stephen mulcahy wrote:
I'd note that battery backed caches on the RAID cards I've looked at in
the last few months seem to come at a *significant* premium over the
base price of the card (50-100% type significant).
Hmmm those must be either really expensive batteries, or really
cheap car
On 23/11/2007, at 4:52 AM, Jeffrey B. Layton wrote:
A quick question about the list. Can I post job openings to the
list or
is that in bad taste?
I guess that depends on whether the list moderators:
(a) are looking for a job in the HPC area, or
(b) are not.
Personally, I don
First post in this group, even if I'm subscribed since a few months
As other people indicated, personally I use Matlab mainly becasue a lot of
toolboxes are missing or incomplete in Octave. It is a considerable effort to
migrate your code. I tried once, but I have too much code and the missing
Hi all
the November issue of the Communications of the ACM has an
nice paper of Basili and Zelkowitz where they report the following
data, from the paper
Hochstein, L., Carver, J., Shull, F., Asgari, A., Basili, V.,
Hollingsworth, J., and Zelkowitz, M. Parallel programmer productivity: A
Jakob Oestergaard wrote:
One thing that HW RAID has, though, that SW RAID fundamentally cannot get, is a
battery backed write cache. This can, for many workloads, prove to give a very
significant performance advantage (for sync heavy workloads you can get several
orders of magnitude speedup).
Joe Landman wrote:
RAID10 is not that hard in software, the hotswap is harder. Cold-swap
*may* be possible (last I looked it should work, haven't tried it
recently on SATA). The issue is whether or not the driver drags the
kernel screaming and kicking into a kernel panic when a device is
rem
Alan Louis Scheinine a écrit :
>
>There is a particular kind of application, single-client
> and serial process, for which a striped file system using
> RAM disk would be very useful. Consider reading small
> blocks at random locations on a hard disk. The latency
> of the HDD could be a few
Hi,
We had the same problems with a cluster of 40 nodes. The motherboard has
problems with great IO. We have some test programs they used only the cpu and
make no or less IO. These programmes runs and runs. But when you have a program
like Gaussian with a big IO then this can happen.
At last w
