Whoever wrote that piece hasn't a clue.
Predict: The Next Generation Console Tech
- Thread starter Acert93
- Start date
- Status
Blade47167
Newcomer
And I for one hope they don't use that feature. I hate the Wii controls. I never feel like I'm in the game and always have to worry about a little swing here and there. Not to mention accuracy is sub-par. Say its the tv I've played it on but I've played on 5 different TV's now to the same effect.
I much prefer the analog stick or the d-pad when playing pretty much any game. I don't mind the six-axis as it doesn't have to be used. My favorite implementation so far is probably using it to hit someone with the end of your gun and the R&C Tornado. Either way making a controller like that for the next Playstation or X-Box would be a mistake.
Agree. While waggle has been a successful venture for Nintendo, there's a large group of gamers who aren't looking for waggle in their games. Looking at previews for wii fit makes me cringe.
this article seems full of shit...
I can't see next gen running 64bit OS so My bet would be on 4GB of GDDR3 in a UMA design.
For the system design... it's way to early to know.
I tend to think that MS will do its best to convince Intel for the cpu and go for a full if Intel solution proves efficient, but it will be a tough battle.
the second option is AMD/ATI, but I am that confident in AMD capability to push something performant at this time, maybe bulldozer and upcoming gpu are worth it.
The way the article is trying to guess is ridiculous, shinking will get harder next gen, and everybody learn that pushing a to expansive product is not a good idea.
Between Intel is already @45nm... I can help it if 360 do well and finish with a healthy market share, Intel will jump in cause it will help them to gain groud as a gpu gpgpu vendor!
My bet wintel again and again.
I can't see next gen running 64bit OS so My bet would be on 4GB of GDDR3 in a UMA design.
For the system design... it's way to early to know.
I tend to think that MS will do its best to convince Intel for the cpu and go for a full if Intel solution proves efficient, but it will be a tough battle.
the second option is AMD/ATI, but I am that confident in AMD capability to push something performant at this time, maybe bulldozer and upcoming gpu are worth it.
The way the article is trying to guess is ridiculous, shinking will get harder next gen, and everybody learn that pushing a to expansive product is not a good idea.
Between Intel is already @45nm... I can help it if 360 do well and finish with a healthy market share, Intel will jump in cause it will help them to gain groud as a gpu gpgpu vendor!
My bet wintel again and again.
Heh, that article sounds like me five years ago when I didn't have a clue about hardware 
CPU GHz means next to nothing. Just for fun compare the Spec results of 1.6Ghz Itanium2 vs 4.7GHz Power6. Core count is also a tricky thing and will highly depend on what kind of architecture it would use (beefy OoO x86 vs in-order Power).
GPU GHz is another meaningless meter. I also doubt it would have that many GPUs, at best there could be 2 GPUs in MCM. Of course there will be a whole lot of shader processors in each one of those.
I'd put RAM to around 2-4GB GDDR5 or perhaps XDR2 if it isn't too expensive. Not sure if UMA or NUMA. NUMA would surely provide additional bandwidth and lower latency but wouldn't be as flexible.
There will certainly be an optical drive, though I have no idea what exactly. It will depend on who survives, BD or HD-DVD.
HDD won't probably be bigger than 1TB, it could be much less. IIRC HDD manufacturers have been talking about 4TB drives coming in 2012 or later, no way so big one would be in a console that fast. In addition there could be a couple of gigs of flash memory acting as a better-than-HDD cache.
CPU GHz means next to nothing. Just for fun compare the Spec results of 1.6Ghz Itanium2 vs 4.7GHz Power6. Core count is also a tricky thing and will highly depend on what kind of architecture it would use (beefy OoO x86 vs in-order Power).
GPU GHz is another meaningless meter. I also doubt it would have that many GPUs, at best there could be 2 GPUs in MCM. Of course there will be a whole lot of shader processors in each one of those.
I'd put RAM to around 2-4GB GDDR5 or perhaps XDR2 if it isn't too expensive. Not sure if UMA or NUMA. NUMA would surely provide additional bandwidth and lower latency but wouldn't be as flexible.
There will certainly be an optical drive, though I have no idea what exactly. It will depend on who survives, BD or HD-DVD.
HDD won't probably be bigger than 1TB, it could be much less. IIRC HDD manufacturers have been talking about 4TB drives coming in 2012 or later, no way so big one would be in a console that fast. In addition there could be a couple of gigs of flash memory acting as a better-than-HDD cache.
archangelmorph
Veteran
No next gen console released in 2012 will ever ship without an optical drive..
Killing the retail disc market is just not an option for retailers (preowned games & hardware is where they make their money) & since this fact makes them quite a significant stakeholder in the development process i's pretty obvious that at least some form of physical media (not sure if there is anything cheaper and with greater storage capacity than optical discs?) will remain...
Possibly Holographic Versatile Discs?
Killing the retail disc market is just not an option for retailers (preowned games & hardware is where they make their money) & since this fact makes them quite a significant stakeholder in the development process i's pretty obvious that at least some form of physical media (not sure if there is anything cheaper and with greater storage capacity than optical discs?) will remain...
Possibly Holographic Versatile Discs?
Last edited by a moderator:
I agree. Downloadable games may exist, but it won't be a replacement. Everyone expects game sizes to balloon beyond what DVD9 currently offers (for the 360). I don't know if Joe Blow wants to wait several hours maxing his bandwidth to get a game. Not everyone has a 10Mbit+ connection or even the monthly bandwidth allowance for 10GB+ per game. It's probably more important not to just assume that everyone will hook up the console to the internet either - maybe a majority do, but in the sale of millions of consoles, even leaving out 5-10% is a pretty decent chunk.
It certainly would be interesting to see which format Microsoft ends up using. Perhaps by 2012, it will be feasible to use a dual-format drive to read both Blu-Ray and HD DVD, thereby leaving it up to publishers to decide on the media.
It certainly would be interesting to see which format Microsoft ends up using.
Perhaps by 2012, it will be feasible to use a dual-format drive to read both Blu-Ray and HD DVD, thereby leaving it up to publishers to decide on the media.You could get around that by mandating seperate traditional and motion sensing control schemes for all games and then allowing the gamer to turn off the motion sensing aspect if desired.
My system of guestimation (process node and technology advancement) we did here was a lot better, and we did that over a year ago. That article is horrible!!
Memory footprint, bandwidth (both chip-to-chip and chip-to-memory), and coherence between local memory pools as well as system memory is going to be large issues. Another issue is going to be load times; a system with 4GB of memory and an optical delivery device is going to be painful. I believe the ratio between system memory footprint and drive speed actually is worst this gen than the previous. e.g.
Xbox: 2x-5x CAV (2.6MB-6.6MB/s); 64MB memory = 10-to-24 seconds
Xbox 360: 12x DVD (15.85MB peak); 512MB memory = 32 seconds (even higher at non-peak)
An 8x Blu-ray drive (36MB/s) is going to take over 110 seconds to fill 4GB. You don't need to fully populate your memory to begin playing a lot of games, but when you are looking at 2 minutes in ideal circumstances to load content to memory... ouch. Streaming content is going to be a big challenge. I would bet on some permenant storage, even possibly a fair chunk of flash-like memory as a cache.
The next 2 years will tell us a lot. There are a lot of rumblings of embedded memory migrating into GPUs in more native/intuitive ways (i.e. more friendly that a large pool that may require software based tiling). I wouldn't be surprised if MS and Sony diverge some. e.g.
Sony: Cell focused design (e.g. two 2/32 Cells) with a "traditional" GPU
MS: GPU centric design with the GPU(s) being a more general resource accessible by the CPUs through extensions (i.e. the direction AMD is going)
Of course there is IBM and Intel... and Samsung... looming.
Personally I am curious if MS/Sony will take the "Nintendo" route... and I think the biggest question right now isn't the hardware, but the content generation tools. Asset creation, design, scripting, etc need to become more accessible. You can have all the power in the world, but if the tools don't allow your entire team to extract potential it is lost.
The place for engines (id Tech, CryEngine 2, Unreal Engine, Source, Gamebryo, etc) and middleware (Havok, Novodex, Speedtree, etc) will become more prominent. I would be surprised if MS/Sony didn't make a strong play toward either aquiring and/or developing a series of APIs and tools specifically designed to offer a general framework to develop games. Many consumers are indicating they are seeing diminishing returns and art is playing a huge factor in what titles look good and which look bad.
Memory footprint, bandwidth (both chip-to-chip and chip-to-memory), and coherence between local memory pools as well as system memory is going to be large issues. Another issue is going to be load times; a system with 4GB of memory and an optical delivery device is going to be painful. I believe the ratio between system memory footprint and drive speed actually is worst this gen than the previous. e.g.
Xbox: 2x-5x CAV (2.6MB-6.6MB/s); 64MB memory = 10-to-24 seconds
Xbox 360: 12x DVD (15.85MB peak); 512MB memory = 32 seconds (even higher at non-peak)
An 8x Blu-ray drive (36MB/s) is going to take over 110 seconds to fill 4GB. You don't need to fully populate your memory to begin playing a lot of games, but when you are looking at 2 minutes in ideal circumstances to load content to memory... ouch. Streaming content is going to be a big challenge. I would bet on some permenant storage, even possibly a fair chunk of flash-like memory as a cache.
The next 2 years will tell us a lot. There are a lot of rumblings of embedded memory migrating into GPUs in more native/intuitive ways (i.e. more friendly that a large pool that may require software based tiling). I wouldn't be surprised if MS and Sony diverge some. e.g.
Sony: Cell focused design (e.g. two 2/32 Cells) with a "traditional" GPU
MS: GPU centric design with the GPU(s) being a more general resource accessible by the CPUs through extensions (i.e. the direction AMD is going)
Of course there is IBM and Intel... and Samsung... looming.
Personally I am curious if MS/Sony will take the "Nintendo" route... and I think the biggest question right now isn't the hardware, but the content generation tools. Asset creation, design, scripting, etc need to become more accessible. You can have all the power in the world, but if the tools don't allow your entire team to extract potential it is lost.
The place for engines (id Tech, CryEngine 2, Unreal Engine, Source, Gamebryo, etc) and middleware (Havok, Novodex, Speedtree, etc) will become more prominent. I would be surprised if MS/Sony didn't make a strong play toward either aquiring and/or developing a series of APIs and tools specifically designed to offer a general framework to develop games. Many consumers are indicating they are seeing diminishing returns and art is playing a huge factor in what titles look good and which look bad.
Isn't hardware cap going to have an impact next gen because we're approaching the minimum size of traditional fabrication processes? If we have 45nm end of this gen, there won't be much further to go with process shrinks to get more on a chip. Then the only solution for more power will be more silicon = more cost. If console design involves one eye clearly on the <$200 price-point and not losing money, that'll determine a maximum silicon budget at the lowest node at that price-point, which by my advanced parametric guessifications is well below the current ~600 mm^2 for CPU+GPU. What's 500 mm^2 now needs to get down to...100 mm^2 to be mainstream say. If you only have one shrink available in your console's life, you're going to have to start at twice that. That'd be like XB360 starting this gen with a 200mm^2 budget. What CPU and GPU would they have got for that? Pretty uninspired!
I think hardware is going to be limited, and it'll be everything else needed to make the difference. It'll also raise questions about future consoles, when hardware can't improve within a fixed price envelope. I expect the future then might be geared towards server based gaming, with consoles being more like dumb terminals. Technology invariably turns full circle, and we haven't done that yet!
Good points Shifty, something we discussed in the other thread (anyone want to dig that up?) If the next consoles come in at 32nm (~8x reduction), the roadmap to 22nm and 16nm will really dictate a lot. This also may require some other forms of revenue models (advertising?) to help reduce hardware costs being passed onto consumers. I believe someone (Gubbi?) in the past commented that if a wall materializes one thing that may shift is pricing as R&D will become less of a component of the cost. If such a situation did occur, could a mass-production off the shelf part be more cost effective by virtue of mass production?
This also may be a factor in who MS and Sony pair with on the manufacturing side. e.g. Right now it seems like Intel is substantially ahead of the market. Lets say Intel charges a 30% premium per chip, but you are getting those chips on a lower process node (smaller, less power/heat) this could be a sizable advantage.
We are at the tail end of 2007 and my guestimate would be we won't see a new console from Sony or MS until 2011. While Carmack indicates the wheels are already turning on the new platforms, I don't think we will have a good picture of technology and process nodes until 2009.
This also may be a factor in who MS and Sony pair with on the manufacturing side. e.g. Right now it seems like Intel is substantially ahead of the market. Lets say Intel charges a 30% premium per chip, but you are getting those chips on a lower process node (smaller, less power/heat) this could be a sizable advantage.
We are at the tail end of 2007 and my guestimate would be we won't see a new console from Sony or MS until 2011. While Carmack indicates the wheels are already turning on the new platforms, I don't think we will have a good picture of technology and process nodes until 2009.
It was closed: http://forum.beyond3d.com/showthread.php?t=31379
Perhaps some thread pruning and a re-opening may be in order.
- Status
Similar threads
- Locked
