🚀 go-pugleaf

prener@watson.ibm.com (Dan Prener) writes:
vhs@rhein-main.de (Volker Herminghaus-Shirai) writes:
>> Subject line almost says it all. Many modern processors have an
>> instruction to swap the bytes in a word. Can one quantify the
>> win of using this instruction vs. the equivalent sequence of
>> instructions? How much silicon is needed? What applications
>> win the most and how much? There must be *some* resaon why the
>> RISC folks put it in, right?

>The win is approximately zero for general code.  But it is significant
>for emulating a processor with the opposite byte order.

This makes the whole issue of NextStep 486 strange.  If the purported
reason that NextStep is only supported for the 486 is to sell more 486's
then why don't they sell a 386 version with all the bswap's replaced by
the long and slow series of 386 instructions which it takes to emulate
it?  That way, they could show that the 486 was much superior to the 386
by demonstrating how much faster NextStep runs on the 486 as compared to
the 386?  That would be a more subtle way of selling 486's.  It might
even carry over to people who want to run things other than NextStep
because, "Well, if it runs that faster, it will run my program faster
too, right?(wrong)" :)

Looks like another bad marketing plan.

--David Willmore
willmore@iastate.edu