OPECG-2009

OPECG 2009 Mode 1 Hands-on Session:
Progamming on Multi-Core Processors
  • Multi-Core Processors – Architecture & Programming Environment
  • Performance Enhancement through Software Multi-threading on Multi-Core Processors
  • Programming using Intel Threading Building Blocks (TBB) on Multi-Core Processors
  • Shared Memory Programming (Pthreads & OpenMP)
  • Message Passing Programming (MPI-1 & MPI-2)
  • Data Parallel Programming (f90/f95) Issues on Multi-Core processors
  • Complier Optimisations on Multi-Core processors
  • Multi-threaded Programming using Memory Allocators
  • Performance Metrics - Measure Performance on Multi-Core Processors
  • Multi-Core Programing Software tools (Thread Checker, Thread Performance Analyzer)
  • System and Application Benchmarks on Multi-Core Processors
  • Tuning & Performance of Numerical and Non-Numerical Computations
  • Mixed mode of Programming (MPI-OpenMP, MPI-Pthreads, MPI-TBB MPi-f90) & Inter Operable Issues on Multi-Core processors
  • An Overview of Hybrid Adaptive Cluster Computing-Programming Paradigms
  • Special Class of Application Kernels, and Numerical Libraries on Multi Core Processors,
OPECG 2009 Mode 2 Hands-on Session
Progamming on GPGPUs /GPU Computing & Hybrid Computing
(CUDA enabled NVIDIA GPUs;   AMD-ATI Stream Computing SDK )
  • An Overview of GPU Computing - NVIDIA CUDA Hardware a& Software Capabilities
  • Programming based on NVIIDA CUDA Software Development Kit, and CUDA APIs
  • GPGPUs - Stream Programming (AMD) : Stream Programming Architectcure; AMD-ATI SDK; Brook+ Programming
  • Examples - GPGPUs - AMD Stream computing Software Development Kit (SDK), Brook+ Programming
  • Mixed mode of Progrmaming : Programming using Intel Thread Building Blocks (TBB) & GPU Computing - NVIDIA CUDA Programming
  • Numerical and Non-Numerical Computations on Multi-core Systems with GPU accelerators using different Programming Paradigms
  • Mixed mode of Progrmaming : Programming using Intel Thread Building Blocks (TBB) & GPU Computing - NVIDIA CUDA Programming
  • Selected Application Kernels in Scientific Computations on Multi-Core Processors & GPUs