Each student in the course will have access to 3 resources: Stampede, Bang and the CSEClass GPU workstations. These platforms should be used for course work only.
The programming environments on each resource have been set up assuming that you have defined the PUB environment variable to point at a certain folder on each respective machine. In particular, PUB should be defined in the file .bashrc, which is is located in your home directory, as follows
| Machine | Command |
|---|---|
| Stampede | export PUB=/home1/00660/tg458182/cse262-wi15 |
| CSEClass | export PUB=/class/public/cse262-wi15 |
| Bang | export PUB=/share/class/public/cse262-wi15 (environment not yet set up) |
Stampede, which is located at TACC, The Texas Advanced Computing Center, is a large mainframe containing two kinds of processing nodes: Nvidia Kepler GPUs and Intel Phi manycore processors. Here is a brief guide to Getting Started with Stampede
There are 5 CSEClass GPU workstations numbered CSEClass03 through CSEClass07. Your login ID and password will be the same as your UCSD network email account name, that is, your AD (Active Directory) credentials. Here is documentation for the CSEClass machines.
Your login ID and password will be the same as your UCSD AD credentials. Here are instructions for using Bang.
If you are running a program with a graphical interface (i.e. gnuplot, or the CUDA profiler), connect to the remote machine from a Linux platform using the -X option of ssh:
$ ssh -X yourlogin@ccom-bang-login.ucsd.edu
If you are connecting from a Windows platform, and are using a ‘secure shell client’ AND ‘exceed’ (or xwin32, or cygwin) to log into a remote machine you must ensure that ‘X11 tunneling’ is set. To do this, you must have installed Cygwin /X with X11. Remote visual display can be slow when off campus.
Here is a description of the memory hierarchy of the multicore processors used in the course. This list will continue to grow Here is how to Explore your processor’s attributes.
Here are some on-line resources for documentation, tutorial, reference materials and course materials.
Introduction to Parallel Computing, Blaise Barney. A comprehensive, terse set of notes covering a wide range of material. You can use this document to chart your progress in learning about parallel computation.
LLNL’s High Performance Computing Training and Workshops. A collection of matierals covering a wide range of topics.
Colorado School of Mines High Peformance Computing Workshop Many code examples + slides.
[SSE] (coming)
Numerical Recipes in C or Fortran, 2nd Ed., by Press et al., Cambridge University Press.
A multigrid tutorial, slides by William Briggs.
UC Berkeley Applications of Parallel Computers, taught by Jim Demmel and others. See the links to course resources including web pages from prior years. Jim Demmel wrote extensive notes in spring 1996 which are a valuable source of information.
How to write Fast Numerical Code, Markus Püschel.
Cambridge University course notes. A wide range of subjects in addition to HPC.
Algorithms in the Real World, Guy Blelloch.
IBM Red books. Some handy references, with discussions of general interest in parallel and high performance computing. A Search portal.
Performance Tuning for the Origin 2000 Though written for the Origin 2000, this material remains an excellent source of information about performance tuning, especially for shared memory computers.
Models of Parallel Computation. SGI Publication Topics in Parallel Computation
Topics in IRIX Programming (Long SGI Manual, PDF, esp. Ch. 11 on pthreads)
Netlib repository of numerical software and related reports. (If you clone this window and click here, you will get a one-page listing of all the libraries available within netlib.)
U. Florida. Sparse matrix algorithm research, matrix collection and software (U. Florida) algorithm research, matrix collection and software.
NIST. Matrix market
Scientific Computing: An Introductory Survey, 2nd Ed., by Michael Heath, McGraw-Hill, 2002. Includes links and software modules accessible via applets.
Introduction to Linear Algebra, 4th Ed., by Gilbert Strang.