Re: fastest lisp machine?

Rainer Joswig <[email protected]> Tue, 14 Jan 2014 22:06:59 +0100
Newsgroups gmane.lisp.mcl.general
Message-ID <[email protected]>
This is a multi-part MIME message.

--===============4973004045695142423==
Content-Type: multipart/alternative; boundary="=_reb-r3D663197-t52D5A6F8"

This is a multi-part MIME message.

--=_reb-r3D663197-t52D5A6F8
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=windows-1252

Hi,

CL-HTTP runs best in LispWorks. Generally the 64bit LispWorks is extremely =
fast (like twice compared to the 32bit version) and has excellent SMP suppo=
rt. Generally I think that LispWorks is the best CL implementation. Sorry, =
MCL users. ;-)

Note that you also need to tune the typical OS to accept enough concurrent =
connections, etc.

Note that one also might need to tune the GC settings of LispWorks for dema=
nding production use a bit.
One drawback of all Lisp implementations (besides ABCL) is that they have n=
o concurrent GC (one that runs concurrently to the other Lisp processes). T=
hus one might need to check the GC behavior under load.

The LispWorks port of CL-HTTP is also more capable and more optimized than =
the other variants.

There are other good choices for multi-core servers - I would usually recom=
mend Clozure CL. I use CL-HTTP with Clozure CL on an ARM based quad-core co=
mputer. But the port of CL-HTTP to Clozure CL is a bit primitive: it lacks =
optimizations and capabilities. For a simple home server it can be enough, =
though. Clozure CL itself has SMP support, has a good GC, is widely ported =
and the developers know their job. Useless to say, I'm a fan of Clozure CL.=
 ;-)
See:  http://lispm.dyndns.org/ccl

If you look at the new Mac Pro, then something like LispWorks and CCL gives=
 you support for the multi-core CPU. What they don't give you is direct sup=
port for computations on the GPUs (graphics cards). It might be useful/poss=
ible to run simulation code on the GPUs - not sure how you would best appro=
ach that from Lisp.

Regards,

Rainer Joswig





Am 14.01.2014 um 05:26 schrieb Mark Klein <m_klein-DPNOqEs/[email protected]>:

>=20
> I=92d like to get the fastest lisp machine possible (within reason) for r=
unning the cl-http server as well as number-crunching-heavy simulations. Wh=
ich hardware should I get? Are there are versions of common lisp that can t=
ake advantage of the multi-core multiprocessor architectures the new hardwa=
re (e.g. the new Apple desktop) is coming out with?
>=20
> -------------------------------
> Mark Klein
> http://cci.mit.edu/klein
>=20
> Principal Research Scientist
> Center for Collective Intelligence
> Massachusetts Institute of Technology
>=20
> Visiting Researcher
> Dynamic and Distributed Information Systems Group
> University of Zurich
>=20
> _______________________________________________
> info-mcl mailing list
> [email protected]
> http://clozure.com/mailman/listinfo/info-mcl


--=_reb-r3D663197-t52D5A6F8
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
	charset=windows-1252

<html><head></head><body style=3D"word-wrap: break-word; -webkit-nbsp-mode:=
 space; -webkit-line-break: after-white-space;">Hi,<div><br></div><div>CL-H=
TTP runs best in LispWorks. Generally the 64bit LispWorks is extremely fast=
 (like twice compared to the 32bit version) and has excellent SMP support. =
Generally I think that LispWorks is the best CL implementation. Sorry, MCL =
users. ;-)</div><div><br></div><div>Note that you also need to tune the typ=
ical OS to accept enough concurrent connections, etc.</div><div><br></div><=
div>Note that one also might need to tune the GC settings of LispWorks for =
demanding production use a bit.</div><div>One drawback of all Lisp implemen=
tations (besides ABCL) is that they have no concurrent GC (one that runs co=
ncurrently to the other Lisp processes). Thus one might need to check the G=
C behavior under load.</div><div><br></div><div>The LispWorks port of CL-HT=
TP is also more capable and more optimized than the other variants.</div><d=
iv><br></div><div>There are other good choices for multi-core servers - I w=
ould usually recommend Clozure CL. I use CL-HTTP with Clozure CL on an ARM =
based quad-core computer. But the port of CL-HTTP to Clozure CL is a bit pr=
imitive: it lacks optimizations and capabilities. For a simple home server =
it can be enough, though. Clozure CL itself has SMP support, has a good GC,=
 is widely ported and the developers know their job. Useless to say, I'm a =
fan of Clozure CL. ;-)</div><div>See: &nbsp;<a href=3D"http://lispm.dyndns.=
org/ccl">http://lispm.dyndns.org/ccl</a></div><div><br></div><div>If you lo=
ok at the new Mac Pro, then something like LispWorks and CCL gives you supp=
ort for the multi-core CPU. What they don't give you is direct support for =
computations on the GPUs (graphics cards). It might be useful/possible to r=
un simulation code on the GPUs - not sure how you would best approach that =
from Lisp.</div><div><br></div><div>Regards,</div><div><br></div><div>Raine=
r Joswig</div><div><br></div><div><br></div><div><br></div><div><br></div><=
div><br></div><div><div><div>Am 14.01.2014 um 05:26 schrieb Mark Klein &lt;=
<a href=3D"mailto:m_klein-DPNOqEs/[email protected]">m_klein-DPNOqEs/[email protected]</a>&gt;:</div><br class=
=3D"Apple-interchange-newline"><blockquote type=3D"cite"><div style=3D"word=
-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-whit=
e-space;"><div><br></div><div>I=92d like to get the fastest lisp machine po=
ssible (within reason) for running the cl-http server as well as number-cru=
nching-heavy simulations. Which hardware should I get? Are there are versio=
ns of common lisp that can take advantage of the multi-core multiprocessor =
architectures the new hardware (e.g. the new Apple desktop) is coming out w=
ith?</div><br><div apple-content-edited=3D"true">
<div style=3D"letter-spacing: normal; orphans: auto; text-align: start; tex=
t-indent: 0px; text-transform: none; white-space: normal; widows: auto; wor=
d-spacing: 0px; -webkit-text-stroke-width: 0px; word-wrap: break-word; -web=
kit-nbsp-mode: space; -webkit-line-break: after-white-space;"><div style=3D=
"letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px=
; text-transform: none; white-space: normal; widows: auto; word-spacing: 0p=
x; -webkit-text-stroke-width: 0px; word-wrap: break-word; -webkit-nbsp-mode=
: space; -webkit-line-break: after-white-space;"><span class=3D"Apple-style=
-span" style=3D"border-collapse: separate; border-spacing: 0px;"><div style=
=3D"word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: af=
ter-white-space;"><span class=3D"Apple-style-span" style=3D"border-collapse=
: separate; font-family: Helvetica; font-style: normal; font-variant: norma=
l; font-weight: normal; letter-spacing: normal; line-height: normal; orphan=
s: 2; text-indent: 0px; text-transform: none; white-space: normal; widows: =
2; word-spacing: 0px; border-spacing: 0px; -webkit-text-decorations-in-effe=
ct: none; -webkit-text-stroke-width: 0px;"><div style=3D"word-wrap: break-w=
ord; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;"><spa=
n class=3D"Apple-style-span" style=3D"border-collapse: separate; font-famil=
y: Helvetica; font-style: normal; font-variant: normal; font-weight: normal=
; letter-spacing: normal; line-height: normal; orphans: 2; text-indent: 0px=
; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; =
border-spacing: 0px; -webkit-text-decorations-in-effect: none; -webkit-text=
-stroke-width: 0px;"><div style=3D"word-wrap: break-word; -webkit-nbsp-mode=
: space; -webkit-line-break: after-white-space;"><span class=3D"Apple-style=
-span" style=3D"border-collapse: separate; font-family: Helvetica; font-siz=
e: 12px; font-style: normal; font-variant: normal; font-weight: normal; let=
ter-spacing: normal; line-height: normal; orphans: 2; text-indent: 0px; tex=
t-transform: none; white-space: normal; widows: 2; word-spacing: 0px; borde=
r-spacing: 0px; -webkit-text-decorations-in-effect: none; -webkit-text-stro=
ke-width: 0px;"><div style=3D"word-wrap: break-word; -webkit-nbsp-mode: spa=
ce; -webkit-line-break: after-white-space;">-------------------------------=
<br>Mark Klein<br><a href=3D"http://cci.mit.edu/klein">http://cci.mit.edu/k=
lein</a><br><br>Principal Research Scientist<br>Center for Collective Intel=
ligence<br>Massachusetts Institute of Technology<br><br>Visiting Researcher=
<br>Dynamic and Distributed Information&nbsp;Systems Group<br>University of=
 Zurich</div></span></div></span></div></span></div></span></div></div>
</div>
<br></div>_______________________________________________<br>info-mcl maili=
ng list<br><a href=3D"mailto:[email protected]">[email protected]</a>=
<br>http://clozure.com/mailman/listinfo/info-mcl<br></blockquote></div><br>=
</div></body></html>
--=_reb-r3D663197-t52D5A6F8--


--===============4973004045695142423==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
info-mcl mailing list
[email protected]
http://clozure.com/mailman/listinfo/info-mcl

--===============4973004045695142423==--