gcl-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gcl-devel] efficiency of package symbol tables


From: Camm Maguire
Subject: Re: [Gcl-devel] efficiency of package symbol tables
Date: 21 Mar 2004 09:52:25 -0500
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2

Greetings!  Found a spare moment to confirm this.  It appears that
Peter is right, and that our package hashing is OK.  This is made
clearer by the new time macro I've put in.  Here are my timings on the
latest 2.6.1:

=============================================================================
>(time (fill-symtab 10000))


real time       :      0.660 secs
run-gbc time    :      0.050 secs
child run time  :      0.000 secs
gbc time        :      0.400 secs
NIL

>(time (fill-symtab 10000))


real time       :      0.490 secs
run-gbc time    :      0.090 secs
child run time  :      0.000 secs
gbc time        :      0.400 secs
NIL

>(time (fill-symtab 10000))


real time       :      0.560 secs
run-gbc time    :      0.070 secs
child run time  :      0.000 secs
gbc time        :      0.490 secs
NIL

>(time (fill-symtab 10000))


real time       :      0.700 secs
run-gbc time    :      0.070 secs
child run time  :      0.000 secs
gbc time        :      0.630 secs
NIL

>(time (fill-symtab 10000))


real time       :      0.750 secs
run-gbc time    :      0.080 secs
child run time  :      0.000 secs
gbc time        :      0.660 secs
NIL

>(time (fill-symtab 10000))


real time       :      0.900 secs
run-gbc time    :      0.080 secs
child run time  :      0.000 secs
gbc time        :      0.820 secs
NIL

>(time (fill-symtab 10000))


real time       :      0.990 secs
run-gbc time    :      0.100 secs
child run time  :      0.000 secs
gbc time        :      0.880 secs
NIL

>(time (fill-symtab 10000))


real time       :      1.030 secs
run-gbc time    :      0.070 secs
child run time  :      0.000 secs
gbc time        :      0.960 secs
NIL

>(time (fill-symtab 10000))


real time       :      1.210 secs
run-gbc time    :      0.100 secs
child run time  :      0.000 secs
gbc time        :      1.110 secs
NIL

>(time (fill-symtab 10000))


real time       :      1.310 secs
run-gbc time    :      0.080 secs
child run time  :      0.000 secs
gbc time        :      1.230 secs
NIL

>(time (fill-symtab 10000))


real time       :      1.310 secs
run-gbc time    :      0.110 secs
child run time  :      0.000 secs
gbc time        :      1.200 secs
NIL

>(time (fill-symtab 10000))


real time       :      1.420 secs
run-gbc time    :      0.090 secs
child run time  :      0.000 secs
gbc time        :      1.330 secs
NIL

>(time (fill-symtab 10000))


real time       :      1.810 secs
run-gbc time    :      0.130 secs
child run time  :      0.000 secs
gbc time        :      1.630 secs
NIL

>(time (fill-symtab 10000))


real time       :      1.720 secs
run-gbc time    :      0.080 secs
child run time  :      0.000 secs
gbc time        :      1.630 secs
NIL

>(time (fill-symtab 10000))


real time       :      1.740 secs
run-gbc time    :      0.110 secs
child run time  :      0.000 secs
gbc time        :      1.620 secs
NIL

>(time (fill-symtab 10000))


real time       :      1.480 secs
run-gbc time    :      0.070 secs
child run time  :      0.000 secs
gbc time        :      1.410 secs
NIL

>(time (fill-symtab 10000))


real time       :      1.690 secs
run-gbc time    :      0.060 secs
child run time  :      0.000 secs
gbc time        :      1.620 secs
NIL

>(setq si::*notify-gbc* t)

T

>(time (fill-symtab 10000))

[GC for 32 STRING pages..(T=10).GC finished]
[GC for 32 STRING pages..(T=11).GC finished]
[GC for 32 STRING pages..(T=11).GC finished]
[GC for 32 STRING pages..(T=11).GC finished]
[GC for 32 STRING pages..(T=12).GC finished]
[GC for 32 STRING pages..(T=12).GC finished]
[GC for 32 STRING pages..(T=12).GC finished]
[GC for 32 STRING pages..(T=12).GC finished]
[GC for 32 STRING pages..(T=11).GC finished]
[GC for 32 STRING pages..(T=11).GC finished]
[GC for 32 STRING pages..(T=11).GC finished]
[GC for 32 STRING pages..(T=11).GC finished]
[GC for 32 STRING pages..(T=11).GC finished]
[GC for 32 STRING pages..(T=11).GC finished]
[GC for 32 STRING pages..(T=12).GC finished]

real time       :      1.800 secs
run-gbc time    :      0.110 secs
child run time  :      0.000 secs
gbc time        :      1.690 secs
NIL

>
=============================================================================


Please let me know if I'm overlooking something.  And please let me
thank both Bruno and Sam again for bringing this to our attention.

BTW, Peter you are of course exactly right regarding the mess of our
error system.  I'm hoping we can clean at least most of this up in
2.7.  In fact, I still have your old postings on the issue!

Take care,

Peter Wood <address@hidden> writes:

> On Sat, Mar 20, 2004 at 02:05:17PM -0500, Camm Maguire wrote:
> > Greetings, and please let me thank both you and Sam for bringing this
> > to our attention.  I'll try to take a look, most likely most 2.6.2.
> > 
> > Take care,
> > 
> > Bruno Haible <address@hidden> writes:
> > 
> > > Mike Thomas wrote:
> > > > I tried duplicating the symbol table problem on a Windows XP box with a
> > > > pre-release of the ANSI GCL 2.6.2 but could not get that explosion in 
> > > > time
> > > > you report after more than 30 iterations, as shown below.
> > > 
> > > Still, you get a heavy slowdown, by a factor of 7 in your case:
> > > 
> > > > >(time (fill-symtab 10000))
> > > >
> > > > real time : 0.367 secs
> > > > run time  : 0.367 secs
> > > > NIL
> > > > ...
> > > > >(time (fill-symtab 10000))
> > > >
> > > > real time : 2.467 secs
> > > > run time  : 2.467 secs
> > > > NIL
> > > 
> > > So the bug is still there in your 2.6.2-pre build.
> > > 
> > > Bruno
> > > 
> 
> Hi
> 
> I'm pretty certain this is not a (package hash) bug, but is arising
> due to inadequate allocation settings.  The first reply I sent to
> Bruno's post was a red herring, but the second documents how with more
> generous allocation settings, the times do _not_ deteriorate
> significantly.
> 
> My posts took place in the middle of the mailing-list delays, so you
> might not have seen the second.  If you want I will send it to you
> again.
> 
> Regards,
> Peter
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah




reply via email to

[Prev in Thread] Current Thread [Next in Thread]