Event Detail

Registration

Sorry, the registration period for this event is over.

Parallel Programming Workshop

 

Instructor

Rolf Rabenseifner, HLRS

Neil Stringfellow, CSCS

Registration fee

This fee covers administrative charges, all coffee breaks as well as slides, CD and manuals provided during the course.

 Academia

·      Employees/ students from educational/ research institutions: CHF 150.00

·      Attendees from industry: CHF 1500.00

Technical requirements

All participants must bring their own laptop to work on. Required software must be pre-installed before the start of the course. Please see below for software requirements.

 

Software Requirements

 

Microsoft Windows (XP/2003/Vista)

X Window Server available from Sourceforge website.

Please install the following three executables:

  Xming-6-9-0-31-setup

  Xming-fonts-7-4-0-3-setup

  Xming-mesa-6-9-0-31-setup

You should also get a PuTTY-SSH client preinstalled.

Linux

Ensure you have an SSH distribution installed and have access to secure shell, secure copy and ftp commands:

  which ssh

  which scp

  which sftp

If this is not the case download OpenSSH from the following website.

Check you are running an X Window System (X11) with an OpenGL implementation that supports 2D and 3D graphic applications: 

  who -r

  glxinfo | grep -1 opengl

The output of the first command for the current runlevel should return the value "5". 

Test it running the basic OpenGL Gears 3D application provided by default to any OpenGL implementation:

  glxgears

A window will open with a 3 colored gears rotating all-together. 

See http://www.x.org/wiki/ for downloading/ configuring/ installing X.org. 

Mac OS X (10.5 or later)

You will find the "X11" application under Applications/Utilities/X11

ssh should be available by default.

Program

                        Day 1

Message Passing Interface (MPI-1, part 1)

8:30

Registration

9:00

Introduction

9:10

Parallel Architectures and Programming Models (2)

10:10

Coffee break

10:25

MPI-1 - Introduction to the Message Passing Interface (3+3a)

10:25

Chap. 1 MPI Overview (talk)

10:45

Chap. 2 MPI Process model (talk + practical)

12:00

Lunch break

13:00

Chap. 3 Messages and Point-to-Point Communication (talk + practical)

14:15

Coffee break

14:30

Chap. 4 Non-Blocking Communication (talk + practical)

15:45

Coffee break

16:00

Chap. 5 Derived Data types (talk + practical)

17:00

Parallel debugging (5) (talk)

17:30

End

 

                        Day 2

Message Passing Interface (MPI-1, part 1)

8:30

Introduction to CSCS Computing Resources (lecture by CSCS)

9:00

Chap. 6 Virtual Topologies (3, continued) (talk + practical)

10:10

Coffee break

10:25

Chap. 7 Collective Communication (talk + practical)

11:00

Coffee break

11:15

Chap. 8 Other MPI-1 features (talk)

11:40

Heat conduction program, a parallelization example with MPI (6) (talk)

12:00

Lunch break

13:00

OpenMP - Overview and execution model (7+7a) (talk + practical)

14:00

Coffee break

14:15

OpenMP - Work sharing directives (talk + practical)

15:15

OpenMP - Data environment (talk + practical)

16:00

Coffee break

16:15

OpenMP - Summary and Pitfalls (talk)

17:15

Assure/ThreadChecker - detection of OpenMP race conditions (8) (talk)

18:15

End

 

                        Day 3

MPI-2, PETSc and VAMPIR

8:30

MPI-2 overview (10) (talk)

8:45

MPI-2 one-sided comm. (12) (talk)

9:15

Parallel file I/O (basics) (11) (talk + practical)

10:15

Coffee break

10:30

Parallel file I/O (fileviews) (talk + practical)

11:30

Parallel file I/O (access methods) (talk)

12:00

Lunch break

13:00

Parallel programming models on hybrid systems / MPI+OpenMP (23) (talk)

14:00

Coffee break

14:15

PETSc Tutorial (41) (talk)

15:00

Laplace-Example with PETSc (42a+c)

15:45

Performance analysis at CSCS (lecture by CSCS)

16:30

End


Back to listing