Amazon’s Web Services team is offering grants for researchers who want to explore the use of cloud computing.  It provides AWS credits for the researcher to use on their services:  Elastic Compute Cloud (EC2), Elastic Map-Reduce, Simple Storage Service (S3), SimpleDB, and others.  For research computing use-cases, Amazon offers a variety of options:  high-memory instances with 17GB to 68GB of memory, “high-CPU” instances with 2GB to 7GB of memory, and their new 23GB/high-performance computing instance.

The application process is extremely simple — a couple of check-boxes on what services you’re interested in, and a 4000 character limit on the description.  To apply, go to the following URL and click on the “Research” tab:

• http://aws.amazon.com/education/

SCSC staff are available if you need assistance with your application.

Following the suggestions made at our last DSCR information session, we have changed how high-priority queues are accessed on the DSCR — note that you still have the same access you had previously, we hope that this change will make it easier to submit jobs to ANY high-priority queue that you have access to.

Previously, if you were a member of multiple groups, you would have to issue a ‘newgrp’ command before launching jobs onto a second group of high-priority machines — without ‘newgrp’, all of your jobs would land on your “home” group’s machines.

You no longer need to issue the ‘newgrp’ command — a job submitted as high-priority will now run on ANY machine in ANY group that you have access to.

If you need or want to run your jobs on specific machines, you may still do so.  See the following page for more information:

• https://wiki.duke.edu/display/SCSC/SGE+Multiple+Groups

We hope this configuration change makes it easier for users to access larger groups of machines when there are deadlines in your work.  As always, if you do not have a specific target-date in mind, low-priority jobs have access to all 720 machines in the DSCR — a lot of compute-power!