> From: Harper Theo <farnboro02.datasci.co.uk at BTHARPER>
> 
> I have just finished converting a Procyon Common Lisp application to run
> using ACL/Win 3.0.
> 
> However, I have seen a marked speed differential. [...]

 * * *

> From: Richard Barber <scientia.com at Richard_Barber>
> 
> [...] we rewrote the garbage collector, more than doubled its speed,
> and allowed it to recover from the "stack overflow because objects
> are too deeply nested bug".  As far as I know Franz still use the
> original garbage collector in ACL/Win 3.0.

  There have been a number of improvements and bug fixes to the aclwin
garbage collector since the original code.  Among them is a fix for
the "too deeply nested" bug.  This fix should have resulted in a
speed-up, but the fix took place a long time ago and I don't remember
our trying to determine how much faster the gc was after the fix.

  If it is the case that the time difference is due to garbage
collection, this may not be such a big problem as there are ways to
tune the aclwin garbage collector (cf., online reference "GC
Information" section).  If you would like to send comparison timings
to <franz.com" at "pc-support> (please use the preassigned label pcspr3291
in the subject line), we may be able to make some suggestions.  Also,
please send us information about the hardware/software configuration
with which you're doing the comparison.

	Charley

---
Charles A. Cox, Franz Inc.        1995 University Avenue, Suite 275
Internet: <franz.com at cox>           Berkeley, CA  94704
WWW: http://www.franz.com/        Phone: (510) 548-3600; FAX: (510) 548-8253