Linax1.cims.nyu.edu will by offline briefly down due to an upgrade to RHEL6.  The scheduled outage is expected to last less than 5 minutes at approximately 6:30pm on Thursday May 3rd.  We urge you to test any software on a RHEL6 system (see compute servers for a list of available systems) to make sure it's compatible.

Feel free to use the access.cims.nyu.edu servers to connect from remote locations during the outage.

If you have problems or questions please contact helpdesk@cims.nyu.edu.

CIMS mail servers are now configured to use procmail as the local delivery agent (the program that is used for delivering messages to local mailboxes).   This means you can no longer use a ~/.forward file to control whether or not procmail is used for delivering messages to your mailbox.  Procmail is now always used for local mail delivery.  You can still use a .forward file to forward mail.  If you do not want to procmail to perform any mail processing whatsoever you should rename your ~/.procmailrc file to something else.  For example:
  mv ~/.procmailrc ~/.procmailrc.OFF

The following old versions of software are slated for removal at the end of the semester.  In preparation they are being moved preparation for their removal at the end of the semester.  If you rely on a specific version of software mentioned below, it will stop working.

• CUDA: 2.x, 3.x is being moved to /usr/local/pkg/cuda/old/ until deletion (current install is 4.x)
• Java: 1.6.x from /opt is being moved to /opt/pkg/java/old until deletion (current install is 1.7.x)
• Maple: 11, 12 is being moved to /opt/pkg/maple/old until deletion (current install is 14)
• Mathematica: 6.x, 7.x is being moved to /opt/pkg/mathematica/old until deletion (current install is 8.0)
• Matlab: R2008-R2010  is being moved to /opt/pkg/matlab/old until deletion (current install is R2011a)
• QT: 4.3.x is being moved to /usr/local/pkg/qt/old until deletion (current install is 4.6.x)
• Sage: 2.11 is being moved to /opt/pkg/sage/old until deletion (current install is 4.7)
• Tecplot: 2008-2009 is being moved to /opt/pkg/tecplot/old until deletion (current install is 2011)
• Visit: 2.1.x is being moved to /opt/pkg/visit/old until deletion (current install is 2.2.x)

If you rely on one of these software packages, you will need to update your software to use the newest version installed.  If you have problems with the current version installed, please contact us via email helpdesk@cims.nyu.edu. 

~~~~~~~~~~~~~~~~~~~~~~~~~~~
Speeding Up MATLAB Applications
~~~~~~~~~~~~~~~~~~~~~~~~~~~

Thursday, November 17

8:50 am -  10:50 am

New York University

Warren Weaver Hall, Room 102


Presenter: Brett Shoelson, Principal Application Engineer

-Agenda-

This seminar will review a few of the products included in the New York University
campus wide License of MATLAB and various Toolboxes.
Specifically, we will discuss and demonstrate simple ways to improve and optimize
your code that can boost execution speed by orders of magnitude.  We will also
address common pitfalls in writing m-code, explore the use of the MATLAB Profiler to
find bottlenecks, and briefly introduce our Parallel Computing Toolbox and
Distributed Computing Server to solve computationally and data-intensive problems on
multicore computers and clusters.

Highlights include:
*             Understand memory usage and vectorization in MATLAB
*             Address bottlenecks in your programs
*             Optimize file I/O to streamline your code
*             Transition from serial to parallel MATLAB programs

CIMS Community is a new forum for CIMS users to post and read items of interest.  It is restricted to those with a CIMS account.  We hope that it will prove a useful space for the CIMS community.  Please direct any feedback or feature requests to Helpdesk.

The Euler Cluster is back online after being upgraded to the the latest operating system.

MathWorks free webinar on October 14th at 2:00 p.m. EDT.

---------------------------------------------------
Parallel Computing with MATLAB
---------------------------------------------------

In this session, you will learn how to solve computationally and data-intensive problems using multicore processors, GPUs, and computer clusters. We will introduce you to high-level programming constructs that allow you to parallelize MATLAB applications without CUDA or MPI programming and run them on multiple processors. We will show you how to overcome the memory limits of your desktop computer and solve problems that require manipulating very large matrices by distributing your data. We will also illustrate how you can run the same application on a single machine using the Parallel Computing Toolbox and on a large scale computing resource such as a cluster, using the MATLAB Distributed Computing Server.

This session will cover:

• - Toolboxes with built-in support for parallel computing
• - Creating parallel applications to speed up independent tasks
• - Programming with distributed arrays to work with large data sets
• - Scaling up to computer clusters, grid environments or clouds

Register for this webinar.

Crunchy3.cims.nyu.edu is currently down due to reconfiguration/reinstall.  You can get a full list of available systems from the Compute Grid Status page.

Feel free to use any of the available systems.

CIMS compute server users can now check real time server status such as CPU, load, memory, disk, and network usage from our webpage.  Before running compute jobs, login and check each grid's status to see where best to run your jobs.  Links are provided on the compute servers webpage.

With the switch to RHEL6 our Linux computers are now compatible with Dropbox and due to high demand it is now available on all our Linux Desktops.  Simply run 'dropbox' from a command line interface.

Crunchy3.cims.nyu.edu is back online.

Crunchy3.cims.nyu.edu is temporarily offline due to a hardware problem.  We'll notify you when it is back online.