This is First in Four Part Tutorial built loosely around Ian Foster's Book
|
This Introduction covers
|
Later talks cover
|
001 Designing and Building Parallel Programs 002 Outline 003 Outline 004 Why Parallel Computing? 005 The Multicomputer: an Idealized Parallel Computer 006 Multicomputer Architecture 007 Multicomputer Cost Model 008 How do Real Parallel Computers Fit the Model? 009 Distributed Memory MIMD Multiprocessor 010 Shared Memory MIMD Multiprocessor 011 Distributed Shared Memory (DSM) 012 Workstation Clusters 013 A Simple Parallel Programming Model 014 Properties 015 Parallel Algorithm Design 016 A Design Methodology 017 Partitioning 018 Communication 019 Agglomeration 020 Mapping 021 Example: Atmosphere Model 022 Atmosphere Model: Numerical Methods 023 Atmosphere Model: Partition 024 Atmosphere Model: Communication 025 Atmosphere Model: Agglomeration 026 Atmosphere Model: Mapping 027 Modeling Performance 028 Bandwidth and Latency 029 Measured Costs 030 Typical Communication Costs 031 Example: Finite Difference 032 Time for Finite Difference 033 Using Performance Models 034 Design Alternatives: Finite Difference 035 Design Alternatives (2) 036 Finding Model Discrepancies 037 Impact of Network Topology 038 Competition for Bandwidth 039 Bandwidth-Constrained Model Versus. Observations 040 Tool Survey 041 High Performance Fortran (HPF) 042 HPF Example 043 HPF Analysis 044 Message Passing Interface (MPI) 045 MPI Example 046 MPI Analysis 047 PCF 048 PCF and OpenMP 049 PCF Example (SGI Variant) 050 OpenMP Example 051 PCF/OpenMP Analysis 052 Portable, Extensible Toolkit for Scientific Computations (PETSc) 053 PETSc Example 054 PETSc Analysis