ARSC HPC Users' Newsletter 345, August 04, 2006

ARSC Announces New System Procurement

More details will be available later, but we wanted to let ARSC users and other newsletter subscribers know:

The Arctic Region Supercomputing Center will be installing a Sun Microsystems high performance computing system running SuSE Linux. The system will consist of 952 AMD Opteron dual core 2.6 GHz processors (1904 total cores), yielding a theoretical system peak floating point performance of 9.88 teraflops. Four gigabytes of memory will be available per core. The compute nodes and the disk storage servers will be connected by a high performance Infiniband switch. In addition to this compute capability, Sun's recently-introduced x4500 disk technology will provide 144 raw terabytes of disk storage. The x4500 systems will use the Lustre file system to deliver high performance shared storage.

We anticipate installation of the Sun system this fall.

Baseline Configuration

The Baseline Configuration (BC) Project of the HPCMP has been working to define consistent working environments across resources of the Modernization Program. The intent of this project is to allow users to compute productively and collaboratively when using the HPC resources at multiple centers.

Towards this end, ARSC has made several changes in the past few months on our allocated resources, iceberg and klondike.

show_usage

The show_usage command displays standard allocation information about your projects. This information includes:

  • Hours allocated to each project
  • Hours remaining for each project
  • Foreground and background use for each project
  • Percent of foreground allocation remaining for each project
  • Percent of the fiscal year remaining.

The "-s" option will show allocation information for all allocated ARSC systems.

common queue names

Both klondike and iceberg now accept the standard BC queue names as well as "historical" ARSC queue names.


  BC queue name   priority          
  --------------  ---------------------
  standard        Standard priority
  urgent          Urgent priority (requires special permission)
  high            High priority (requires special permission)
  debug           Debug priority
  background      Background priority
  challenge       Challenge type (requires challenge project status)

common environment variables

Several environment variables are now set:

  • ARCHIVE_HOST : the hostname for the system housing long term storage.
  • WORKDIR : the globally accessible temporary filesystem (same as WRKDIR )
  • ARCHIVE_HOME : the location of long term storage on ARCHIVE_HOST (same as ARCHIVE )
  • JAVA_HOME : the base directory where java is installed if available.
  • PET_HOME : the directory containing tools installed by the PET CE group.

2006 Faculty Camp Talks

ARSC Faculty Camp is an intensive series of seminars and hands-on experiences, presented by ARSC staff, UAF/ARSC Joint Faculty, and current users, which provides participants with assistance and expertise while they develop individual projects.

All ARSC users and potential users are invited to attend any lectures listed below.


Monday, 8/7
        10:00 AM        Performance Programming              WRRB 009
                          Tom Baring
         2:00 PM        Debugging                            WRRB 009
                          Ed Kornkven

Tuesday, 8/8
        10:00 AM        Parallel Programming with OpenMP     WRRB 009
                          Tom Logan
         2:00 PM        OpenMP Hands-On Session              WRRB 009
                          Tom Logan        

Wednesday, 8/9
        10:00 AM        Visualization of 3D Datasets         WRRB 009
                          Sergei Maurits
         2:00 PM        Visualization of 3D Datasets         WRRB 009
                          Sergei Maurits
        
Thursday, 8/10
        10:00 AM        Parallel Programming with MPI        WRRB 009
                          Tom Logan
         2:00 PM        MPI Hands-On Session                 WRRB 009        
                          Tom Logan

Friday, 8/11
        10:00 AM        Using netCDF                         WRRB 009
                          Kate Hedstrom

Monday, 8/14
        10:00 AM        IDL                                  WRRB 009
                          Sergei Maurits
         2:00 PM        IDL Hands-On                         WRRB 009
                          Sergei Maurits

Tuesday, 8/15
        10:00 AM        Introduction to Amira                WRRB 009
                          Sergei Maurits
         2:00 PM        Visualization Open Session           WRRB 009
                          Sergei Maurits
        
Wednesday, 8/16
         2:00 PM        ARSC Future Technologies             WRRB 009
                          Greg Newby

ARSC Science Seminars

Topics: "High-Level Languages (HLLs) for High-Performance Reconfigurable Computers (HPRCs): A Comparison Study on Cray-XD1" and "UPC/Multi-Core Benchmarking" When: Thursday August 10th, 1:00 - 2:30 PM Where: Location: WRRB 010 Seminar Description:

Presentations by George Washington University doctoral students in Electrical and Computer Engineering, highlighting their work during 10 weeks at ARSC this summer. Presenters are students of Dr. Tarek El-Ghazawi: Kun Xi, Yiyi Yao, Esam El-Araby, Mohamed Taher and Abdullah Kayi.

Contact: Greg Newby 907-450-8663

Quick-Tip Q & A


A:[[ I want to search for certain keywords in a monstrous text file.
  [[ But for each "hit," I want to see, not just the line containing 
  [[ word, but a few lines **preceding** it as well.  I think I could
  [[ the write a perl script to do this, but is there another way?

#
# Thanks to Rich Griswold, Liam Forbes, Matthew Page, Lee Higbie
#   and Jed Brown, all of whom know that GNU grep is a better answer.
#   Here's a blend of their answers:
#

The GNU grep utility has a couple of options that should work for you.

  '-A NUM' prints NUM lines after each group of matching lines
  '-B NUM' prints NUM lines before each group of matching lines
  '-C NUM' prints NUM lines before and after each group of matching 
           lines

--

So:

  $ grep -B5 "cow tipping" file.ext

would show you 5 lines before each match of "cow tipping" in the file.

--

  Especially with context matching, `--color=auto' is handy.

--

Editor's note: These options are only in GNU grep, which is available
on ARSC systems with the following caveats:

  --On the ARSC IBM systems, GNU's family of grep commands is available
    under /opt/freeware/bin/.
  --On the Cray X1, they're in the coreutils module (execute,
    "module load coreutils" to load this module).
  --GNU grep isn't available on the ARSC SGIs.


  

Q:[[ I'm looking for a file that I changed a few minutes ago. 
  [[ Unfortunately I seem to have forgotten the name.  Is there a easy 
  [[ way to search for any files in directory or subdirectory that have 
  [[ changed in the last 10 minutes?
      

[[ Answers, Questions, and Tips Graciously Accepted ]]


Current Editors:
Ed Kornkven ARSC HPC Specialist ph: 907-450-8669
Kate Hedstrom ARSC Oceanographic Specialist ph: 907-450-8678
Arctic Region Supercomputing Center
University of Alaska Fairbanks
PO Box 756020
Fairbanks AK 99775-6020
E-mail Subscriptions: Archives:
    Back issues of the ASCII e-mail edition of the ARSC T3D/T3E/HPC Users' Newsletter are available by request. Please contact the editors.
Back to Top