Self-modifying code

[T. William Wells]

In article <60175 at sun.uucp>, petolino%joe at Sun.COM (Joe Petolino) writes:
> >>    Why are an Icache plus a Dcache better than just
> >>    a big shared cache as big as both?
> >
> >In terms of hit/miss ratios, a unified cache is clearly better.
>
> I beg to differ...
>                                  ...Take this with a grain of NaCl: we only
> tried a few test programs, and each was small enough to fit into the cache.

Another relevant point about separate caches is this:
instructions and data have different access patterns.  A cache
designed for the one is not necessarily going to be right for the
other; a cache designed for both may not do as well as the
separate ones.  So this possibility has to be balanced against
the general benefit of having a larger pool to work with.