>>>>> "Robert" == Robert Dodier writes:
Robert> OK, I have modified the CMUCL MT rng (rand-mt19937.lisp) so
Robert> Clisp and GCL are happy with it.
Cool.
Robert> The down side is that it's substantially slower than the
Robert> unmodified mt19937. I cut out calls to the CMUCL vm stuff,
Robert> which include a central calculation coded in x86 assembler.
Robert> I assumed that's not portable, maybe I'm wrong about that.
I'm a bit surprised. How did you compare the generation speed between
your version and the unmodified mt19937? The x86 assembler is for
x86, of course. But on all other platforms the portable version is
used, and I've looked at the generated sparc code; I don't think an
assembly version would be much faster.
Ray