[AlStrong]

Originally Posted by Squilliam

To me it makes more sense given their 'talent' lies in the GPU arena given their close relationship with Direct X and GPU development that they would focus more individual attention on that part as they did this generation.

P.S. Does IBM stand a chance of making the nextXbox CPU? Their POWER 7 does look like quite the sturdy chip at 45nm, so maybe it'll fit the power budget at 32nm or lower and everyone loves redunkulous cache sizes with ED-RAM, right?

Power7 probably doesn't make too much sense for them. As you say, they're best on the GPU side. I mean, aside from recent endeavours to shove as much graphics tasks to Cell, is there anything that really needs a massive CPU? (There is the mention of *ahem* physics in the GPU Architecture Job listing. ) So... hey, here's an ultra-beefy GPU with unified shaders, lol-threads, textures, triangles, tessellation, etc.... Just throw things at it, it'll take care of the scheduling and threading and what-not. Don't worry about needing to make up performance by off-loading to the CPU... Of course, I am speaking within the confines of a CPU-GPU budget for a console. And well, this does sound awfully familiar now doesn't it.

There is also the PC-console porting to consider as well. Just thinking aloud, but how many devs have actually made significant use of the CPU on PC based on SPU-graphics work, not just multi-threading needs as dictated by Waternoose's triple hyperthreaded cores. (Just in case it needs to be said, Tile-Based Deferred Shading can be done on the GPU http://visual-computing.intel-resear...red_rendering/ )

[Ruskie]

Originally Posted by AlStrong

Power7 probably doesn't make too much sense for them. As you say, they're best on the GPU side. I mean, aside from recent endeavours to shove as much graphics tasks to Cell, is there anything that really needs a massive CPU? i.e. Here's an ultra-beefy GPU with unified shaders, lol-threads, textures, triangles, tessellation, etc.... Just throw things at it, it'll take care of the scheduling and threading and what-not. Don't worry about needing to make up performance by off-loading to the CPU... Of course, I am speaking within the confines of a CPU-GPU budget for a console. And well, this does sound awfully familiar now doesn't it.

There is also the PC-console porting to consider as well. Just thinking aloud, but how many devs have actually made significant use of the CPU on PC based on SPU-graphics work, not just multi-threading needs as dictated by Waternoose's triple hyperthreaded cores.

Agreed.Nice,beefy,juicy GPU with a lot of bells and whistles would be another great design choice by MS They really outdid themselves with 360,its really designed for developers...to bad it had major RROD flaw.

[hoho]

They could just design a GPU that is about as general-purpose as Cell SPE's are and get nearly perfect setup. No need for super-fast CPU then

[liolio]

I'm not expecting Ms taking this route but if SPUs shine at something it's perfs per Watts they are pretty outstanding in this regard. I would bet they grossly outdo GPU at "mundane" calculations.

I don't expect either SOny to take that road but I've a feeling that something really "outlaw ish" in regard to nowadays standard hardware designs could surprise, in efficiency and performances.
(I think of those posts from Fafalada and Panajev2001a as back up for the assumption).

Another things so we look @ late 2013 or 2014 for next generation hardwares, it kind of sucks... (or better be impressive).

[AlStrong]

Originally Posted by liolio

I'm not expecting Ms taking this route but if SPUs shine at something it's perfs per Watts they are pretty outstanding in this regard. I would bet they grossly outdo GPU at "mundane" calculations.

I don't expect either SOny to take that road but I've a feeling that something really "outlaw ish" in regard to nowadays standard hardware designs could surprise, in efficiency and performances.
(I think of those posts from Fafalada and Panajev2001a as back up for the assumption).

mmm.... indeed. I do wonder how the current gen CPUs actually compare to more recent designs in the low power sector (Atom, Llano, ARM). Of course, even 45nm XCPU or Cell are still going to be an order of magnitude greater in power, but what if one scaled those particular designs up

Another things so we look @ late 2013 or 2014 for next generation hardwares, it kind of sucks... (or better be impressive).

At least we'll see what happens on the memory side of things. It's been pretty boring with GDDR5 for awhile. No hints on the next iteration.

At least as far as the 360 design philosophy went, it would be rather interesting to see the CPU be attached to the GPU right off the bat (because it'd start from a low-power design), and they could budget more for a potential off-die eDRAM.

-----------------

On a side note, regarding the job listing for future 360 developments (3rd link), I suppose we're not really done with the eDRAM. Based on measurements, we're still at 65nm for the Slim.

http://forum.beyond3d.com/showpost.p...&postcount=239

The expected scaling from 64mm^2 @ 80nm to 65nm would be 42mm^2. We got ~45mm^2. Fair enough. Nothing is perfect, plus there are the ROPs and such to build. A hypothetical shift to 40nm would put that down to around 17-20mm^2. (TSMC 40nm production began in 2010)

hm.... A bit curious the article mentions 45nm, not 40nm for the CGPU. Anyways, scaling up for next gen using a similar design would still be... peculiar to imagine.

a) double or triple the ROPs, increased functionality e.g. texturing
b) quadrupling die size (@ imaginary 40nm) would bring it back to 360's original 80mm^2 for eDRAM @ 90nm.

30-40MB eDRAM @ 40nm, ~80mm^2

Give it a few more years for 28nm to ramp up... Could maybe hit 70-75MB @ 80mm^2 *shrug* Fun to think about it.

-----------------------------
Self-reminder to read sometime

[Rangers]

Originally Posted by eastmen

Patcher is claiming a 360 price drop in june. How many more does MS really have , i believe the core verison of the system will now be at $150 if the price drop is true. I think once we hit $100 for the low end its time for ms to put out something new

And yet the ASP of X360 was $307 or something last month, of course helped by Kinect SKU's, but in the end an sku is a sku.

Seems a lot of price room left to me...

I would think ideally oneday MS might want to get rid of non-kinect SKU's, if they can get the Kinect sku price low enough. Ideally. So that takes a long time too.

[fehu]

Originally Posted by AlStrong

Power7 probably doesn't make too much sense for them. As you say, they're best on the GPU side. I mean, aside from recent endeavours to shove as much graphics tasks to Cell, is there anything that really needs a massive CPU? (There is the mention of *ahem* physics in the GPU Architecture Job listing. ) So... hey, here's an ultra-beefy GPU with unified shaders, lol-threads, textures, triangles, tessellation, etc.... Just throw things at it, it'll take care of the scheduling and threading and what-not.

i'm nerdixcited about more general purpouse gpu, but when i ask there's always the objection "you want to use every piece of silicon in the gpu only for graphic stuff because is this this that sell games"

nerdeluded

[Rangers]

Originally Posted by fehu

i'm nerdixcited about more general purpouse gpu, but when i ask there's always the objection "you want to use every piece of silicon in the gpu only for graphic stuff because is this this that sell games"

nerdeluded

Hmm, are people more excited when cell is used for "general purpose", or graphics as in Dice presentations?

There's your answer

[liolio]

Originally Posted by AlStrong

mmm.... indeed. I do wonder how the current gen CPUs actually compare to more recent designs in the low power sector (Atom, Llano, ARM). Of course, even 45nm XCPU or Cell are still going to be an order of magnitude greater in power, but what if one scaled those particular designs up

Cell (or more accurately SPUs) is in a class of its own at culling it beats a core i7 @2,6GHz badly (almost twice as fast) vs Xenon lose two three times faster and for the later I guess shading is not an option. Something that stroke me after watching dice presentation is that the Cell has actually 8 SPUs SOny decided to disabled one and relies on one SPU (and something) for security), I can only wonder what the figures would be if Dice ran their code on a full blown Cell
When we considering perfs of the Cell in the PS3 actually we're grossly considering 1 PPU + 5SPU. Going from 5 or even 6 to 8 SPUs is not a tiny jump we're looking at least +33% improvements.

OK I pass on the end of your post my wife is dragging me to bed.

[AlStrong]

Originally Posted by liolio

OK I pass on the end of your post my wife is dragging me to bed.

ಠ_ಠ

[Squilliam]

Originally Posted by AlStrong

mmm.... indeed. I do wonder how the current gen CPUs actually compare to more recent designs in the low power sector (Atom, Llano, ARM). Of course, even 45nm XCPU or Cell are still going to be an order of magnitude greater in power, but what if one scaled those particular designs up

I guess we'll find out when we get the first reports back as to the performance of the NGP CPU, wouldn't we? I too would like to know however...

At least we'll see what happens on the memory side of things. It's been pretty boring with GDDR5 for awhile. No hints on the next iteration.

GDDR5? XDR2? You have the proven performance and price effectiveness of GDDR5 vs Unproved XDR2, any guesses if it were be between these two?

At least as far as the 360 design philosophy went, it would be rather interesting to see the CPU be attached to the GPU right off the bat (because it'd start from a low-power design), and they could budget more for a potential off-die eDRAM.

Given TSMCs 40nm capacity and ability to make ED-RAM on 40nm and with a process roadmap good to 28nm according to them, could we see the next Xbox with ED-RAM combined at 28nm on a bulk process? I think if this is a possibility we may see the next Xbox 360 revision having all three chips combined on the 40nm process.

[liolio]

Originally Posted by AlStrong

ಠ_ಠ

lol

I don't think that you're wrong actually I was just making a difference beween xenon and its three Px, the PPU in Cell, Atom, ARm, llano and SPUs. If we speak of only the std CPus I would bet that only llano is signifanctly faster (per cycle) than the other architectures. How Arm will do in the future? As everybody I want to know but I'm cautiously optimistic Intel and AMD faced problems and failures while trying to raise the bar in performances.

At least we'll see what happens on the memory side of things. It's been pretty boring with GDDR5 for awhile. No hints on the next iteration.

At least as far as the 360 design philosophy went, it would be rather interesting to see the CPU be attached to the GPU right off the bat (because it'd start from a low-power design), and they could budget more for a potential off-die eDRAM.

Well GDDR5 is boring but even on a 128bits bus it seem to offer enough bandwidth for the resolution next gen systems may be aimed at especially as deferred renderer are getting used more and more.
-------------------
Interesting a slim² 360 may be on the way, I'm not sure they will consider putting Xenon xenos and the daughter die on the same chip, but say they keep xenon+xenons on the same die @45nm (SOI) and move edram from 65nm to 45m would they have room to put this on top of RAM modules on the same package?

[liolio]

Originally Posted by Squilliam

Given TSMCs 40nm capacity and ability to make ED-RAM on 40nm and with a process roadmap good to 28nm according to them, could we see the next Xbox with ED-RAM combined at 28nm on a bulk process? I think if this is a possibility we may see the next Xbox 360 revision having all three chips combined on the 40nm process.

I wonder about it MS just swallowed the RD cost to implement xenos on SOI 45nm process. Move everything to bulk @ a process that won't give them much in regard to the number of dies per wafers seems like superflous expanse at this point. Moving Xenon+Xenos on GF 32nm process late in 2012 as well as edram to 40/45nm late in 2012 seems like a more likely scenario to me.

[liolio]

Originally Posted by Rangers

Hmm, are people more excited when cell is used for "general purpose", or graphics as in Dice presentations?

There's your answer

What got me exited is what Intel engineers working on Larrabee think of this and what they consider as design if they don't give up on the X86 ISA. They went in such lengths either on the software and hardware front, four way SMT, "Software multithreading", software rasterizer, super wide SIMD, etc.
I feel more and more like they overshoot what was needed to make X86 relevant as a counter to GPU computing. I hope they won't give up on the project but I badly wondering at what the team in charge has to be thinking now.

[Entropy]

Originally Posted by liolio

What got me exited is what Intel engineers working on Larrabee think of this and what they consider as design if they don't give up on the X86 ISA. They went in such lengths either on the software and hardware front, four way SMT, "Software multithreading", software rasterizer, super wide SIMD, etc.
I feel more and more like they overshoot what was needed to make X86 relevant as a counter to GPU computing. I hope they won't give up on the project but I badly wondering at what the team in charge has to be thinking now.

Does Intel need a counter to GPU computing?
Do they lose a lot (any?) revenue due to GPU computing?

[liolio]

Originally Posted by Entropy

Does Intel need a counter to GPU computing?
Do they lose a lot (any?) revenue due to GPU computing?

So what was the point of investing in Larrabee? May they wanted to kill it while it is/was still a potential treat.

[MarkoIt]

More:
http://www.linkedin.com/jobs?viewJob...*2_*2_*2_*2_*2

They are really starting now.
BTW, Xenon project was started in early 2003 (right?).. and Xbox360 was launched in late 2005. Maybe there are chances for a late 2013 launch.

[fehu]

Originally Posted by Rangers

Hmm, are people more excited when cell is used for "general purpose", or graphics as in Dice presentations?

There's your answer

i don't understand what you are impling
even the cpu is used for graphic so nextgen all the gpu will be used for graphic? so a little beefier cpu and not too much elaborated gpu?

[Squilliam]

Originally Posted by liolio

I wonder about it MS just swallowed the RD cost to implement xenos on SOI 45nm process. Move everything to bulk @ a process that won't give them much in regard to the number of dies per wafers seems like superflous expanse at this point. Moving Xenon+Xenos on GF 32nm process late in 2012 as well as edram to 40/45nm late in 2012 seems like a more likely scenario to me.

Why would they port it from bulk to SOI when they can use a bulk process from TSMC? Although im not sure if its worth it for a marginal shrink and combining the ED-RAM onto the one die.

Originally Posted by MarkoIt

More:
http://www.linkedin.com/jobs?viewJob...*2_*2_*2_*2_*2

They are really starting now.
BTW, Xenon project was started in early 2003 (right?).. and Xbox360 was launched in late 2005. Maybe there are chances for a late 2013 launch.

I would say they are ramping up now. They probably couldn't employ such people until such time as they'd worked out many of the base architectural details. I guess this means they're at the 1/3rd mark in terms of being ready to release a next generation console. This means they could possibly rush it out for 2012 though the release date ought to be 2013 if everything goes to plan.

[liolio]

Originally Posted by Squilliam

Why would they port it from bulk to SOI when they can use a bulk process from TSMC? Although im not sure if its worth it for a marginal shrink and combining the ED-RAM onto the one die.

Vahalla is made on SOI 45 nm process if my memory serves right that's why I stated that they would have to reimplement everything on Bulk.

[eastmen]

late 2013 might give them acess to 18nm for the parts . Could see a monster lead in performance compared to last gens leap.

[Shifty Geezer]

At what cost? Launching a $400 monster power console at 18nm will enable what scale reductions and cost savings for hitting cheaper pricepoints? If there's little room to shrink, the launching price will have to remain closer to final price, which suggests something more Wii like - a smaller box launching with a more conservative spec.

[liolio]

I agree on top of it I would wait to see when TSMC 28nm process enter in mass production before expecting them to delivers 18nm in mass production from two years from now.

[eastmen]

Originally Posted by Shifty Geezer

At what cost? Launching a $400 monster power console at 18nm will enable what scale reductions and cost savings for hitting cheaper pricepoints? If there's little room to shrink, the launching price will have to remain closer to final price, which suggests something more Wii like - a smaller box launching with a more conservative spec.

Hasn't the sell through price remained close to the launch price for the 360 ?

The asp for MS was in the low $300 range and the 360 launched at 300/400 in dec of 2005. It seems to have worked quite well for them.

Sure they were able to add kinect but next generation if they move to a DD only platform hardrive space may become a huge selling point later in its life span. Launch systems may hit with 1TB hardrives while towards the end of the generation we could see all the way up to 10TB being an option this could keep the asp high , new revisions and shrinks could keep the asp high.

with 18nm you should have a good 2-3 process drops in its life time along with fabs upgrading to larget wafers and some new mask tech coming down and some other innovations that will be implemented in the coming years.

I don't think we will hit 18nm and suddenly stop shrinkin chips.

[AlStrong]

The smaller process node is only cheaper when production ramps up and there's a high capacity for it... that's not going to be the case when it's first introduced.

Considering the delays with 28nm, I don't see how we're going to hit 18nm that quickly or inexpensively.