POWER (Performance Optimization With Enhanced Risc) is an extremely nice architecture compare to the x86. It's firmware system was so ahead of the BIOS disaster (EFi might change this a bit, yet). I remember loading and booting kernels from NFS because you could load the network card drivers[1]. This is 2006, EFI was not known/used widely. The CPU itself has amazing features too.[2]
I guess we are seeing another occurrence of Gresham's law applied to the technical field. "Bad architecture drives out good". In this case bad is cheaper, and this is the only thing we care about after all.
You cannot praise POWER for OpenFirmware. OpenFirmware came out of SUN, and allows for the development of CPU-independent drivers. IIRC, SUN needed that because they shipped workstation variants that either had a 68k CPU or a SPARC CPU.
I seem to remember an article that pointed out some of the POWER addressing modes made it hard to optimize so it always would run slower than other chips. I would really love to hear what the story actually is.
I guess we are seeing another occurrence of Gresham's law applied to the technical field. "Bad architecture drives out good". In this case bad is cheaper, and this is the only thing we care about after all.
1. Open Firmware allows the system to load platform-independent drivers directly from the PCI card, improving compatibility. http://en.wikipedia.org/wiki/Open_Firmware
2. http://www.spscicomp.org/ScicomP16/presentations/Power7_Perf...