Heterogeneous GPGPU computing promises tremendous acceleration of applications. This programming workshop includes hands-on sessions, application examples and an introduction to CUDA, CAPS, cuBLAS, cuFFT, the Portland Group Fortran Compiler, pycuda, and R. The intended audience includes scientists which want to port their simulation software to GPGPUsas well as people interested in a short overview of the available programming techniques.

Please Note: this is a first draft of the programme which may change before the Workshop: :

Day 1: Intro@CUDA and hands-on;  CAPS HMPP and PGI Accelerator Compiler

Day 2: CUBLAS & CUFFT + Hands-on; Scripting CUDA

The National High-End System HLRB II provides nearly ten thousand cores and the Linux cluster systems at LRZ more than 4,000 cores to scientists and students. Furthermore, powerful remote visualization facilities are available. Based on hands-on examples, an easy-to-follow introduction to basic Linux usage, specific information on the hard- and software of the LRZ cluster systems, the visualization systems and usage of the grid middleware (Globus Toolkit) is given. Grid certificates can be provided if needed.

Dr. M. Allalen, Dr. F. Jamitzky, Dr. H. Satzger, Dr. H. Heller

Please register via the LRZ registration form (please choose course HPCI1W11)

The course gives an introduction into the simulation of molecules based on several software packages on the supercomputers at LRZ Garching. (Maestro, Desmond, VMD, NAMD, Gromacs) This also includes an introduction to the remote visualization services at LRZ as well as hands-on sessions. The course focuses on biomolecules and targets the life science community. Hands-on sessions featuring example applications are given.

Program:

Day 1: Gromacs (ScalaLife) and Dalton (ScalaLife)

Day 2: Intro@MD, CPMD hands-on, Remote visualization services, VMD/NAMD (LRZ)

Day 3: Advanced feature of the graphical user interface Maestro (Schrodinger). This information is also available on Schrodinger web server: http://www.schrodinger.com/events/254/1

Primary drivers for the increased focus on parallel computing are new hardware trends (multi-core), larger data sets, and increased computational requirements stemming from more sophisticated methodologies. This course demonstrates the efficient use of R in parallel computing. In the beginning advanced R programming skills
(vectorization, apply functions, profiling) will be repeated. After a short theoretical course for parallel computing the parallel program design, methods and techniques for parallel computing, and parallel thinking are communicated with several examples. Main part of the course is the practical application of the R packages snow, multicore and foreach. Nevertheless, there are exercises for using batch systems and other R packages for parallel computing (snowfall, NWS). In the course the computer resources - including the HLRB2 with 9728 processors - from the Leibniz Rechenzentrum (LRZ, www.lrz-muenchen.de) can be used. The
participants are guided to use these resources and how to write applications for getting resources at the LRZ. The course extends the content of the publication "State-of-the-Art in Parallel Computing with R", Schmidberger, et.al. published in 2009 at the JSS journal.

Basic knowledge in using Linux systems (a short course can be found at http://www.linux.org/lessons/beginner/)

Note: While the bulk of the system will become available in summer 2012, LRZ is already operating a migration system with a similar architecture.

• Introduction to the x86 Westmere and Sandy Bridge Processor Architecture
• Usage of Intel Compilers: switches, optimization, profiling, OpenMP, advanced features
• Usage of the Intel debugger
• Explicit and compiler-directed vectorization with SSE and AVX

• Programming for NUMA
• Intel Cilk Plus
• Fortran standard support and extensions in Intel Fortran
• Intel Inspector XE (correctness checking for memory and threading)
• Threading Building Blocks, OpenCL and other parallel programming models
• Static Security Analysis
• Performance Libraries (MKL, IPP)
• VTune Amplifier XE (performance analysis)

• Intel Performance Monitoring Unit (PMU) and other performance analysis tools
• Intel Cluster Tools Overview and Usage
• Intel Many Integrated Core (MIC) Architecture
• A case study illustrating the use of the Intel toolchain

Thursday schedule:

• Using LoadLeveler for batch queuing
• IBM's parallel environment
• Performance Tools (High Performance Toolkit, Eclipse User Interface) including a demonstration.

On day 1, an introduction to Totalview including memory debugging and exercises is provided

On day 2, CUDA debugging and the Replay Engine are covered. In the afternoon, the Threadspotter tool for identifying performance issues is presented

To consolidate the lecture material, each day's approximately 4 hours of lecture are complemented by 3 hours of hands-on sessions.

The scalalife winterschool on molecular modelling focuses on the scalalife software packages (www.scalalife.eu) and aims to give an introduction to students in the field of life sciences and HPC. The developers of the software packages Gromacs, Dalton and Discrete give lectures and hands-on tutorials. Additional software packages used in the life scinece community are also explained in lectures and tutorials.

Day 1: HPC resources for life sciences in the scalalife project

Day 2: Gromacs

Day 3: Dalton

Day 4: Discrete

Day 5: Additional software packages

Weekend: Skiing trip to the Alps.

Participants should have knowledge in programming and life science software.

RRZE building, University campus Erlangen, Martensstr. 1

LRZ Building, University campus Garching, near Munich. (via video conference if there is sufficient interest)

Day 1

• Introduction to High Performance Computing
• HPC systems at LRZ and in Germany
• Technical aspects of software engineering: development process, usage of libraries, memory management, tools for code documentation, development environments
• Tools: Using Secure Shell, User interfaces, Source code versioning systems and Make

Day 2

• Basic parallel programming with MPI and OpenMP
• Processor Architectures: Register, Cache, Locality, Performance metrics

Day 3

• Principles of code optimization (unrolling, blocking, dependencies, C++ issues, bandwidth issues, performance projections)
• Using performance libraries
• Advanced MPI programming

Day 4

• Parallel algorithms: data parallelism, domain decomposition, task parallelism, master-worker, granularity, load balancing, scalability models
• Advanced OpenMP programming
• Parallel Architectures: multi-core, multi-socket, ccNUMA, cache coherence and affinity, tools for handling memory affinity

Day 5

• Architecture-specific optimization strategies: compiler switches, avoiding cache thrashing, exploiting SIMD capabilities
• MPI exercises

Each day is comprised of approximately 5 hours of lectures and 2 hours of hands-on sessions.

 Day 1

• Intel Tracing Tools: MPI tracing and correctness checking
• Intel Threading tools for OpenMP correctness checking and profiling
• Profiling on SGI Altix systems: histx and lipfpm
• Introduction to Scalasca

 Day 2

• Parallel application performance analysis with Scalasca
• Parallel input/output with MPI-IO

 Day 3

• I/O tuning on high performance file systems
• Portability of binary files, big/little endian issues
• Using I/O libraries (pNetCDF, HDF5)

Day 4

• Introduction to the PGAS languages Coarray Fortran and UPC: Partitioned global address space languages have emerged as an alternative to other parallel programming models, promising a shorter development cycle due to improved programmability while keeping the performance level on par with MPI. This course introduces the parallel  facilities integrated into the Fortran language (coarrays) and the C language (unified parallel C), respectively. A hands-on session allows to experiment with the new concepts, using prototype implementations on LRZ's HPC systems.

• Introduction to Eclipse IDE
• Introduction to CDT
• Hands-on with CDT
• Short introduction and demo of Photran

Different modern Krylov Subspace Methods (CG, GMRES, BiCGSTAB ...) as well as highly efficient preconditioning techniques are presented in the context of real life applications.

Hands-on sessions (in C and Fortran) will allow users to immediately test and understand the

• basic constructs of iterative solvers
• the Message Passing Interface (MPI)
• the shared memory directives of OpenMP.

This course is organized by the University of Kassel, the high performance computing centre of Stuttgart (HLRS) and IAG.

• object oriented features, submodules
• design patterns
• generation and handling of shared libraries
• mixed language programming
• standardized IEEE arithmetic
• I/O extensions from Fortran 2003
• parallel programming with coarrays
• source code versioning system (subversion)

To consolidate the lecture material, each day's approximately 4 hours of lecture are complemented by 3 hours of hands-on sessions.