🚀 go-pugleaf

In article <1992Aug07.203524.7205@sco.COM> md@sco.COM (Michael Davidson) writes:
>Can you conceive of any reason why Intel would *bother* to produce a P5
>processor if it *didn't* run 486 code (which is all that will be available
>for it to run intially) faster than a 486 running at the same clock speed?

Yes, I can.  The 386SX is a little slower than the 286 when executing
some code at the same clock speed.  It's certainly no faster.  Now, I
wonder why the 386SX is selling *a little bit* faster than the 286?

>Of course it is perfectly possible that recompilation would be necessary
>in order to take full advantage of the new processor - as Guy Harris pointed
>out a superscalar version of the 486 would show a performance gain at the
>same clock speed but would require recompilation to get the optimal instruction
>sequences.

you have given the reason yourself.  I expect Intel would be quite
happy to take a *small* performance hit on existing code if this meant
that they could continue with speed/other improvements with a new
instruction set/mode.

Gordon.