RE: MISC not for embedded?
But what about the argument the Scenix company is using? They are
now marketing the speed and excellent debugging features built
in as a way to develop more software peripherals. They are approaching ten
times the speed of an ordinary 68hc11 with lots of parallel functioning
peripherals, yet they have an easy $7 price. They don't say to do speech
processing with it, (yet), they say throw that extra speed into
code that emulates serial/parallel ports, and sigma delta ADCs and DACs.
If the development model of
creating code to do peripheral functions works on a scale of uarts and SPI
ports already, then imagine multiple processors with a rapid develop/debug
cycle to do the tough jobs which need many IO's. MISC does fit that model
and goes farther since the bigger available word width makes ease of
handling
math operations on precise converted data streams from sensors and to
motors.
For a really fast develop/debug cycle, a valuable function to tightly
integrate with a MISC is non-volatile memory like Ferroelectric RAM (Only,
right now, FRAM is not available as far as I know). Another value in
embedded controlling is cheap packages (read low pin counts). A MISC
version with some internal memory, plus a serial I/O to a fast serial
memory, where the serial IO speed is faster than the processor speed by the
right amount to match the memory read/writes to the ALU would be a boon to
me. How fast can you get serial memory now?
> -----Original Message-----
> > What embeddable processors need is:
> > - Many on-chip peripherals
> > - On board (small) RAM en (big) [[[E]E]P]ROM.
> > - Low cost (so high production volume)
> > - Good support from the manufacturer
> >
> > Speed is often not essential.
