"Xbox 2 Patent?"

[Brimstone]

It's quite disappointing to notice a lot of patents I posted here go completely unnoticed. And then they bounce back here from another (loosy) forum

Turn that frown upside dowon Nao. Your posts are some of the most highly regarded on the forum. It is lame that the guy from Team Xbox didn't cite you for the patent though, when it's clearly obvious you came across it first. Beyond3D has to be one of the most ripped off sites on the web for 3d graphics and technology.

 

[nAo]

Thank you Brimstone..I'm touched

 

[Shifty Geezer]

IS there any point to these patents? They all seem to talk about the same thing - having multiple CPU do processing work. What exactly is the new and unique idea they want to secure as their own idea?

 

[AlNom]

IS there any point to these patents? They all seem to talk about the same thing - having multiple CPU do processing work. What exactly is the new and unique idea they want to secure as their own idea?

Just the method... so that noone else can do the same thing

 

[AzBat]

nAo,

Check my comments...

http://www.eurogamer.net/article_discussion.php?article_id=57519

Tommy McClain

 

[one]

Well well, ripping things are common in today's blog-filled web... more so when you can search public patent database.
Next time, before E3, let's find Revolution patents earlier than anyone else

 

[Shifty Geezer]

IS there any point to these patents? They all seem to talk about the same thing - having multiple CPU do processing work. What exactly is the new and unique idea they want to secure as their own idea?

Just the method... so that noone else can do the same thing

So what's new bout this MS approach? A glance at the diagram suggests a straightforward multiprocessor set-up, with perhaps specialisation of certain processors. The idea of one processor for general purpose and the others specialised for geometry work sounds a bit pants IMO. Why not just use the multiprocessors like unified shaders and divvy up the processing power as needed between physics, AI and graphics?

Of course optimised geometry processors would be quicker, but aren't we expecting Power derivatives with a few custom functions? Or is the Xenon processor still quite an uncertain beast that might have non-standard specialism? Is the 6 core design 1 CPU and 5 'support' processors?

 

[skilzygw]

" Particularly, Allard described a new technique called “procedural synthesisâ€￾; special programs that create realistic bricks, trees, and other environmental objects, leveraging game artists from spending long hours creating repetitive geometry. "

For people who actually know(cause i dont) doesnt this just sound like what Unreal Engine 3 can do in its software? I remember reading something about the engine being able to for example draw in foliage or apply the correct bump mapping etc..

 

[AzBat]

Well well, ripping things are common in today's blog-filled web... more so when you can search public patent database.
Next time, before E3, let's find Revolution patents earlier than anyone else

LOL! I understand, but I still don't think it's right for site like GamesIndustry.biz or TeamXbox to not do the proper research and give proper credit to the source. Oh well.

Tommy McClain

 

[nAo]

Thank you Azbat

 

[Inane_Dork]

So what's new bout this MS approach?

Besides the multiprocessor/multithreaded nature of it, nothing. It's been done already on the Gamecube.

Of course optimised geometry processors would be quicker, but aren't we expecting Power derivatives with a few custom functions? Or is the Xenon processor still quite an uncertain beast that might have non-standard specialism? Is the 6 core design 1 CPU and 5 'support' processors?

CELL, are we thinking?

I rather doubt MS wants processor specialization. I would guess they go for equal processors and dynamic balancing between them. Keep in mind that the GPU should also be capable of a fair amount of vertex processing. Between the CPU pool and the GPU, the Xenon should push *puts pinkie finger to lip* millions *removes pinkie* of polygons.

 

[Brimstone]

Well well, ripping things are common in today's blog-filled web... more so when you can search public patent database.
Next time, before E3, let's find Revolution patents earlier than anyone else

LOL! I understand, but I still don't think it's right for site like GamesIndustry.biz or TeamXbox to not do the proper research and give proper credit to the source. Oh well.

Tommy McClain

Even www.hardocp.com has a link to the Team Xbox story.

Posted by Steve 7:37 AM (CST)
Xbox 2 Patent?
Team Xbox have what they believe to be Xbox 2 patent information (along with a diagram) that seems to confirm the multi-core CPU information leaked last year. Very cool stuff if it does pan out to be next-gen Xbox hardware. Here is a quote from the article:

When someone posted the news in the Team Xbox forums they at least mentioned Beyond3D. Yet, when Team Xbox writes an article about it a few days later, they can't even be bothered to mention any type of source.


http://forum.teamxbox.com/showthread.php?t=318027

 

[loekf2]

it might be old now, but i hadn't seen the diagram

Anybody also noticed that the GPU features:

- memory controllers -> embedded DRAM and external DRAM ?
- the CPUs have to access via a shared L2 cache DRAM via the GPU. I hope the L2 cache and X-bar are smart enough to allocate the right bandwidth for these CPUs and GPU.

BTW... why is there a coherency module in the GPU ? Seems a bit odd to me.

 

[Guden Oden]

- memory controllers -> embedded DRAM and external DRAM ?

The diagram doesn't show any embedded memory at all, and multiple memory controllers have been a standard feature since the geforce 3 on any decent 3D accelerator.

the CPUs have to access via a shared L2 cache DRAM via the GPU.

Nah. Via a crossbar located ON the GPU, just like the current xbox. No difference there...

BTW... why is there a coherency module in the GPU ? Seems a bit odd to me.

Um, well as both the GPU and CPU are capable of changing data in memory, they'd both have to be aware of what the other are doing and update their caches appropriately too or else risk corrupting data for the other.

 

[AzBat]

Sorry to bring this back from the dead, but thought some of you might be interested in this...

U.S. Patent and Trademark Office officially issued Microsoft their Xbox 2 patent today...

"System and method for parallel execution of data generation tasks"

Patent Number: 6,862,027

Issue Date: March 1, 2005

Published Date: December 30, 2004

Filed Date: June 30, 2003

Abstract:
A CPU module includes a host element configured to perform a high-level host-related task, and one or more data-generating processing elements configured to perform a data-generating task associated with the high-level host-related task. Each data-generating processing element includes logic configured to receive input data, and logic configured to process the input data to produce output data. The amount of output data is greater than an amount of input data, and the ratio of the amount of input data to the amount of output data defines a decompression ratio. In one implementation, the high-level host-related task performed by the host element pertains to a high-level graphics processing task, and the data-generating task pertains to the generation of geometry data (such as triangle vertices) for use within the high-level graphics processing task. The CPU module can transfer the output data to a GPU module via at least one locked set of a cache memory. The GPU retrieves the output data from the locked set, and periodically forwards a tail pointer to a cacheable location within the data-generating elements that informs the data-generating elements of its progress in retrieving the output data.

Inventors:
Andrews; Jeffrey A. (Sunnyvale, CA); Baker; Nicholas R. (Cupertino, CA); Goossen; J. Andrew (Issaquah, WA); Abrash; Michael (Kirkland, WA)

http://patft.uspto.gov/netacgi/nph-....WKU.&OS=PN/6,862,027&RS=PN/6,862,027

I don't see anything new in the patent. There shouldn't have been anyway. Only interesting part is that it was issued pretty quickly.

Tommy McClain

 

[wco81]

Who are those inventors and why are there two of them in Silicon Valley?

ATI guys?

 

[PC-Engine]

Yeah, seeing the filed date June 30, 2003, they seemed to have studied a lot about the other CPU architecture by then.

Yeah funny how they didn't go the streaming processor route like CELL and instead went with a SMP model.

 

[one]

Yeah, seeing the filed date June 30, 2003, they seemed to have studied a lot about the other CPU architecture by then.

Yeah funny how they didn't go the streaming processor route like CELL and instead went with a SMP model.

I find it's more interesting that each element in the Cell processor, SPE and the Cell processor itself, has its own local memory near, and they form NUMA, rather than SMP.

 

[PC-Engine]

Yeah, seeing the filed date June 30, 2003, they seemed to have studied a lot about the other CPU architecture by then.

Yeah funny how they didn't go the streaming processor route like CELL and instead went with a SMP model.

I find it's more interesting that each element in the Cell processor, SPE and the Cell processor itself, has its own local memory near, and they form NUMA, rather than SMP.

Well interesting doesn't really pay the bills...

 

[Ty]

Yeah funny how they didn't go the streaming processor route like CELL and instead went with a SMP model.

I find it's more interesting that each element in the Cell processor, SPE and the Cell processor itself, has its own local memory near, and they form NUMA, rather than SMP.

Well interesting doesn't really pay the bills...

True, but neither does funny.