On 10-May-1998, S. Alexander Jacobson <[EMAIL PROTECTED]> wrote:
> On Fri, 8 May 1998, Fergus Henderson wrote:
> > Note that consolidating multiple passes into single passes is not
> > always a win. For example, if your machine has 10 available registers,
> > and each pass uses 8 of them, then combining the two passes may mean
> > that some variables can longer fit in registers, which may lead to
> > things being less efficient.
>
> Isn't it somewhat unreasonable to expect the typical Haskell
> programmer to know how many registers a particular target CPU has and for
> that matter how many registers a particular algorithm actually uses?
> Can't the compiler figure this one out?
Yes, ideally the compiler would figure it out. I was just pointing out
that figuring out whether or not it will be a win is not as easy as it
might first appear.
--
Fergus Henderson <[EMAIL PROTECTED]> | "I have always known that the pursuit
WWW: <http://www.cs.mu.oz.au/~fjh> | of excellence is a lethal habit"
PGP: finger [EMAIL PROTECTED] | -- the last words of T. S. Garp.
