Personal tools
You are here: Home User Support Training Last Training Events

Last Training Events

 

DEISA Training Courses: 14 - 16 September 2010 at EPCC, Edinburgh

Introduction to the DEISA Infrastructure

Performance Optimisation on Cray XT Systems

Parallel Programming with Coarray Fortran


DEISA is running a series of three training courses at EPCC, The University of Edinburgh, in September 2010. All the courses will be based around a series of lectures with associated practical programming exercises.

Courses are free to all EU academics, and you can register for all the courses or just for selected days. Note that the first course is primarily aimed at DEISA users with accepted DECI-6 projects; the other courses are of general interest to all HPC users.

Registration form: http://www.epcc.ed.ac.uk/training-education/course-registration-form/
Location of EPCC:
http://www.epcc.ed.ac.uk/contact-us/maps/

Schedule of the Courses

Tuesday 14th September: Introduction to the DEISA Infrastructure

This course will cover the basic aspects of the DEISA distributed supercomputer environment and the software tools that are used to access it. It is primarily aimed at DEISA users who have had projects accepted under DECI-6, and will give them an opportunity to log on to their execution platform(s) and start porting their applications.
DEISA staff will be on hand to help with installing DEISA tools, accessing machines, compiling and executing jobs.

09:30 Overview of DEISA project and HPC systems
10:00 Basic DEISA software stack
10:30 DEISA certificates
11:00 COFFEE
11:30 Use of gsissh and gridftp
13:00 LUNCH
14:00 Introduction to UNICORE
15:00 Practical session (1)
15:30 TEA
16:00 Practical session (2)
17:30 CLOSE

 

Wednesday 15th September: Performance Optimisation on Cray XT Systems

Several of the DEISA systems are Cray XT machines, including the UK national supercomputer HECToR hosted by EPCC. This course will cover all aspects of achieving optimal performance on the Cray XT architecture, including single CPU, shared-memory multicore and inter-node message-passing issues.
It will also cover the use of Cray tools to perform detailed analyses of runtime behaviour.
This course is delivered in collaboration with Cray staff from the Cray Centre of Excellence based at EPCC. Access to Cray XT systems will be provided for all attendees.

09:30 Introduction to the Cray XT Programming Environment
11:00 COFFEE
11:30 Cray MPI Optimisation
12:30 LUNCH
13:30 Cray Performance Tools
14:30 TEA
15:00 Practical Session
17:30 CLOSE

 

Thursday 16th September: Parallel Programming with Coarray Fortran

Coarrays are a feature of the new Fortran 2008 standard that enable parallelism using a small number of additional language elements. A new array declaration syntax allows for remotely accessible variables, with data allocated across multiple "images". The execution model is that of a Partitioned Global Address Space (PGAS) language.
Programming using coarrays has many potential advantages compared to using MPI. Amongst these are simplicity, compiler checking and scope for automatic optimisation of communications by the compiler. Coarrays can also be introduced incrementally to existing MPI codes to improve performance-critical kernels. This course introduces coarrays and covers the related Fortran language features that support them.
This course is delivered in collaboration with Cray staff from the Cray European Exascale Research Initiative based at EPCC. Access to Cray systems and compilers will be provided for all attendees, although practical exercises may also be undertaken on a personal multicore laptop using the g95 compiler.

09:30 Coarrays and the PGAS model
10:00 Practical: Logging on and compiling
10:30 Basic coarray syntax
11:15 COFFEE
11:45 Further coarray features
12:15 Practical: Image processing
13:00 LUNCH
14:00 Advanced features
14:30 Practical: Image processing (cont)
15:30 TEA
16:00 Comparison to UPC
16:30 User experiences
17:30 CLOSE

 

For further details and to register see:
http://www.deisa.eu/usersupport/training/training-events/
The registration form is available directly at:
http://www.epcc.ed.ac.uk/training-education/course-registration-form/
For the location of EPCC see:
http://www.epcc.ed.ac.uk/contact-us/maps/

5 - 6 May 2010, University College, London, UK

  • Introduction to High Performance Computing(day 1).

  • Introduction to the DEISA Infrastructure (day 2).

DEISA is running two training courses at University College, London, in early May 2010. Both will be based around a number of practical programming exercises. No prior knowledge is assumed for either of the courses.

The first course on Wednesday 5th May is an "Introduction to High Performance Computing". It will cover the fundamentals of modern HPC architectures and the two major parallel programming models: shared variables and message passing. Practical sessions will involve running existing parallel programs to investigate issues such as performance and scalability.

The second course on Thursday 6th May is an "Introduction to the DEISA Infrastructure". This will cover the basic aspects of the DEISA distributed supercomputer environment and the software tools that are used to access it, including the Application Hosting Environment (AHE). Practical sessions will involve installing software on the desktop and using it to access the DEISA systems.

Courses are available free for academic attendees. If the courses become over-subscribed, preference will be given to members of the Virtual Physiological Human Network of Excellence.

 

Those attending are encouraged to use their own laptops for both courses.

To register, please fill in the form at http://www.epcc.ed.ac.uk/training-education/course-registration-form/
Registration closes on Wed April 28th

Accomodation

see pdf file with detailed hotel information

 

Location:

Room LG27, Chemistry Department, University College of London, 20 Gordon Street, London WC1H 0AJ, UK

Travel Guide: http://ccs.chem.ucl.ac.uk/directions/

Programme:

Day 1: Introduction to High Performance Computing (Wed 5th May 2010)

10:00 Introduction to Parallel Computing
11:00 COFFEE
11:30 Image processing Practical
12:15 HPC Hardware
12:45 LUNCH
13:45 HPC Architectures
14:30 Message-Passing Concepts
15:00 Practical: Traffic (MP)
15:30 Shared-Variables Concepts
16:00 TEA
16:30 Practical: Traffic (SV)
17:00 Comparison of models
17:30 CLOSE

Day 2: Introduction to the DEISA Infrastructure (Thu 6th May 2010)

10:00 Overview of DEISA project
10:20 DEISA software
11:00 COFFEE
11:20 Use of Certificates
11:45 Use of Globus services
13:00 Lunch
14:00 Introduction to AHE
15:15 AHE practical
15:30 TEA
16:00 AHE practical (cont.)
17:00 CLOSE



Training provided by DEISA partners

In addition to the DEISA-specific courses, many of the DEISA partners run their own training in areas relevant to HPC and Computational Science. Please see the links below for more details.

LRZ http://www.lrz-muenchen.de/services/compute/courses/
CSC http://www.csc.fi/english/csc/courses/
FZJ http://www.fz-juelich.de/jsc/courses/
IDRIS https://cours.idris.fr/
EPCC http://www.epcc.ed.ac.uk/training-education/
HLRS http://www.hlrs.de/no_cache/organization/sos/par/services/training/

 

Recent Training Courses

3 - 4 November 2009, SARA, Amsterdam

  • Introduction to the DEISA Infrastructure.

  • The Scalasca Performance Analysis Toolset.

DEISA is running two training courses at the SARA Computing and Network Centre, Amsterdam, in November 2009. Both courses will be based around a number of practical programming exercises. Registration is closed

The first course on Tuesday 3rd November is an Introduction to the DEISA Infrastructure. This will cover the basic aspects of the DEISA distributed supercomputer environment and the software tools that are used to access it.

The timetable is:

09:00 Overview of DEISA project and HPC systems
09:30 Basic DEISA software stack
10:00 Introduction to UNICORE
10:30 Coffee break
11:00 UNICORE practical session
12:30 Lunch break
13:30 Introduction to DESHL
14:00 DESHL practical session
15:00 Coffee break
15:30 Overview of Globus services
16:15 Globus practical
17:30 Close

No prior knowledge is assumed for this course.

The second course on Wednesday 4th November will cover The Scalasca Performance Analysis Toolset.

Scalasca is an open-source, portable toolset that can be used to analyse the behaviour of parallel applications and to identify opportunities for performance optimisation. It has been specifically designed for use on large-scale HPC systems such as those in DEISA, but is also well-suited for small and medium-scale HPC platforms. Scalasca supports an incremental performance-analysis procedure that integrates runtime summaries with in-depth profiles of parallel execution. DEISA staff will be on hand to help users with the Scalasca practical exercises and to help analyse the execution of their own programs on the DEISA platforms.

The timetable is:

09:00 Introduction to parallel performance analysis
10:00 Introduction to the Scalasca toolset
10:30 Coffee break
11:00 Hands-on practical: using Scalasca with provided example application
12:00 Scalasca usage case studies
12:30 Lunch break
13:30 Hands-on practical: analysing user applications
15:00 Coffee break
15:30 Hands-on practical: analysing user applications (continued)
17:30 Close

We assume that those attending this course will already have their own parallel application that currently runs on some other HPC system.

Please note that registration is closed

 

13 - 14 January 2009: FZJ, Jülich, Germany

 

DEISA is running two training courses at the Jülich Supercomputer Centre in January 2009. Both courses will be based around a number of practical programming exercises.

 

  • The first course on Tuesday 13th January is an "Introduction to the DEISA Infrastructure". This will cover the basic aspects of the DEISA distributed supercomputer environment and the software tools that are used to access it.
    No prior knowledge is assumed for this course.
    Timetable:
    09:00 Overview of DEISA project and HPC systems
    09:30 Basic DEISA software stack
    10:00 Introduction to UNICORE
    10:30 COFFEE
    11:00 UNICORE practical session
    12:30 LUNCH
    13:30 Introduction to DESHL
    14:00 DESHL practical session
    15:00 COFFEE
    15:30 Overview of Globus services
    16:15 Globus practical (or continue UNICORE and DESHL practicals)
    17:30 CLOSE
  • The second course on Wednesday 14th January is an "Introduction to the Blue Gene P" which is aimed at users planning to port their codes to this new architecture. The course will introduce the basic features of the Blue Gene design and cover the details of the three Blue Gene machines within the DEISA infrastructure. People on the course will be given access to the Blue Gene P at FZJ, JUGENE, which is the fastest supercomputer in Europe. DEISA staff will be on hand to help users to port their codes and perform test runs.
    We assume that those attending this course will already have a parallel application that has previously been run on some other system.
    Timetable:
    09:00 Introduction to the Blue Gene architecture
    10:00 The IDRIS Blue Gene system
    10:30 COFFEE
    11:00 The FZJ JUGENE System
    11:30 Basic practical session (log on, compile, submit a test job)
    12:30 LUNCH
    13:30 Performance Tools
    14:00 Hands-on practical: porting user applications
    15:00 COFFEE
    15:30 Hands-on practical: porting user applications (continued)
    17:00 CLOSE

For more details and an application form (including accommodation in Jülich), please see this URL.

 

17 - 18 September 2008: Portici, Naples

 

DEISA is running two training courses at the ENEA Research Centre in Portici near Naples, Italy. Both courses will be based around a number of practical sessions for which users must supply their own wireless laptop.

 

  • The first course on Wednesday 17 September is an "Introduction to the DEISA Infrastructure". This will cover the basic aspects of the DEISA distributed supercomputer environment and the software tools that are used to access it.
    No prior knowledge is assumed for this course.
  • The second course on Thursday 18th September is on "Parallel IO using MPI-IO". This covers the fundamental concepts and usage of the parallel IO functions contained in the Message-Passing Interface standard. This area is becoming increasingly important as scientific datasets increase in size, and reading from and writing to disk starts to become a serious bottleneck on large parallel machines.
    The course assumes that you already have basic practical knowledge of the MPI programming model and that you are able to write simple programs using basic MPI functions such as send, receive and reduce.

For more details and a registration form, please see this URL.

 

Previous Training Events

Training Archive (2005-2008)

Document Actions