SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?


To: chipguy who wrote (242162)10/9/2007 3:08:55 PM
From: graphicsguruRead Replies (2) | Respond to of 275872
 
Chipguy, your example of a slight twist in a CMOS inverter sounds a whole
lot more defensible than the SGI patent. Evidently, it had some value, and
was different from what people did previously -- so it wasn't completely
obvious.

The SGI patent is rather different. I know ATI regularly
goes and speaks to customers and asks them what they want. And I
would be very surprised if ATI can't find some documents from customers
asking for exactly the features in the SGI patent (pre application). Most
likely, they also have design documents that also pre-date the patent, where
they've evaluated the cost-benefit of going to fp, and decided that it's
still n years away. They certainly have long-term roadmap documents
that show fp in their future plans -- some of these likely pre-date the
application. There's also an excellent chance that a thorough search will
turn up an obscure piece of research graphics hardware using fp.

I'm well aware that "massive originality" is not a requirement for patents,
but "non-obviousness" is. And the SGI patent fails in this regard.
Whether or not a court will agree is a bit of a crapshoot, but I'd bet on
ATI prevailing.

So I have three reasons for believing this is not terribly significant:
my own judgement about non-obviousness, my own judgement about
prior art, and the fact that ATI did an evaluation, and decided to litigate.

By the way, I hold over ten patents with more pending, so I do know a
thing or two about them.



To: chipguy who wrote (242162)10/9/2007 5:04:41 PM
From: MagratheaRespond to of 275872
 
IMO AMD's best bet is to
show prior art involving hardware realization of the invention.


In what ways does the x387 floating-point coprocessor fail the test of prior art here?

-Magrathea