Index
Prev
Index
building MPI applications
Building MPI Applications
distributed programs
Launching a Distributed Program
environment variable setting
Setting MPI Environment Variables
frequently asked questions
Frequently Asked Questions
internal message buffering
Internal Message Buffering in MPI
MPI
components
MPI Components
overview
MPI Overview
mpirun
argument file
Using a File for
mpirun
Arguments
command
Syntax of the
mpirun
Command
for distributed programs
Launching a Distributed Program
for local host
Launching Programs on the Local Host Only
MPT
components
Overview
overview
Overview
Optimizations
application
Application Optimizations
collective calls
Optimized Collective Calls
constructs to avoid
Avoiding Certain MPI Constructs
eliminating retries
Eliminating Retries
less restrictive single-copy
Less Restrictive Single Copy Using the XPMEM Driver
low latency, high bandwidth
Optimized Point-to-Point Calls
NUMA placement, application
NUMA Placement
NUMA placement, runtime
NUMA Placement
one-sided operations
MPI One-Sided Operations
point-to-point calls
Optimized Point-to-Point Calls
profiling
Additional profiling
reducing runtime variability
Reducing Runtime Variability
runtime
Runtime Optimizations
single-copy
Single Copy Optimization
single-copy with XPMEM driver/BTE
Single Copy Using the XPMEM Driver and the BTE
tips
Tips for Optimizing
traditional single-copy
Traditional Single Copy Optimization and Restrictions
using MPI statistics
Using MPI Statistics
using PCP
Performance-CoPilot
using perfex
perfex
using profiling
Using Profiling
using Speedshop
Speedshop
program development
MPI Program Development
program segments
Launching a Distributed Program
sprocs
Thread-Safe MPI
threads
collectives
Collectives
exception handlers
Exception Handlers
finalization
Finalization
initialization
Initialization
internal statistics
Internal Statistics
probes
Probes
query functions
Query Functions
requests
Requests
signals
Signals
thread-safe systems
Thread-Safe MPI
troubleshooting
Frequently Asked Questions