On 4 Apr 2009, at 10:29 pm, Jason Riedy wrote:
And Joe Landman writes:
Good performance:
--
GlusterFS
PVFS2
I don't suppose you've experimented with Ceph or POHMELFS? I just
attempted to build Lustre support for experimenting and remembered
why I avoid it.
And Lustre
And Greg Lindahl writes:
An example would be HDFS, the Hadoop Distributed FS.
Last time I checked, it strongly encouraged marshaling into text.
The libraries didn't have an obvious way to handle binary data,
but that was from a somewhat cursory read by someone less than
fully interested in Java
2009/4/5 Jason Riedy ja...@acm.org:
I'm more shocked that no one has written up using cfengine for
managing laptops. It seems a perfect model. With the more open
development model, perhaps it'll come back. But its competitors
are more web 2.0 cool.
Indeed. We are all just lumbering
Definitely, it's a non-cluster question, but may be
I'll get an answer: is there any remarkable difference
in FPU performance between Extreme and non-Extreme
Intel CPUs, e.g. QX9650 and Q9650?
Dmitry Zaletnev
___
Beowulf mailing list,
On Thu, Apr 02, 2009 at 03:16:22PM +0200, Simon Hogg wrote:
Robert G. Brown wrote:
IIRC Google doesn't use server grade anything. They use OTC parts and
do a running computation on failure rates and optimize price performance
dynamically. They are truly industrial scale production here.
Would a Sun/SGI merger have been a marriage made in heaven?
No but rackable + supermicro + sgi would be interesting.
___
Beowulf mailing list, Beowulf@beowulf.org
To change your subscription (digest mode or unsubscribe) visit
On Apr 2, 2009, at 11:10 PM, Vincent Diepeveen wrote:
[...]
At computerchess tournaments, end of 90s, i saw a representative of
Sun brag loud about Sun
machines, meanwhile he ran for his own software at a x86 pc.
[...]
I heard Sun engineers had to kick and scream before Sun management
Greg Lindahl wrote:
On Thu, Apr 02, 2009 at 09:05:26PM -0700, Ellis Wilson wrote:
Though entertainingly put, it would be an error to say ECC is a
requirement for everyone in a cluster group. I can think of more
than just a few purposes for clusters that truly do not require the
accuracy
On Sat, 4 Apr 2009, Joe Landman wrote:
Mark Hahn wrote:
Not every cluster FS we see is Lustre.
so what are the non-lustre scalable options for cluster FS?
Good performance:
--
GlusterFS
PVFS2
I have not been able to get even adequate performance out of GlusterFS,
which
On Sat, 4 Apr 2009, Andrew M.A. Cater wrote:
Interesting. Since I'm stuck using Red Hat and IBM, I've been hit by
this on a 10TB storage shelf. Red Hat will only offer me ext3 and 8TB.
IBM storage on a Megaraid card which handles the disks as one physical
volume
I'm using CentOS and the
On Sun, 5 Apr 2009, John Hearns wrote:
2009/4/5 Jason Riedy ja...@acm.org:
I'm more shocked that no one has written up using cfengine for
managing laptops. It seems a perfect model. With the more open
development model, perhaps it'll come back. But its competitors
are more web 2.0 cool.
On Sat, Apr 04, 2009 at 01:50:59PM -0500, Matt Lawrence wrote:
On Sat, 4 Apr 2009, Andrew M.A. Cater wrote:
Interesting. Since I'm stuck using Red Hat and IBM, I've been hit by
this on a 10TB storage shelf. Red Hat will only offer me ext3 and 8TB.
IBM storage on a Megaraid card which handles
On Sat, Apr 04, 2009 at 12:19:02PM -0700, Greg Lindahl wrote:
Exactly like shared-bus multiprocessors.
The incremental method of solving this is what Opteron/Nehalem does.
None of these had to deal with hundreds or thousands of cores in a
single socket yet (arguably GPGPU is pushing the
On Sun, Apr 05, 2009 at 08:49:22PM +0200, Eugen Leitl wrote:
On Sat, Apr 04, 2009 at 12:19:02PM -0700, Greg Lindahl wrote:
Exactly like shared-bus multiprocessors.
The incremental method of solving this is what Opteron/Nehalem does.
None of these had to deal with hundreds or thousands
Definitely, it's a non-cluster question, but may be
I'll get an answer: is there any remarkable difference
in FPU performance between Extreme and non-Extreme
Intel CPUs, e.g. QX9650 and Q9650?
of course not - the chips are the same inside, so deliver
exactly the performance you expect from
On 5 Apr 2009, at 4:00 pm, Jason Riedy wrote:
A similar situation exists in the node management space, where
existing solutions like CFengine were pretty much ignored by HPC
people.
Ha! Cfengine was pretty much ignored by *everyone*, including
its author for quite some time. Promising (pun
On 5 Apr 2009, at 6:32 pm, Andrew M.A. Cater wrote:
On Sat, Apr 04, 2009 at 01:50:59PM -0500, Matt Lawrence wrote:
On Sat, 4 Apr 2009, Andrew M.A. Cater wrote:
Interesting. Since I'm stuck using Red Hat and IBM, I've been hit by
this on a 10TB storage shelf. Red Hat will only offer me ext3
On Sun, Apr 5, 2009 at 9:48 PM, Mark Hahn h...@mcmaster.ca wrote:
Definitely, it's a non-cluster question, but may be
I'll get an answer: is there any remarkable difference
in FPU performance between Extreme and non-Extreme
Intel CPUs, e.g. QX9650 and Q9650?
of course not - the chips are
On Sun, Apr 05, 2009 at 10:00:44PM +0100, Tim Cutts wrote:
It still is. Our machines run it once a day.
I'm curious, do you also run it at boot? The one thing my usual scheme
lacks is an easy boot-time should I rejoin the herd or not? check.
-- greg
I'll get an answer: is there any remarkable difference
in FPU performance between Extreme and non-Extreme
Intel CPUs, e.g. QX9650 and Q9650?
of course not - the chips are the same inside, so deliver
exactly the performance you expect from their operating clock.
same cache, same clock, same
DEADLINE EXTENDED TO APRIL 13, 2009
Call for SC09 Tutorial Proposals
Experts in high performance computing are invited to share their expertise
with the High Performance Computing (HPC) community by submitting proposals
for tutorials at the SC09 conference to be held in Portland, Oregon,
Hi:
I wanted to carry out a test of parallel installation of a molecular
dynamics program (amber) operating directly on the parallel computer
(the ssh daemon was stopped)
The command (as user)
make test /dev/null
asked to confirm the RSA key fingerprint, unable to establish the
authenticity of
Tim Cutts wrote:
But you know what really gets me is that so many companies we've dealt
with don't officially support Debian or Ubuntu, but when you talk to
:)
You might be talking to the wrong companies. Seriously ...
We support Ubuntu. We'll support Debian with updated kernels/apps
On Sun, Apr 05, 2009 at 09:24:46PM +0200, Francesco Pietra wrote:
I can't explain better the above /dev/null, just taken from amber's
manual.
It just means the makefile has some ssh commands in it that need the -n flag.
As for your problem, perhaps it is sshing to localhost? ssh treats that
Not sure if you've seen this yet, but Sun appears to have rejected the
IBM bid as being insufficient. The claim is that they wanted out of the
exclusivity bit, likely to try to get a better deal from someone else.
So it seems IBM yanked the deal.
Of course this is all hearsay, but it is on
On Sun, Apr 05, 2009 at 10:10:18PM +0100, Tim Cutts wrote:
But you know what really gets me is that so many companies we've dealt
with don't officially support Debian or Ubuntu, but when you talk to
their engineers, you discover that they do their actual development on
it, and port to
On 5 Apr 2009, at 11:24 pm, Greg Lindahl wrote:
On Sun, Apr 05, 2009 at 10:00:44PM +0100, Tim Cutts wrote:
It still is. Our machines run it once a day.
I'm curious, do you also run it at boot? The one thing my usual scheme
lacks is an easy boot-time should I rejoin the herd or not? check.
27 matches
Mail list logo