Get Link
|
Sync TOC
|
<<
|
>>
Intel® MPI Library User's Guide for Linux* OS
Legal Information
Optimization Notice
Introduction
Introducing Intel® MPI Library
Intended Audience
Conventions and Symbols
Related Information
Usage Model
Installation and Licensing
Installing Intel® MPI Library
Licensing for Intel® MPI Library Runtime Environment and Software Development Kit
Compiling and Linking
Compiling an MPI Program
Adding Debug Symbols
Other Compilers Support
Running Applications
Running an MPI Program
Intel® MPI Library Configurations
Multi-threaded Applications
Selecting Fabrics
I_MPI_FABRICS
Debugging and Testing
GDB*: The GNU* Project Debugger
TotalView* Debugger
DDT* Debugger
Logging
Getting Debug Information
Tracing an Application
Checking Correctness
Gathering Statistics
Testing the Installation
Compiling and Running a Test Program
Process Management
Selecting a Process Manager
Scalable Process Management System (Hydra)
Multipurpose Daemon* (MPD*)
Controlling MPI Process Placement
Tuning with mpitune Utility
Cluster-Specific Tuning
Application-Specific Tuning
Setting Time Limit
Setting a Fabrics List
Setting a Range for the Number of Processes
Setting a Limitation for Hosts Usage
Restoring mpitune from the Last Saved Session
Tuning Applications Manually
Job Schedulers Support
Altair* PBS Pro*, TORQUE*, and OpenPBS*
IBM* Platform LSF*
Parallelnavi NQS*
SLURM*
Univa* Grid Engine*
SIGINT, SIGTERM Signals Intercepting
General Cluster Considerations
Defining which Nodes to Use
Password-less ssh Connection
Heterogeneous Systems and Jobs
Troubleshooting
General Troubleshooting Procedures
Examples of MPI Failures
Communication Problems
Environment Problems
Other Problems
Using the Intel® MPI Library with the Intel® Many Integrated Core (Intel® MIC) Architecture
Libraries
Multiple Cards
Using Intel® MPI Library on Intel® Xeon Phi™ Coprocessor
Building an MPI Application
Running an MPI Application