ImgTec Launches Programmable Shader Graphics Family

Thruth be told Img tech just hold too many tbdr patents for anyone to go toe to toe with them in the performance department.

I think they've wrapped up the mobile realm for good.

 

If the TBDR advantages in small embedded/UMA architectures can't be surpassed, then theoretically at least other IHVs could easily consider using TBDRs of their own. As long as it wouldn't infringe IMG's patents, I don't see a problem there.

Did you ever hear that a startup company wouldn't theoretically be able to build an IMR, because either ATI or NVIDIA or others hold "too many patents"?

PowerVR has their own exclusive patents for it's own architecture, yet that doesn't exclude for anyone else (yes at least in theory) to build a deferred renderer if they'd want to.

The real reason why major IHVs would be reluctant to jump to DR, would be rather that they have an already large and existing technology portofolio and there it's much easier to try to use that one for ultra small dies, then develop a new fundamentally different core from scratch.

 

I think an img pr guy said himself that because they have patented the technique worldwide, it's very very difficult for competitors to do without infringing patents.

The fact that we haven't seen defered rendering by anyone else suggests that they are right. It is indeed very difficult to do without being taken down to the courts.

Alot of companies in the past have claimed to be able to have done it- there's a company in the US called PixelSquirt that claims it does it, there was gigapixel and there are others too.

Yet however nobody has ever brought one out to market and the simple reason is that if they did, IMGtec would have a field day with lawsuits.

These groups may hold their own patents but all the workable and worthwhile ones are held by Img.

They pioneered it, they invented it, it is therefore rightfully their's and only their's.

I hope SEGA SAMMY buys them out.

 

Did you say PR?

Those companies' failure to release a design of hardware is according to you the possibility of patent infringement?

a. W/o released hardware and a chance to have a closer look at it, you can't even know - and yes even IMG - if any patents have been really infringed.

b. Gigapixel lost the XBox Microsoft deal because they were unable to deliver on time.

What would you call the Falanx Mali by the way?

I don't want to even hear it. Not only don't they have the resources for a stunt like that, but I'd rather prefer companies like that to lay off their fingers from anything considering hardware design.

 

Regarding the falanx mali I'm not too sure what it is, as TBR? a TBDR?

Also Img can't do anything untill a product utilizing the chip hit's shelves.

 

Tile based + early-Z.

Which would mean what exactly for your former assumptions? (yes assumptions). I wonder why hasn't anyone exlusively patented the Z-buffer yet?

There's a difference between patenting one's own implementation and claiming to have patented exclusively the whole approach. Granted there might be pitfalls for any non-experienced designers with TBDR, the only other question is if they'll be able to circumpass them without infringing any patents in the end. With as many talented and experienced engineers out there I don't have much reason to believe that it's entirely impossible though.

 

One design is supposedly suspected of possibly infringing PowerVR IP, but most companies who don't get their processor designs brought to market are impeded by a lack of resources, either of themselves or of their product partners.

 

By the way...
Assembly 2005 Mobile demo compo winner was made for Dell Axim x50v.

Too bad that I missed the competition, but the annual meeting with Nicklas and other Beyond3D'ers was at the same time.

 

I wish that the next Game Boy would use a PowerVR SGX core.

but assuming that ATI gets the nextgen Game Boy graphics contract (or already has it) I hope ATI can produce something equally as good as SGX. I have not been keeping track of ATI's advancements in mobile graphics for handhelds.

does ATI have anything that can compete with SGX now or on the horizon for the near future?

 

Besides the current Imageon (which is more a multimedia than 3D oriented chip IMO) nothing else is known about the future. I would dare to speculate though, that ATI might also go unified with their second attempt.

 

Which is frankly a phenomenon we see way too often with any IP company. Even IMG/PowerVR hasn't been immun to those.

That was the GP-LP that Gigapixel had planned for the PDA/mobile market:

http://users.otenet.gr/~ailuros/GPLP.pdf

 

They will be available, but the time is different question. In demo scene delays used to be more rule than exception. Nowadays things have been going rather well.

based on paper specs, no one has anything that can compete with SGX right now. G40 is the closest thing, but it's slower and less flexible. (though both have full OpenGL ES 2.0 support, SGX goes beyond that.)

 

We have now also Falanx's Mali200 paperspecs; that's three in total so far for the second generation of PDA/mobile specs. I guess it's the turn of the big boys now

 

Yes.

Cheers,

AJ

 

... and ATI also has a patent on unified shader architecture, which would mean that ImgTec licenses that from ATI or ...

 

A CPU is a unified shader architecture.

Perhaps you meant to say "ATI also has a patent on a unified shader architecture"

 

No

 

If the Series 5 GPU that SGX is based on also has a unified architecture, then Img applied for it's patents before Ati because the Series 5 GPU was ready to go last summer.

 

Actually, technically they appear to have a patent application in that area.