On 1/16/09 12:03 PM, "Robert G. Brown" wrote:
> On Fri, 16 Jan 2009, Greg Lindahl wrote:
>
>> On Thu, Jan 15, 2009 at 04:26:00PM -0500, Lechner, David A. wrote:
>>
>>> Sometimes the newer technology uses less power and is cheaper to
>>> operate(anyone ever create a KW/MFLOP vs. Time curve?
The question was raised as "When should all these servers be upgraded or replaced
again?"
3-5 years, IMO. if you replace hardware in <3 years, you're obviously
burning money. that's defensible sometimes, but always pretty dubious -
or else you need an inflated sense of self-importance like o
i added the mailing list to this since you did not hit reply to all and i
have been the only one getting the replies. i think that is not fair and you
should be allowed to contact the manufacturer directly. i did that with
corsair cuz of some fault ram and im rma ing the paried set that i have back
On Fri, 16 Jan 2009, Greg Lindahl wrote:
On Thu, Jan 15, 2009 at 04:26:00PM -0500, Lechner, David A. wrote:
Sometimes the newer technology uses less power and is cheaper to
operate(anyone ever create a KW/MFLOP vs. Time curve? has that
really gone down? )
I'd think that you would have l
stephen mulcahy wrote:
What I'm seeing in my research is that there are a bunch of switches
around the €600 mark (such as D-Link's DGS-1248T) and then a bunch of
switches around the €1300 mark (such as HP's Procurve 2810-48G). If you
A customer recently replaced the D-Link model switch wit
Hi,
Thanks to everyone for their responses to my KVM query - I need to
digest those a bit more but the general picture I'm getting is that KVMs
are expensive and don't add much to the overall management.
On a related note, I'm also looking at a gigabit switch for the same
cluster. I saw some
As has been mentioned before, having traditional KVMs can become rather
problematic when you have high node density or a lot of nodes. You have to put
those cables somewhere. We use terminal servers that connect to the serial
ports on the nodes and can be accessed over the network.
We are ha
On Thu, Jan 15, 2009 at 04:26:00PM -0500, Lechner, David A. wrote:
> Sometimes the newer technology uses less power and is cheaper to
> operate(anyone ever create a KW/MFLOP vs. Time curve? has that
> really gone down? )
I'd think that you would have looked that up before you posted! I
don't
We use D-Link DKVM-8E's and chain them together quite satisfactorily
to a single console in one room. It works okay for 16 nodes. I
don't know why it wouldn't work for 40, but there's no way to access
them remotely.
Mike
At 08:17 AM 1/16/2009, you wrote:
Hi,
Do people here normally use K
Note that single threaded performance doesn't say a thing,
because when just 1 core runs, nehalem automatically overclocks 1 core.
A very nasty feature.
My experience is that Shanghai scales 4.0 nearly versus nehalem 3.2,
because of the overclocking of 1 core.
So seeing a 9% higher IPC is not v
stephen mulcahy wrote:
Hi,
Do people here normally use KVMs for management of compute nodes or
forego those and either have a grad student with a keyboard+monitor on a
trolley or rely entirely on remote management cards and/or IPMI?
My problem with old fasion KVMs is that all those extra
cab
stephen mulcahy wrote:
Hi,
Do people here normally use KVMs for management of compute nodes or
forego those and either have a grad student with a keyboard+monitor on a
trolley or rely entirely on remote management cards and/or IPMI?
IPMI 2.0 with kvm over IP whenever possible. Crash cart
Speaking of Linux Mag, nice article by Douglas summarizing GPU computing
for HPC. But Doug, maybe have someone read it for typos?
Ralph Finch, P.E.
California Dept. of Water Resources
Delta Modeling Section, Bay-Delta Office
Room 215-13
1416 9th Street
Sacramento CA 95814
916-653-7552
rfi...@wat
Hello Again to all you Beowulf users!
Similar to Doug Edline's poll
I was a subscriber in the late 1990s, but had to sign off the list for several
years as my job changed.
I used to follow many of your posts with such interest, and hope you oldtimers
(and newer subscribers as well) are all d
Hi,
Do people here normally use KVMs for management of compute nodes or
forego those and either have a grad student with a keyboard+monitor on a
trolley or rely entirely on remote management cards and/or IPMI?
For those of you that use KVMs, what ones do you normally use?
I'm taking a look a
Hi folks:
Thought you might like to see this. I rewrote the interior loop for
our Riemann Zeta Function (rzf) example for SSE2, and ran it on a
Nehalem and on a Shanghai. This code is compute intensive. The inner
loop which had been written as this (some small hand optimization, loop
unr
All:
In keeping with our timely reporting pledge, ClusterMonkey
has just posted pictures form the BeoBash
http://www.clustermonkey.net//content/view/244/1/
Thanks again sponsors: AMD, NVidia, ClusterCorp,
Penguin/Sclyd Computing, Terascala, and Xand Marketing.
And while you are there, take th
i havent done much digging in the transistor count but from people i have
talked to in regards to the i7s the i7s are true quad cores im not sure what
that means for the core 2 quads that came before the i7.
On Fri, Jan 16, 2009 at 2:01 PM, Mark Hahn wrote:
> normally according to moores law new
normally according to moores law new technology is released every 18 months.
no: it observes that _transistor_count_ (at efficient cost)
doubles every _2_years_. the wikipedia page is very good.
however, ML largely ignores the fact the various measures
that display exponential improvement are
Hi all,
I have made a set of OFED 1.4 infiniband for debian sid. The packages contain
the utilities, libraries and kernel modules not currently packaged for debian.
ftp://ftp.sanger.ac.uk/pub/gmpc/repository/infiniband/
sources.list entry:
deb ftp://ftp.sanger.ac.uk/pub/gmpc/repository/infinib
20 matches
Mail list logo
