original speed because of defects.

Rumor has it that they will reduce clock speed of the GPU by 200 MHz to increase yield which would reduce the lower gamming performance of the XBone even further and might even lead to games no longer being able to run at 1080p. GPU and SRAM are so tightly connected that the decrease in clock speed will also affect the SRAM (it runs at the same clock as the GPU) and so the whole system performance will be significantly affected.

The alternative would have been to scarp the current box and do a respin but that would have delayed the XBone by six months a delay they can't afford when Sony comes out this year.

Sony on the other hand never wanted to do anything but a game console so they started with a straightforward design but only 2 GB of GDDR5 betting on the chips getting cheaper as more graphics cards use them. They are now at a price point where they can offer 8 GB of GDDR5 Ram even though that didn't seem possible half a year ago. This also caught MS completely by surprise.

If the rumor is true and the reduction in clock speed is coming then the PS4 will probably be twice as powerful as the XBone as far as raw computing and graphics power is concerned.

The same source also claimed that the XBone OS is pretty much in an alpha state right now and that MS is delayed on OS, games and certain other elements of the XBone which might have led to the uneven presentation

[Kageru]

The comments from the lady in the video quoted above more or less admitted they are way behind on the software side. "My team is working round the clock and we've come a long way and will be even better by launch" is quite an admission that they are hitting time pressures.