From: Larry A. Taylor
Subject: Allegro for NT garbage collection: 3.0 and 5.0
Date:
Message-ID: <6o3dnt$8l$1@dropit.pgh.net>
This follows on the recent discussion about generational
garbage collection. The contributers seemed to think that
Allegro and most modern CLs did generational garbage collection.
Does anybody know if ACL 3.0, which runs on NT, has a generational
garbage-collection scheme? 3.0 Differs in many respects from the
Unix ACL. Allegro 4.3.2, however, was a close port from Unix.
If there is anything wrong with GC on ACL 3.0 for NT, is it fixed
in 5.0? ACL 5.0 is supposed to be based on the Unix versions, but
with the Windows interfaces we need.
LAT
From: Lyman S. Taylor
Subject: Re: Allegro for NT garbage collection: 3.0 and 5.0
Date:
Message-ID: <6o4dpk$snc@pravda.cc.gatech.edu>
In article <···········@dropit.pgh.net>,
Larry A. Taylor <·······@wisdomcorp.com> wrote:
>This follows on the recent discussion about generational
>garbage collection. The contributers seemed to think that
>Allegro and most modern CLs did generational garbage collection.
>
>Does anybody know if ACL 3.0, which runs on NT, has a generational
>garbage-collection scheme?
If you enter the expression:
(room t )
at the listener prompt, you might be able to see for yourself. It may be
labeled ephemeral instead of generation. Hopefully, the implementation
would report information about the generations if it has them. It
isn't required, but seemingly printing "maximal information" would include
this info.
--
Lyman S. Taylor "Computers are too reliable to replace
(·····@cc.gatech.edu) humans effectively."
Commander Nathan Spring, "Starcops"
I believe that ACL 3.0 was originally based on Procyon Common Lisp.
This had (has?) a generational garbage collector. Running (room t) as
proposed in my version of Procyon gives the following information:-
> (room t)
351,444 bytes free before next GC.
0 garbage collections so far.
Total GC time: 0.00 seconds (0 clock ticks)
Time since startup: 9.88 seconds (9879 clock ticks)
Used heap space 5,257,104 bytes
Free heap space 128,764,024 bytes.
4 generations.
Size of first generation 524,288 bytes.
Promote after surviving 3 GCs.
GC next generation after 6 GCs.
Root table size 16384.
Root table entries left 15680.
nil
>
>
In article <··········@pravda.cc.gatech.edu>, Lyman S. Taylor
<·····@cc.gatech.edu> writes
>In article <···········@dropit.pgh.net>,
>Larry A. Taylor <·······@wisdomcorp.com> wrote:
>>This follows on the recent discussion about generational
>>garbage collection. The contributers seemed to think that
>>Allegro and most modern CLs did generational garbage collection.
>>
>>Does anybody know if ACL 3.0, which runs on NT, has a generational
>>garbage-collection scheme?
>
> If you enter the expression:
>
> (room t )
>
> at the listener prompt, you might be able to see for yourself. It may be
> labeled ephemeral instead of generation. Hopefully, the implementation
> would report information about the generations if it has them. It
> isn't required, but seemingly printing "maximal information" would include
> this info.
>
>
>
>
>
--
Anthony Willoughby. Principal, Ascham Consulting.
Object oriented specification and design.
Email: ···@ascham.demon.co.uk. Home page: www.ascham.demon.co.uk
Tel: +44(0)1223-836995. Mobile: 0973-801016
From: Johannes Beck
Subject: Re: Allegro for NT garbage collection: 3.0 and 5.0
Date:
Message-ID: <1o1jo6.rd.ln@cygnus.x1>
Anthony Willoughby schrieb in Nachricht
<················@ascham.demon.co.uk>...
>I believe that ACL 3.0 was originally based on Procyon Common Lisp.
>This had (has?) a generational garbage collector. Running (room t) as
>proposed in my version of Procyon gives the following information:-
>
>> (room t)
>351,444 bytes free before next GC.
>0 garbage collections so far.
>Total GC time: 0.00 seconds (0 clock ticks)
>Time since startup: 9.88 seconds (9879 clock ticks)
>Used heap space 5,257,104 bytes
>Free heap space 128,764,024 bytes.
>4 generations.
>Size of first generation 524,288 bytes.
>Promote after surviving 3 GCs.
>GC next generation after 6 GCs.
>Root table size 16384.
>Root table entries left 15680.
>nil
>>
>In article <··········@pravda.cc.gatech.edu>, Lyman S. Taylor
><·····@cc.gatech.edu> writes
>>In article <···········@dropit.pgh.net>,
>>Larry A. Taylor <·······@wisdomcorp.com> wrote:
>>>This follows on the recent discussion about generational
>>>garbage collection. The contributers seemed to think that
>>>Allegro and most modern CLs did generational garbage collection.
>>>
>>>Does anybody know if ACL 3.0, which runs on NT, has a generational
>>>garbage-collection scheme?
>>
>> If you enter the expression:
>>
>> (room t )
>>
>> at the listener prompt, you might be able to see for yourself. It may
be
>> labeled ephemeral instead of generation. Hopefully, the implementation
>> would report information about the generations if it has them. It
>> isn't required, but seemingly printing "maximal information" would
include
>> this info.
The file "allegro.ini" contains some settings for GC which seem to indicate
that GC
on ACL3.0 is generational.
--
[Allegro CL]
HeapSize=32.00
GC_Space=0.25
No_of_Generations=4
Promote_After_GCs=3
GCs_Per_Generation=6
Dynamic_GC=0
Explicit_GC=-1
ImageName=
--
Bye,
Johannes Beck