Seems like Dirk Meyer had AMD on the right track given the financial performance and their new products. I have seen some new server teardowns that use its CPU-GPU combo that seemed to yield screaming performance (with hardware-level hypervisor) in a multicore environment. Of course, the price premiums over the value they deliver and the performance will have to be looked at. Also, @Mapou is right in saying that the memory bandwidth bottleneck will be the deciding factor.
Dr. MP Divakar
It does not matter who replaces Meyer. AMD made a big mistake by focusing on competing with Intel when it should have been working on a competitive tablet processor. Both Intel's and AMD's hybrid strategy of combining last century's boomer-geek CPU and GPU technologies is wrongheaded. Larrabee was the first big failure of that strategy and AMD's Fusion will be the next.
One of the big problems with boomer geek technology is their infatuation with the Turing machine as God's gift to humanity. The Turing computing model is turning out to be a complete catastrophe when it comes to parallel programming. Not to mention that it has been a disaster with regard to software reliability and productivity from the beginning.
However, the worst problem with boomer geek technology is the memory bandwidth bottleneck. This will kill any and all progress in the age of massive parallelism. This is the nasty problem that will kill Intel and everyone else and it is the problem Dirk Meyer should have been focusing on.
Google "How to Solve the Parallel Programming Crisis" for insights into possible solutions.
It should be succeeded by an executive from Qualcomm, nVidia o Broadcom, they are leading performers in what AMD is trying to become.
Another theory; a conflict of interest about wafer pricing and volumes being outsourced to Globalfoundries?
TSMC is offering much better prices per wafer, then why continue to subsidize Globalfoundries now that AMD owns less than 14% of Glofo.
A rational board acting on the best interest of AMD shareholders would demand large scale fabrication outsourcing to TSMC.
AMD has had some sucess in the microprocessor market against Intel, but they aren't any where close to equals. They have been diversifying and moving to fab-lite(less?). I don't see them as a threat to Intel's market share anymore. I guess the CEO is always to blame.
Blog Make a Frequency Plan Tom Burke 17 comments When designing a printed circuit board, you should develop a frequency plan, something that can be easily overlooked. A frequency plan should be one of your first steps ...