Nvidia Making x86 CPU With Ex-Transmeta Brains?
Remember the Crusoe? Perhaps now it'll live on at Nvidia.
We've been hearing rumors of Nvidia's interest in entering the x86 CPU since last year, and now the rumblings are back.
With Nvidia not having the license to produce chipsets for the latest generation of Intel chips, the graphics company may have a further reason to make its own. Intel is also moving towards integrating graphics cores onto its CPUs, providing another threat to Nvidia's graphics business. Putting both of those things together, analysts believe that could be enough for Nvidia to get into the x86 game.
''We believe Nvidia could enter the x86 CPU business,'' said analyst Doug Freedman of Broadpoint AmTech, in an EETimes story. ''Nvidia could become a supplier of x86 CPUs by necessity to preserve both GPU and chipset revenue.''
For a while, there was speculation that Nvidia would acquire or invest heavily into VIA Technologies for the company's CPU properties, but now Broadpoint AmTech believes it'll be an in-house job.
''We believe internally developed x86 solutions are more likely than external acquisitions (i.e. Via Technologies),'' Freedman said.
What could make things even more interesting is the analyst's belief that Nvidia has picked up talent from Transmeta: ''We believe that Nvidia has hired former Transmeta staff extensively, and that instruction code "morphing" requirements have declined as more x86 instructions have come off of patent coverage,'' he said.

Yeah so its all good. I actually hope they do so but they better not start creating incompatibility issues, if they do so, with later on drivers with other companes "non intentionally" if you get my drift ...
http://www.tomshardware.com/forum/265251-28-emulate-cores
I posted a topic about that idea and everyone seemed to dismiss the idea. I still think it would be good on certain applications depending if the whole program could be decoded and executed. Much faster than only looking at 100-200 instructions of a 10,000 instruction program. Many more things could/should be programmed into a parallel mode of computation.
gpu's would not emulate but assist with on die x86 silicon , is what's this is all about . amd has both , intel has both and nvidia needs its own x86 or similar capability in silicon . it can neither ask amd nor intel for this .
you cannot parallelize or emulate steps when the next one depends on the previous , like recursive calculations . and an softwate emulated pipeline if that's what you mean , would not run at practical speeds . adding to the above post , why emulate when you can co-operate with native on die x86 ?
I wouldn't see much competition brought, because they are only limiting their options to 32 bit, the only way I can see this being affective is if they exceed or match the performance of Amd/Intel and be really really cheap, because why buy an x86 cpu when I can buy one that does x86 and x64.
I would not be surprised if we see Intel moving in a gaming GPU market then, and we can see all Intel/AMD/Nvidia based machines comping out.