[plt-scheme] Futures!

From: Synx (plt at synx.us.to)
Date: Mon Apr 12 20:06:39 EDT 2010

Wow, use of multiple processors? I'll be interested to see where this is
going! I'm a little confused though. I know currently you have to do a
lot of fixnum wangdoodles to get it to consider the procedure of the
future to be "safe" and not "freeze" (thereby serializing any
parallelization), but even with that I can't seem to get futures to show
any demonstrable difference on my machine. I made sure to compile with
--enable-futures, and I'm just on x86_64 Linux so there shouldn't be any
compatability problems.

I have two CPUs, so doing a lengthy operation with two futures should be
(somewhat less than) half as fast as doing the lengthy operation twice.
So I had hoped that the following script would behave in that fashion.
Unfortunately, futures don't seem to make even the slightest bit of
difference in the time it takes to execute the program. Is it that I am
in a multi-process environment, and it just can't spare more than one
processor at a time for each of the two futures? Or am I doing something
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: futures.ss
URL: <http://lists.racket-lang.org/users/archive/attachments/20100412/6f89cb36/attachment.ksh>

Posted on the users mailing list.