Sony's latest legal challenger - aiming to destroy PlayStation!

[nelg]

http://www.freepatentsonline.com/5056000.html

A high speed computer that permits the partitioning of a single computer program into smaller concurrent processes running in different parallel processors. The program execution time is divided into synchronous phases, each of which may require a shared memory to be configured in a distinct way. At the end of each execution phase, the processors are resynchronized such that the composite system will be in a known state at a known point in time. The computer makes efficient use of hardware such that n processors can solve a problem almost n times as fast as a single processor

They might have a reasonable case (in a legal sense).

The real meat of the case most likely lies in the first claim.......

1. An apparatus for parallel data processing over a plurality of phases, comprising:

a plurality of processors, any one of said processors operative as a master processor, said master processor including means for generating interconnection switch configuration control signals;

a communication bus for interconnecting each of said processors, for exchange of at least control and synchronization information among each of said processors;

a plurality of multi-access memory modules;

an interconnection switch coupled to each of said processors and each of said multi-access memory modules, and responsive to said interconnection switch control signals from said master processor, for selectively interconnecting any one or more of said processors with one or more of said multi-access memory modules, and whereby any one of said multi-access memory modules is exclusively interconnected to only one of said processors during any given phase of processing;

each of said processors further including local memory, whereby one or more of said processors processes data in its local memory before, after, and during a phase of processing;

said master processor further including means for generating a processing phase commencement signal over said communication bus to the other of said processors, said commencement signal indicating the start of each of said plurality of phases of parallel data processing during which any one or more of said plurality of processors is exclusively interconnected to said one or more multi-access memory modules; and

each of said processors further including means for generating a completion signal over said communication bus to the other of said processors said completion signal indicating completion of each phase of processing.

 

[Gradthrawn]

Only just found it..

Not really had time to look at it properly..

Maybe some of you guys can evaluate it fully..

Great find. A brief cursory overview would seem to indicate that what they're describing is pretty similar to the Cell. However, some of it is so general in terms of parallel processing that it could apply to many multi-core CPUs.

From the patent...

Each processor has its own local memory and local input/output capabilities, and therefore can act as an independent computer. Each of the processors are connected to a global bus and are connected through an interconnection switch to a shared multiaccess memory (MAM) with multiple memory modules.

This would seem to give their claim some merit.

A MIMD computer is disclosed herein which has a plurality of physically identical processors, one of which can function as a master processor and the others of which can function as slave processors.

This, however, would seem a bit damning. Unless, somehow, the PPU and SPUs are "physically identical."

 

[Carl B]

Well obviously on a high level, the Cell infringes on this. Now - the question then becomes, will the courts side with Parallel Processing Corporation, or will they determine that the patent is bogus and overly broad?

They picked the right court, that's for sure - but if this patent is left to stand, its going to get under a lot of peoples skin. My feeling is that Toshiba and IBM are definitely going to come in with "friend of the court" testimony should this thing actually progress anywhere. We may even see Intel AMD and Sun get involved also, since they're each moving into Cell territory shortly.

I bet that's what this is about also - Parallel Processing Corporation wants a ruling in its favor by attacking what they perceive to be the weakest link in the chain - Sony: a non-traditional complex IC manufacturer that may be unused to defending itself against this type of litigation - and via securing a ruling in their favor, will look to lord over the entire massively-multicore era by using the case as precedent in future lawsuits.

 

[Shifty Geezer]

The principle claim seems to cover Cell as abroad description, but the meat to me seems quite different. Their design seems to be for a processor which takes a job and spreads it across multiple CPUs, whereas Cell is multiple independent devices that, if you really want, you could try to get the PPU to take a program and spread it between them. Indeed the first claim says...

said master processor further including means for generating a processing phase commencement signal over said communication bus to the other of said processors, said commencement signal indicating the start of each of said plurality of phases of parallel data processing during which any one or more of said plurality of processors is exclusively interconnected to said one or more multi-access memory modules

But the PPU isn't there to send a start signal to get the SPU's working, and they act independently.

 

[archangelmorph]

Hang on a second..?

Didn't Sony and IBM patent the Cell??

If so how the hell did it ever get through the patent process with this patent so clearly similar and represented as related prior art?

Surely then this entire lawsuit was settled when the ignorance/carelessness/complacency/whatever it was of the patent office granted Sony the patent for the Cell processor in the first place..?

Either the previous patent was deemed too dissimilar, it wasn't referenced at all (in such a case Sony should be able to counter-sue the patent office.. In an ideal world anyway..) or it was deemed void and of no relevance to the Cell patent..

Either way we all know who the REAL criminals are in this..

(Patent Office.. Grrr...)

 

[archangelmorph]

The principle claim seems to cover Cell as abroad description, but the meat to me seems quite different. Their design seems to be for a processor which takes a job and spreads it across multiple CPUs, whereas Cell is multiple independent devices that, if you really want, you could try to get the PPU to take a program and spread it between them. Indeed the first claim says...

said master processor further including means for generating a processing phase commencement signal over said communication bus to the other of said processors, said commencement signal indicating the start of each of said plurality of phases of parallel data processing during which any one or more of said plurality of processors is exclusively interconnected to said one or more multi-access memory modules

But the PPU isn't there to send a start signal to get the SPU's working, and they act independently.

I thought about this aswell..

Since the Cell is completely programmable and thus isn't restricted to complete synchronised processing across the entire chip, it doesn't make sense for this patent to hold any merit..

 

[Carl B]

Well that's a good point, Sony *does* have a patent on this... and its not the only one between the three companies related to the architecture either. But the whole patent system is a mess, and if the patent's from 1991, it's not hard for these guys to say it was overlooked when determining whether the Cell BE concepts were patentable.

US patent system = broken

 

[Shifty Geezer]

Didn't Sony and IBM patent the Cell??

If so how the hell did it ever get through the patent process with this patent so clearly similar and represented as related prior art?

This is something I was thinking, and perhaps here's a good time for those more clued into international intellectual property to comment

1) If the Patent Office grant a patent when they shouldn't have, and Sony lose the case, can they sue the Patent Office and cast all costs onto them? After all, if the PO hadn't granted the patent and pointed out the prior patent, Sony would have done things differently.

2) If Texas in its infinite wisdom sides with the plaintiff, how does that affect rulings elsewhere in the world? Does a violation in one country constitute a violation in the other countries that are members of the IPT or whatever it is? Do other US States have to follow Texas's lead where other countries don't? Presumably the wholesale destruction of PS3 (and what other devices? They mention PS3 is only one example) and murder of PS3 owners for being part of this infringement and the destruction of their houses would be limited to certain territories. Thus US PS3 owners could flee the country?

 

[Tap In]

http://arstechnica.com/news.ars/pos...lel-processing-patent-threatens-sony-ps3.html

...What often happens in such cases is that the patent stays on file until it's rediscovered and evaluated as a potential source of licensing revenue. Whenever small companies have such potentially valuable patents on file, they'll often either sell them or license them to an enterprising group of patent lawyers who can litigate them. From what we can surmise, this appears to be what happened in this case. The originating company, IPMI, may not have had the resources to litigate the patent itself, so it is possible that the company licensed it to a group of lawyers who set up a corporation specifically for the purpose of exploiting this patent....

as the patent could be construed to cover certain critical aspects of modern parallel computing—to the point that it affects every company that dips its toes in the parallel-processing waters, including IBM, Microsoft, Intel, and AMD. Speaking of IBM, it's a bit strange that Parallel didn't go after Big Blue first.

 

[dobwal]

This is something I was thinking, and perhaps here's a good time for those more clued into international intellectual property to comment

1) If the Patent Office grant a patent when they shouldn't have, and Sony lose the case, can they sue the Patent Office and cast all costs onto them? After all, if the PO hadn't granted the patent and pointed out the prior patent, Sony would have done things differently.

I don't think Sony can sue the US government under these circumstances unless the US government waived sovereign immunity or Sony could sue under the Federal Tort Claims Act (FTCA).

2) If Texas in its infinite wisdom sides with the plaintiff, how does that affect rulings elsewhere in the world? Does a violation in one country constitute a violation in the other countries that are members of the IPT or whatever it is? Do other US States have to follow Texas's lead where other countries don't? Presumably the wholesale destruction of PS3 (and what other devices? They mention PS3 is only one example) and murder of PS3 owners for being part of this infringement and the destruction of their houses would be limited to certain territories. Thus US PS3 owners could flee the country?

I don't think US patents are inherently enforceable throughout the world.

 

[Shifty Geezer]

I don't think Sony can sue the US government under these circumstances unless the US government waived sovereign immunity or Sony could sue under the Federal Tort Claims Act (FTCA).

That's what I thought, but the flip-side is, basically the US government could screw up, cost Sony billions, and dust their hands of it. Japan aren't going to take kindly to that! It'd be a huge ruckus if this case is upheld.

 

[Carl B]

http://arstechnica.com/news.ars/pos...lel-processing-patent-threatens-sony-ps3.html

Well, like I said I think Sony simply presents the weakest target as far as the Texas court system is concerned. Also I wish Ars would give Sony and Toshiba some of the design credit they deserve! IBM reaps too much credit from Western media.

 

[Tap In]

That's what I thought, but the flip-side is, basically the US government could screw up, cost Sony billions, and dust their hands of it. Japan aren't going to take kindly to that! It'd be a huge ruckus if this case is upheld.

I'm not an IP lawyer but I'm pretty sure that the Patent Office has no responsibility to ensure that applications do not infringe on an older patent.

There is no screwing up as far as the patent office goes. It IS left up to the courts to determine if wording/ideas from one patent infringes on another if the original holder files a complaint.

so IOW this is SOP.

 

[nelg]

I'm not an IP lawyer but I'm pretty sure that the Patent Office has no responsibility to ensure that applications do not infringe on an older patent.

Yep. The due diligence falls on the applicant. It is their job, not the patent office's, to do the searching. How they missed this in a search for prior art is beyond me. The patent office will look at the prior art listed in the application and judge if it is new and novel enough to issue a patent.

 

[Carl B]

The patent office *does* search for prior art when considering the merits of an application, and does not simply go with whatever is presented by the applicant. Rather, the applicant simply has a legal obligation to disclose all prior art they may feel pertinent. But the PTO definitely conducts their own search. Take this from someone that has been embroiled in a back-and-forth with the PTO for several years now.

 

[wowfactor]

But the patent office does not guarantee the party that is applying for a patent free from challenges. The party that is applying for a patent may lose their patent if they fail the challenge.

 

[dobwal]

That's what I thought, but the flip-side is, basically the US government could screw up, cost Sony billions, and dust their hands of it. Japan aren't going to take kindly to that! It'd be a huge ruckus if this case is upheld.

At worse Sony would be required to pay license fees for any future Cells sold and back payments on any already on the market.

If the Sony's patent are found invalid it makes no sense for the validated patent holder to stop the production of the Cell as its counterintuitive to the purpose of patents (to profit from them).

 

[3dilettante]

The principle claim seems to cover Cell as abroad description, but the meat to me seems quite different. Their design seems to be for a processor which takes a job and spreads it across multiple CPUs, whereas Cell is multiple independent devices that, if you really want, you could try to get the PPU to take a program and spread it between them. Indeed the first claim says...

said master processor further including means for generating a processing phase commencement signal over said communication bus to the other of said processors, said commencement signal indicating the start of each of said plurality of phases of parallel data processing during which any one or more of said plurality of processors is exclusively interconnected to said one or more multi-access memory modules

But the PPU isn't there to send a start signal to get the SPU's working, and they act independently.

The PPU is capable of doing so, it has the capability of producing what could be construed as commencement signals or messages for SPU execution.
At the very least, the PPU's initializing the SPUs might count enough for that.

The patent at the very least covers the ability of Cell to initialize itself.


From a brief glance, one possible issue is the reference to multi-access memory modules.
The patent claim includes the idea that each processor can have only be linked to one given module during any phase of computation.

If the multi-access memory for Cell is XDR RAM, then it may be possible that the reference to fixed access per module is not infringed, since accesses are by address, and do not match internal divisions by rank, row, or physical chips.

Whether this is enough to invalidate all or part of the claim is something a laywer would have an idea about.

 

[Carl B]

But the patent office does not guarantee the party that is applying for a patent free from challenges. The party that is applying for a patent may lose their patent if they fail the challenge.

Of course.

The PTO and patent issuance is only an abstract pre-positioning for what really matters: the courtroom, and indirectly, the licensing fees and IP controls that are meant to avoid it.

It'll be an interesting case - just from a computing angle - to see develop. Hopefully they get trounced, because this patent really is a zombie come to terrorize all of modern computing should the judge side with them. But that said, it also goes as an almost automatic that the case will be appealed should things get funky.

 

[bystander]

Perhaps this recent ruling could restore some faith in the legal system? (in relation to patents)

Essentially it's a case regarding eBay's Buy it Now button, with the plaintiff demanding an injunction and damages on the Buy it now button. The plaintiff won damages but the court turned down the request for the injunction.

Judge Friedman once again declined to issue an injunction, arguing that MercExchange is what's often referred to in patent law as a "troll" - a company whose main objective is to use its patents in suits against other business, not as a means of getting ahead in the marketplace . "MercExchange's modus operandi appears to be to seek out companies that are already market participants that are infringing, or potentially infringing, on MercExchange's patents and negotiate to maximize the value of a license, entered into as a settlement to, or avoidance of, litigation," he wrote. The company, he argued, can be compensated solely with an award of monetary damages.

"If there is one patent holder out there you didn't know about and you're infringing on their exclusive rights, they can use an automatic injunction to essentially extract the entire value of your product - even if their patent represents only a small part of the product." said Sarah King, an intellectual property lawyer with the San Francisco Howard Rice Nemerovski Canady Falk & Rabkin PC.