At 06:06 PM 11/7/2001 -0500, Ken Fox wrote: >Dan Sugalski wrote: > > At 04:29 PM 11/7/2001 -0500, James Mastros wrote: > > > On Wed, Nov 07, 2001 at 10:15:07AM -0500, Ken Fox wrote: > > > > If Perl can keep the loop index in an integer register, then Parrot > > > > could use fast loop ops. IMHO there's no point in using fast loop ops > > > > if taking the length of @vector is expensive. > > > > > > Huha? I don't see what the problem is with keeping the index in an I > > > register. The problem is with keeping the _current element_ in an I > > > register. > > > > Yup. Hidden things--loop counters, iterators, stuff like that--can and > > probably will use the I registers. > >If the interfaces between "hidden things" and PMCs are expensive >then I don't see the point in putting those things in non-PMC >registers. For example, if a loop index is compared to the >length of an array, Parrot is going to have to convert the index >to a PMC so it can be compared to the PMC array length.
No it isn't. It can get the integer length of the array and stuff it in a register at the beginning of the loop, or do an integer compare when it needs to, depending on the semantics of the loop. Dan --------------------------------------"it's like this"------------------- Dan Sugalski even samurai [EMAIL PROTECTED] have teddy bears and even teddy bears get drunk