I heard about the new Cell microprocessor announcement from IBM, Sony, and Toshiba.
Wow. I don’t think they should call that chip a microprocessor. It should be called a micro-multiprocessor (MMP). Amazing work.
Now, the trick to these announcements is the execution. They better ship that thing this summer or they will have some serious competition on their tail in no time flat. If the past is any indicator, this could be very difficult for these guys.
What do I mean about execution? I mean having quantities ready on time. Being clear about target platforms. Don’t talk about a business PC until they have the OS and toolchain there. If they do come out with a chip for enthusiasts, have an inexpensive motherboard that they can purchase. After shipping the motherboard, continue to support it and improve it until the platform supports itself.
So, what happens if they do pull it off? They get to lead the next fundamental change in computing. These chips are designed for heavy signal processing and computing. This is not a spreadsheet, database, or web-server chip. This is built for hight bit-rate media processing (encoding, encryption, decompression, transformation). I look forward to the day that I get to experience what one of these powerhouse chips can deliver. I want to see them render the vast majority of systems as obsolete.
Things are getting interesting again, but only if they really ship.