On Thu, 12 Apr 2007, H?kon Bugge wrote:
> The dataset is fixed, elapsed time includes
> initialization, write of animation files and
> more. Hence, slower per node performance would
> _scale_ better.
My comparison and measured scalability is based on each node's
speedup relative to their own 2p
At 11:22 AM 4/16/2007, Peter St. John wrote:
While looking around for a laptop I discovered that MS is paying
Google for the keyword "linux" to point to a "Get the Facts!" page
(reminds me of TheTruth.com) with testomonials about why MS is
better than Linux for clustering (!). What struck me wa
registers. AMD had to include x87 for 32-bit compatibility,
but there was no way they were going to beef up its number
of registers for 64-bit mode when they had the better SSE2
alternative.
making x87 better could be done fairly easily: just get rid of
the stack-based register set ;)
HPC f
Robert G. Brown wrote:
On Mon, 16 Apr 2007, Peter St. John wrote:
[...]
vendor to do the setup & support, and they don't say if that compares
economically to, say, paying Joe to do that, only that it saves them time
themselves. Sure, if I had that budget, I'd pay Joe and RGB to come
and wo
On 4/14/07, Greg Lindahl <[EMAIL PROTECTED]> wrote:
The two issues (extra registers and x87) aren't related. And most HPC
people have been avoiding the x87 for a long, long time.
Aren't they related by the introduction of the x86-64 ABI?
x87 provided 8 80-bit registers, while SSE2 on x86-64
wa
On Mon, 16 Apr 2007, Andrew M.A. Cater wrote:
This is directly associated with the wish for a minimal install -- I
have a system sitting upstairs right now that thinks that it has a
problem with a library that a) I've updated repeatedly so that I'm
certain that the image I'm installing from matc
On Mon, 16 Apr 2007 at 9:12pm, Andrew M.A. Cater wrote
On Mon, Apr 16, 2007 at 04:35:27PM -0400, Robert G. Brown wrote:
On Mon, 16 Apr 2007, Joshua Baker-LePain wrote:
I think you underestimate the amount of driver back-porting RH puts into
point releases. When I first got my dual woodcrest
On Mon, Apr 16, 2007 at 04:35:27PM -0400, Robert G. Brown wrote:
> On Mon, 16 Apr 2007, Joshua Baker-LePain wrote:
>
> >On Mon, 16 Apr 2007 at 2:12pm, Robert G. Brown wrote
> >
> >>Try installing two year old Centos AT ALL on six-month-old hardware, and
> >>I think that there is a very high probab
On Mon, 16 Apr 2007, Peter St. John wrote:
The ad is http://www.microsoft.com/windowsserver/facts/default.mspx?R=cf
The testamonials all seem to be equivalent to "I"m a PhD in Rocket Science
but setting up a Linux Cluster is too complex, so using MS saves me time to
do Rocket Science". None of
Peter St. John wrote:
While looking around for a laptop I discovered that MS is paying Google
for the keyword "linux" to point to a "Get the Facts!" page (reminds me
of TheTruth.com) with testomonials about why MS is better than Linux for
clustering (!). What struck me was that I wasn't googl
On Mon, 16 Apr 2007, Joshua Baker-LePain wrote:
On Mon, 16 Apr 2007 at 2:12pm, Robert G. Brown wrote
Try installing two year old Centos AT ALL on six-month-old hardware, and
I think that there is a very high probability that it will require a
much larger investment in time backporting kernels
From: "Peter St. John" <[EMAIL PROTECTED]> wrote
> The ad is http://www.microsoft.com/windowsserver/facts/default.mspx?R=cf
>
> The testamonials all seem to be equivalent to "I"m a PhD in Rocket Science
> but setting up a Linux Cluster is too complex, so using MS saves me
time to
> do Rocket Scie
Andrew,
Thanks, yeah, people seem to be "voting" for Thinkpads.
But, is it 8-year-old SON proof? :-)
Since besides RGB, IBM is the **other** 600-lb Gorilla of Linux Advocacy
(installing on a wristwatch, booting 1k instances thru VMWare on a 390,
etc), it saddens me that Thinkpads only come in a
On Mon, Apr 16, 2007 at 02:22:59PM -0400, Peter St. John wrote:
>
> Be all that as it may, I"m wondering what laptop to get. There are two
> issues: one, that it used to be scary to get all the device drivers for
> linux (maybe not anymore); and two, that one would prefer not to pay the tax
> asso
On Mon, Apr 16, 2007 at 09:39:44AM -0700, Greg Lindahl wrote:
> On Mon, Apr 16, 2007 at 06:44:47AM +, Andrew M.A. Cater wrote:
>
> > 17,740 packages in Debian main. Pure 64 bit distribution. Some
> > Beowulf-type software already packed. Runs out of the box on
> > Alpha/Sun/AMD64 (and will de
On Mon, 16 Apr 2007, Reuti wrote:
Now if we could just get e.g. the SGE folks and PVM folks to drive a
stake once and for all through aimk and convert their entire packages to
build clean for linux via Gnu autoconf/automake etc (for better or worse
-- the GBT are not at all perfect and in many w
On Mon, 16 Apr 2007 at 2:12pm, Robert G. Brown wrote
Try installing two year old Centos AT ALL on six-month-old hardware, and
I think that there is a very high probability that it will require a
much larger investment in time backporting kernels and worse.
I think you underestimate the amount
While looking around for a laptop I discovered that MS is paying Google for
the keyword "linux" to point to a "Get the Facts!" page (reminds me of
TheTruth.com) with testomonials about why MS is better than Linux for
clustering (!). What struck me was that I wasn't googling "linux cluster"
but jus
Robert G. Brown wrote:
On Sun, 15 Apr 2007, Mike Davis wrote:
one desire is supported stability. My number two desire is speed.
Maybe this philosophy comes from all of my years in unix world (21 and
counting), but the idea of standardizing on something that has the
limited longterm support o
Please join moderator and Beowulf cluster co-inventor Donald Becker on
Tuesday, April 17 for the next Bay Area Beowulf Users Group (BayBUG):
Bay Area Beowulf User Group (BayBUG)
April 17, 2007
2:30 - 5:00 p.m.
AMD headquarters Common Building,
Room C-6/7/8
991 Stewart Drive, Sunnyvale
There wi
On Mon, 16 Apr 2007, John Hearns wrote:
Robert G. Brown wrote:
On Sun, 15 Apr 2007, John Hearns wrote:
And re. the future version of Scientific Linux, there has been debate on
the list re. co-operating with CENTos and essentially using CENTos
IMO, most cluster builders will find it more
Andrew M.A. Cater wrote:
Debian: released every 18 months - two years. Guaranteed to support
previous version for a year after release. Seamless upgrade path - well,
nearly :)
17,740 packages in Debian main. Pure 64 bit distribution. Some
Beowulf-type software already packed. Runs out of th
As much as I love distribution debates, I have been really trying to
just let it happen and not get drawn in. I think I did a pretty good
job too! ... Well, as of 5 minutes ago.
How can anyone really take a distribution that does not support its
own packages for any reasonable amount of tim
On Sun, Apr 15, 2007 at 07:26:49PM -0400, Joe Landman wrote:
> This is one of the strongest arguments to certifying to an LSB and not a
> distro.
Er, you can't do that. LSB gives a philosophy for where things should
go. It says nothing about the exact versions. ISV codes have bugs with
some vers
On Monday 16 April 2007 11:13, Mark Hahn wrote:
> I only occasionally have to deal with it, so do not consider my
> opinion to be authoritative. as far as I can tell, it's a tool
> that dates back to the days when everyone as inventing their own
> way to do whole-tree builds. X had a similar thin
On Sun, 15 Apr 2007, Mike Davis wrote:
one desire is supported stability. My number two desire is speed. Maybe this
philosophy comes from all of my years in unix world (21 and counting), but
the idea of standardizing on something that has the limited longterm support
of FC scares me. We regul
opinion: cmake is precisely the same sort of
ingrown-project-creates-own-tool-badly as aimk.
Could you please expand on that a bit? I find to be a pretty good tool.
I only occasionally have to deal with it, so do not consider my
opinion to be authoritative. as far as I can tell, it's a tool
Now if we could just get e.g. the SGE folks and PVM folks to drive a
stake once and for all through aimk and convert their entire packages to
build clean for linux via Gnu autoconf/automake etc (for better or worse
or cmake - does anyone use it here - opinions?
http://www.cmake.org/
opinion:
On Monday 16 April 2007 10:18, Mark Hahn wrote:
> opinion: cmake is precisely the same sort of
> ingrown-project-creates-own-tool-badly as aimk.
Could you please expand on that a bit? I find to be a pretty good tool.
wt
--
Warren Turkal
___
Beowulf mai
I work for Penguin and it's my job to sell Scyld, however the following
also represents my personal
opinion.
The issue you mentioned below is why I love the light weight compute
node concept of Scyld,
which installs on top of redhat enterprise 4.
While it does still install everything that red
Robert G. Brown wrote:
I've always liked the idea of the core remaining a VERY marginal set
that is pretty much "just enough" to bootstrap an install. One of the
Hmmm I indicated this some time ago and got some grief over this.
Few of the distro makers seem to like this concept. I wan
On Monday 16 April 2007 06:32, Reuti wrote:
> or cmake - does anyone use it here - opinions?
>
> http://www.cmake.org/
If you are familiar with the plplot library, they use it.
wt
--
Warren Turkal
___
Beowulf mailing list, Beowulf@beowulf.org
To change
On Sun, 15 Apr 2007, Ed Hill wrote:
From a packaging (not a user perspective) there are a number of ways
that merging Core+Extras was/is a big improvement. Dependencies
between packages (Core items could not depend on Extras) was, for
instance, an annoying problem that now vanishes.
Fair enou
Now if we could just get e.g. the SGE folks and PVM folks to drive a
stake once and for all through aimk and convert their entire
packages to
build clean for linux via Gnu autoconf/automake etc (for better or
worse
-- the GBT are not at all perfect and in many ways suck)...
Hi,
or cmake -
Hi Mike
Mike Davis wrote:
Andy,
Debian is a possibility. I know that my friends in the UK and throughout
Europe like it.
FWIW: Ubuntu is quite good. Several of the versions have an N year
support window (3-5, I don't remember N pre-coffee).
I am migrating my laptop away from SuSE to Ub
On Sun, 15 Apr 2007, Ed Hill wrote:
If all of the bits used by a particular ISV code are user space then
they can use static linkage. And statically linked executables can be
rather portable between Linux distros. Alternatively, they can package
all the needed shared libs in some self-containe
On 16 Apr 2007, at 12:24 pm, Mike Davis wrote:
Andy,
Debian is a possibility. I know that my friends in the UK and
throughout Europe like it.
Debian's what we use on our 1,500 CPU cluster. I'm biassed, since
I'm a Debian Developer, but we found a lot of things easier to get
working on
Geoff Jacobs wrote:
Kyle Spaans wrote:
Wait, we can use openMOSIX and MPI at the same time? I thought they
were separate ideas? For example, MPI for multithreading and message
passing, and openMOSIX for just process migration. Can they be used at
the same time?
1) Spawn MPI processes on the h
Andy,
Debian is a possibility. I know that my friends in the UK and throughout
Europe like it.
We have been running CentOS for our current lifecycle of Opteron
machines. It is a trivial job to get most of the scientific software
that we run operational (including several commercial packages
Robert G. Brown wrote:
On Sun, 15 Apr 2007, John Hearns wrote:
And re. the future version of Scientific Linux, there has been debate
on the list re. co-operating with CENTos and essentially using CENTos
IMO, most cluster builders will find it more advantageous to track the
FC releases inst
Ed Hill wrote:
On Sun, 15 Apr 2007 "Robert G. Brown" <[EMAIL PROTECTED]> wrote:
And re. the future version of Scientific Linux, there has been
debate on the list re. co-operating with CENTos and essentially
If anyone is interested in helping out with (or even just curious
about) the Fedora pac
On Sun, Apr 15, 2007 at 10:01:58PM -0500, Mike Davis wrote:
> Joe is right about the stability factor.
>
> Stability, stability, stability.
>
>
> Mike Davis
>
>
Debian: released every 18 months - two years. Guaranteed to support
previous version for a year after release. Seamless upgrade path
42 matches
Mail list logo
