> As well as fixing all the expressions and core opcodes to
> handle this, you'd have to add more logic in the static
> analyzer to detect when it is possible to do block processing
> (both sfront and SAINT do this, I think)
Actually, sfront doesn't do this (yet) -- while its a clear
win for interpeted or p-code systems, its less clear what the
impact will be in a fully compiled system like sfront, especially
once the effect of the cache is factored in.
One of the things we're planning to do is to hand-code up some sa.c
both blocked and unblocked, and measure them on various systems, to
see if the win is worth putting the feature into sfront ...
--jl
This archive was generated by hypermail 2b29 : Wed May 10 2000 - 12:15:46 EDT