Table of Contents

[SYMBOL] [A] [B] [C] [D] [E] [F] [G] [H] [I] [J] [K] [L] [M] [N] [O] [P] [Q] [R] [S] [T] [U] [V] [W] [X] parallel filesystems      ClusterNFS      Coda      Intermezzo      Lustre      NFS (Network File System)      OpenAFS      PVFS [See PVFS] parallel networks parallel program design      algorithm design      MPI-IO      MPI-IO functions      parallel I/O          impact of hardware      parallel versus serial algorithms      problem decomposition          control decomposition          data decomposition 2nd          data dependency graphs          decomposition strategies          degree of concurrency          tasks and task granularity      random numbers          libraries for generating          SPRNG (Scalable Parallel Random Number Generators)      tasks, mapping to processors          communication overhead          communication, minimizing impact of          load balancing          redundant calculations          task characteristics, evaluating          work pools parallel programming      subdivision of problems 2nd parallel programming libraries 2nd parallel programs      debugging          compiler features, using          ddd          deadlock          diagnostic code          fflush, using with printf          gdb [See gdb]          lint and splint          memory leaks, checking for          printf, tracing with          programming tools          rereading code          symbolic debuggers          synchronization problems      profiling [See profiling] Parallel Virtual File System [See PVFS] parallelism PBS (Portable Batch System) 2nd 3rd [See also OpenPBS] PBS roll (Rocks) pbsnodes command (PBS) pcount striping parameter (PVFS) peer-to-peer networks performance hardware PFC (power-factor corrected) power supplies pfilter pile-of-PCs (POP) pipelining 2nd POP (pile-of-PCs) Portable Batch System [See PBS] Portable Batch System (PBS) roll power factor, calculating power requirements, estimating precompiled kernels      installing prepareclient script printf      debugging, issues of using for private interface private IP address spaces process migration (openMosix) processor array prof profilers      gcov      gprof profiling      asymptotic analysis      code optimization and      gcc and -O compiler flag      justification for      MPE      MPI, gprof, and gcov      OSCAR and      Rocks and      timing C-language code segments          gettimeofday function          MPI, manual timing [See MPI, code timing with]          time system call      timing programs          profilers [See profilers]      writing code versus optimization programming software 2nd      choosing      debuggers      HDF5      LAM/MPI          installing          testing          user configuration          using      library selection          MPI          PVM      MPICH          installing          MPE (Multi-Processing Environment)          on Windows systems          testing          user configuration          using      OSCAR and      programming languages      Rocks and      SPRNG proprietary clusters ps command pseudorandom number generators public interface purify pushimage command PVFS (Parallel Virtual File System) 2nd 3rd      advantages and disadvantages      architecture      client setup          mknod command          pvfstab file          useful utilities      cluster partitioning      downloading      head node configuration      head node, installing on      I/O server setup      metadata server, configuring      needed patches      OSCAR and Rocks      ownerships and permissions, setting      running          daemons, starting up          insmod command          iod-ping and mgr-ping utilities          mount command          ps command          troubleshooting      striping scheme and parameters      using          pvstat utility          ROMIO interface          u2p utility          ÒUsing the Parallel Virtual File SystemÓ documentation      web site pvfstab file (PVFS) PVM (Parallel Virtual Machine) library pvstat utility (PVFS)

Table of Contents