From: Ken Forbus

Subject: Re: Suggestions for tuning GC performance?

Date: 2000-6-15 17:54

Thank you for the summary; we will try out this advice.

         Ken

At 02:06 PM 6/15/00 +0200, Francis Leboutte wrote:
>At 14/06/00 17:48 -0700, Ken Forbus wrote:
>>In an application we are building, several global GC's occur at an >>inconvenient time, e.g., while the user is waiting for something to >>happen. What are the appropriate parameters to tune to (a) ramp up the >>total space available, and (b) make global GC's less frequent, by doing >>more growth when each one occurs? I've looked over the documentation and >>wasn't able to figure this out from what was there, but of course I could >>have overlooked something. Thanks in advance for any insights you can share.
> >Attached 2 recent posts from COMP.LANG.Lisp that should help you. >Especially interesting is the description of new GC features that will be >available in ACL6. >Two more things : >- you could trigger global GC at some well chosen time (e.g. after a long >computation). >- with the appropriate settings the efficiency of a generational GC should >be greater than 95% like in the following: > scavenging...done eff: 99%, copy new: 2931064 + old: 528688 = 3459752 > > > > >-- >Francis Leboutte >Algorithme, Rue de la Charrette 141, 4130 Tilff, Belgium > <algo.be at fl> <acm.org at leboutte> > www.algo.be +32-(0)4.388.3919 > >
From 9/99 to 8/00: Kenneth D. Forbus Xerox PARC, 3333 Coyote Hill Road, Palo Alto, CA, 94304, USA Voice: (650) 812-4891; Fax: (650) 812-4334 Otherwise: Kenneth D. Forbus Professor of Computer Science and Education Computer Science Department, Northwestern University 1890 Maple Avenue, Evanston, IL, 60201, USA Voice: (847) 491-7699; Fax: (847) 491-5258