I'd like to share an interesting historical tidbit I have run across
while browsing the archive of the CLIM mailing list. It's a
performance comparison between Lisp Machines and contemporary general
purpose workstations. In this message:
Date: Wed, 4 Sep 1991 11:33-0400
From: Scott McKay <···@SAPSUCKER.SCRC.Symbolics.COM>
Subject: CLIM philosophy wrt to X.
Message-ID: <····················@EVENING-GROSBEAK.SCRC.Symbolics.COM>
Scott McKay wrote:
As for statistics, CLIM 1.0 has roughly equivalent run-time performance
on an XL400 running Genera (on Genera's native window system), on a 30MHz
486 running CLOE and MicroSoft Windows, and on a Sparcstation 1 (that's
one, not one-plus or two) running Franz Allegro or Lucid. I believe that
the performance on MCL 2.0 on a Mac IIfx is also comparable. By
comparable, I mean that the XL400 wins some of the benchmarks, CLOE wins
some, Allegro wins some, and Lucid wins some. No single platform does
markedly worse than any other in their poorest benchmarks.
This gives an idea of Lisp Machine performance to those, like me, who have
never used one, but who used PCs.
Paolo
--
Paolo Amoroso <·······@mclink.it>
From: Scott McKay
Subject: Re: LispM vs general purpose workstation performance (historical tidbit)
Date:
Message-ID: <La_mb.37329$Fm2.16142@attbi_s04>
"Paolo Amoroso" <·······@mclink.it> wrote in message
···················@plato.moon.paoloamoroso.it...
> I'd like to share an interesting historical tidbit I have run across
> while browsing the archive of the CLIM mailing list. It's a
> performance comparison between Lisp Machines and contemporary general
> purpose workstations. In this message:
>
> Date: Wed, 4 Sep 1991 11:33-0400
> From: Scott McKay <···@SAPSUCKER.SCRC.Symbolics.COM>
> Subject: CLIM philosophy wrt to X.
> Message-ID: <····················@EVENING-GROSBEAK.SCRC.Symbolics.COM>
>
> Scott McKay wrote:
>
> As for statistics, CLIM 1.0 has roughly equivalent run-time performance
> on an XL400 running Genera (on Genera's native window system), on a
30MHz
> 486 running CLOE and MicroSoft Windows, and on a Sparcstation 1 (that's
> one, not one-plus or two) running Franz Allegro or Lucid. I believe
that
> the performance on MCL 2.0 on a Mac IIfx is also comparable. By
> comparable, I mean that the XL400 wins some of the benchmarks, CLOE wins
> some, Allegro wins some, and Lucid wins some. No single platform does
> markedly worse than any other in their poorest benchmarks.
>
> This gives an idea of Lisp Machine performance to those, like me, who have
> never used one, but who used PCs.
>
Cool tidbit.
The XL1200 and the original "Open Genera" on the original
DEC alphas were about 3x faster than the XL400, meaning
that an XL1200 was a full 100MHz machine. Woo hoo!
In article <·····················@attbi_s04>, Scott McKay
<···········@comcast.net> wrote:
> "Paolo Amoroso" <·······@mclink.it> wrote in message
> ···················@plato.moon.paoloamoroso.it...
> > I'd like to share an interesting historical tidbit I have run across
> > while browsing the archive of the CLIM mailing list. It's a
> > performance comparison between Lisp Machines and contemporary general
> > purpose workstations. In this message:
> >
> > Date: Wed, 4 Sep 1991 11:33-0400
> > From: Scott McKay <···@SAPSUCKER.SCRC.Symbolics.COM>
> > Subject: CLIM philosophy wrt to X.
> > Message-ID: <····················@EVENING-GROSBEAK.SCRC.Symbolics.COM>
> >
> > Scott McKay wrote:
> >
> > As for statistics, CLIM 1.0 has roughly equivalent run-time performance
> > on an XL400 running Genera (on Genera's native window system), on a
> 30MHz
> > 486 running CLOE and MicroSoft Windows, and on a Sparcstation 1 (that's
> > one, not one-plus or two) running Franz Allegro or Lucid. I believe
> that
> > the performance on MCL 2.0 on a Mac IIfx is also comparable. By
> > comparable, I mean that the XL400 wins some of the benchmarks, CLOE wins
> > some, Allegro wins some, and Lucid wins some. No single platform does
> > markedly worse than any other in their poorest benchmarks.
> >
> > This gives an idea of Lisp Machine performance to those, like me, who have
> > never used one, but who used PCs.
> >
>
> Cool tidbit.
>
> The XL1200 and the original "Open Genera" on the original
> DEC alphas were about 3x faster than the XL400, meaning
> that an XL1200 was a full 100MHz machine. Woo hoo!
For the record, the XL400's Ivory (processor) had a cycle time of 150ns
which meant it ran at 6.67MHz.
The XL1200 was supposed to be three times faster (i.e., 50ns cycle
time) but the revised Ivory chip turned out to be unstable at that
speed. The XL1200 was shipped with a 65ns cycle time, or 15.4MHz.
From: Joe Marshall
Subject: Re: LispM vs general purpose workstation performance (historical tidbit)
Date:
Message-ID: <ekwyr838.fsf@ccs.neu.edu>
Paolo Amoroso <·······@mclink.it> writes:
> I'd like to share an interesting historical tidbit I have run across
> while browsing the archive of the CLIM mailing list. It's a
> performance comparison between Lisp Machines and contemporary general
> purpose workstations. In this message:
>
> Date: Wed, 4 Sep 1991 11:33-0400
> From: Scott McKay <···@SAPSUCKER.SCRC.Symbolics.COM>
> Subject: CLIM philosophy wrt to X.
> Message-ID: <····················@EVENING-GROSBEAK.SCRC.Symbolics.COM>
>
> Scott McKay wrote:
>
> As for statistics, CLIM 1.0 has roughly equivalent run-time performance
> on an XL400 running Genera (on Genera's native window system), on a 30MHz
> 486 running CLOE and MicroSoft Windows, and on a Sparcstation 1 (that's
> one, not one-plus or two) running Franz Allegro or Lucid. I believe that
> the performance on MCL 2.0 on a Mac IIfx is also comparable. By
> comparable, I mean that the XL400 wins some of the benchmarks, CLOE wins
> some, Allegro wins some, and Lucid wins some. No single platform does
> markedly worse than any other in their poorest benchmarks.
>
> This gives an idea of Lisp Machine performance to those, like me, who have
> never used one, but who used PCs.
Circa 1984 the LMI Lambda had a microcode cycle time of about 250 ns.
(4 MHz!) I think a macro-instruction took at least 2 cycles in the
best case, and that could only happen on the `odd' instructions
because they were in the upper half of the word. A function call took
100 microinstructions (!). I remember that the Tak benchmark took
about 7 seconds. (To contrast, CMUCL 18c on PentiumIII at 1.1GHz does
it in .8)