Linards Ticmanis wrote:
Michael J. Mahon wrote:Any "smart" drive will wind up reading the disk data into an internal buffer and then transferring the data to the host computer when requested. This double move performed by a single processor guarantees that a smart drive cannot move data at the rate the disk can support.Just to go entirely off-topic, this isn't quite right as the Commodore PET/CBM drives had two 6502s in interleaved operation, one for talking to the host computer and one for talking to the actual drive hardware. Those drives were reasonably fast though expensive.
I'm impressed! I would have expected a little more cost-cutting than that. ;-)
But then when they ported their DOS code to a single-CPU design for the cheaper home computer line they kept this logic around (one CPU now alternately impersonating the two CPUs of the old design), which made the DOS really weird (the two "personalities" didn't simply call each other as needed but rather communicated via command lists in RAM) and contributed to its slowness.
A curious design history--thanks for the info! -michael NadaPong: Network game demo for Apple II computers! Home page: http://members.aol.com/MJMahon/ "The wastebasket is our most important design tool--and it's seriously underused."