Re: Re(fcc): TI-H: Radio/Infrared/Laser Communications


[Prev][Next][Index][Thread]

Re: Re(fcc): TI-H: Radio/Infrared/Laser Communications




>NT sucks.  MacOS sucks.  BeOS sucks.  AmigaOS sucks.
>Unix sucks.  Less.
>Perhaps Jon has to do some work on an NT-based program that won't run
>under Wine, and he's not willing to deal with FreeBSD's latest "can't
>write floppies" problem, or the sloth that it Solaris.  Granted, the
>floppy problem has a fix, but I had to vent.
>Linux can do all the things that NT can, only better.  That's why one
>would choose it over NT.
>Finally, CISC has to interpret and breakdown variable-size microcode in
>order to run, while RISC avoids that step, using fewer instructions of
>the same length.  As a result, RISC lends itself better to
>multi-pipeline CPU designs.  Finally, you cannot compare RISC and CISC
>based on clock speed, because the programs are compiled based on the CPU
>itself.  The PIC may use a limited instruction set, but unless it uses
>no microcode, it's not RISC.  To prove a point, Intel x86 designs have
>been moving toward RISC since the 486, further blurring the distinction
>between the two camps.
>Oh, and you can't compare bogomips across platforms.  I saw a Libretto
>100CT get a higher bogomips rating than an Alpha 300+ MHz machine.  It's
>all in the loop optimization.

So what.  Makes good numbers.  :)

RISC CPUs based on instruction cycle are faster...


References: