From: Sascha Matzke
Subject: Real multithreaded Lisp?
Date: 
Message-ID: <1143829039.326800.117000@z34g2000cwc.googlegroups.com>
Hi,

is there any Lisp implementation which is really multithreaded (os
level) for Sun Solaris?

I have a Sun T2000 here for evaluation and would like to try out
something "lispy" besides the normal Java work-related stuff.

I already tested AllegroCL which doesn't support os-threads under Unix
- so having 16 virtual / 4 physical CPUs doesn't really made a
difference. The test maxed out at 6.25% CPU usage (one of the sixteen
available CPUs).

Sascha

From: Peter Seibel
Subject: Re: Real multithreaded Lisp?
Date: 
Message-ID: <m2wteazctj.fsf@gigamonkeys.com>
"Sascha Matzke" <·············@gmail.com> writes:

> Hi,
>
> is there any Lisp implementation which is really multithreaded (os
> level) for Sun Solaris?
>
> I have a Sun T2000 here for evaluation and would like to try out
> something "lispy" besides the normal Java work-related stuff.
>
> I already tested AllegroCL which doesn't support os-threads under Unix
> - so having 16 virtual / 4 physical CPUs doesn't really made a
> difference. The test maxed out at 6.25% CPU usage (one of the sixteen
> available CPUs).

You might check out:

  <http://www.scieneer.com/scl/index.html>>

I've never used them but their claim to fame seems to be support for
SMP threading.

-Peter

-- 
Peter Seibel           * ·····@gigamonkeys.com
Gigamonkeys Consulting * http://www.gigamonkeys.com/
Practical Common Lisp  * http://www.gigamonkeys.com/book/
From: Tolstoy
Subject: Re: Real multithreaded Lisp?
Date: 
Message-ID: <1143830768.336141.126750@i39g2000cwa.googlegroups.com>
Interesting.  No indication on their pages about how to download the
system, or even how to purchase it.  Some sort of stealth marketing
plan?
From: Dmitry Gorbatovsky
Subject: Re: Real multithreaded Lisp?
Date: 
Message-ID: <e0jv9i$fkd$1@emma.aioe.org>
Peter Seibel wrote:

> "Sascha Matzke" <·············@gmail.com> writes:
> 
>> Hi,
>>
>> is there any Lisp implementation which is really multithreaded (os
>> level) for Sun Solaris?
>>
>> I have a Sun T2000 here for evaluation and would like to try out
>> something "lispy" besides the normal Java work-related stuff.
>>
>> I already tested AllegroCL which doesn't support os-threads under Unix
>> - so having 16 virtual / 4 physical CPUs doesn't really made a
>> difference. The test maxed out at 6.25% CPU usage (one of the sixteen
>> available CPUs).
> 
> You might check out:
> 
>   <http://www.scieneer.com/scl/index.html>>
> 
> I've never used them but their claim to fame seems to be support for
> SMP threading.
> 
> -Peter
> 
I also have interest in Scieneer(TM) implementation.
So maybe some one outthere would be so kind
to share an experience.

cheers
From: Alex Mizrahi
Subject: Re: Real multithreaded Lisp?
Date: 
Message-ID: <442d958c$0$15784$14726298@news.sunsite.dk>
(message (Hello 'Sascha)
(you :wrote  :on '(31 Mar 2006 10:17:19 -0800))
(

 SM> is there any Lisp implementation which is really multithreaded (os
 SM> level) for Sun Solaris?

 SM> I have a Sun T2000 here for evaluation and would like to try out
 SM> something "lispy" besides the normal Java work-related stuff.

Armed Bear Common Lisp is based on Java-machine and is multithreaded as long 
as Java is.

)
(With-best-regards '(Alex Mizrahi) :aka 'killer_storm)
"People who lust for the Feel of keys on their fingertips (c) Inity") 
From: Sascha Matzke
Subject: Re: Real multithreaded Lisp?
Date: 
Message-ID: <1143839305.031030.230780@z34g2000cwc.googlegroups.com>
Hi,

Alex Mizrahi wrote:
> Armed Bear Common Lisp is based on Java-machine and is multithreaded as long
> as Java is.

This looks promising and the idea to reuse the "infrastructure" work
already done by the JVM is a good a idea. The fact that

"ABCL's CLOS is intolerably slow and does not handle on-the-fly
redefinition of classes correctly. There is no support for the long
form of DEFINE-METHOD-COMBINATION, and certain other required CLOS
features are also missing. Enough CLOS is there to run ASDF and
CL-PPCRE, if you're in no hurry. There's no MOP worth mentioning."

and that I already made an investment in a AllegroCL licence (mainly
because of AllegroCache - which is really, really great) rule ABCL out
for the moment.

As far as I can see the biggest Lisp vendors (in my knowledge) Franz
and LispWorks both don't really support SMP - which by the way is a
great deficit from my point of view.

CPUs are beginning to become and going to be multi-core for the future
as standard. Add to that the engagement Franz Inc. is putting in all
kinds of web-application / Ajax things - throuput (the ability to
process a huge number of requests not  in the shortest but in an
acceptable amount of time) is becoming more and more important. I think
those vendors should consider to put a fair amount of R&D into making
their products capable of using the future computer infrastructure.



Sascha
From: Alain Picard
Subject: Re: Real multithreaded Lisp?
Date: 
Message-ID: <87k6a89vvf.fsf@memetrics.com>
"Sascha Matzke" <·············@gmail.com> writes:

> I think those vendors should consider to put a fair amount of R&D
> into making their products capable of using the future computer
> infrastructure.

And I expect that that's precisely what they're doing.  At least,
that's more or less what Lispworks 5 is touted to be, IIUC.

Unfortunately, that WAS due in Q1 of this year, then readusted to Q2...
but hey, I'm a software developer, I'm not about to cast the first stone
for missing a shipping date.   ;-)

            --ap
From: Douglas Crosher
Subject: Re: Real multithreaded Lisp?
Date: 
Message-ID: <442DBD1F.2010704@scieneer.com>
Sascha Matzke wrote:
> Hi,
> 
> is there any Lisp implementation which is really multithreaded (os
> level) for Sun Solaris?
> 
> I have a Sun T2000 here for evaluation and would like to try out
> something "lispy" besides the normal Java work-related stuff.
> 
> I already tested AllegroCL which doesn't support os-threads under Unix
> - so having 16 virtual / 4 physical CPUs doesn't really made a
> difference. The test maxed out at 6.25% CPU usage (one of the sixteen
> available CPUs).
> 
> Sascha
> 

Then Scieneer Common Lisp offers multi-threaded support for the SPARC Solaris
platform.  Please contact in private for more information.

Regards
Douglas Crosher