Learn about the new performance assistant chart, which shows potential performance problems in an MPI program.

Get Started Guide

Analyze OpenMP* & MPI Applications

Learn the basic steps to analyze hybrid OpenMP* and MPI code for inefficiencies using an MPI performance snapshot, Intel® Trace Analyzer and Collector, and Intel® VTune™ Profiler. Discover how to use Intel® Parallel Studio to tune hybrid applications by reviewing MPI use inefficiencies and balancing thread load levels.

Instructions (Linux*)

Detect & Resolve Errors with the MPI Correctness Checker

In addition to regular performance analysis, Intel Trace Analyzer and Collector can perform correctness checking on MPI applications to help you find programming mistakes and detect errors in the execution environment. This tutorial demonstrates the correctness checking workflow for sample MPI applications with the following types of errors:

  • Data type mismatches
  • Deadlocks


Lightweight Statistics for MPI

Intel® Performance Snapshot is a scalable, lightweight performance tool for MPI applications. It collects MPI application statistics, such as communication, activity, and load balance, and presents them in a streamlined format. Learn a complete workflow that includes how to:

  • Detect performance issues
  • Find communication hot spots
  • Review your application

Get Started (Linux)

Analyze MPI Applications

To improve performance of some complex applications, it is necessary to analyze their cross-process behavior and single process performance. Use Intel Trace Analyzer and Collector to analyze communications between processes. Intel VTune Profiler helps you find single process performance issues. This workflow demonstration is applied to a sample program and showing you how to:

  • Find communication imbalance issues in your application
  • Find hot spots on the intraprocess level of your application


Detect & Remove Unnecessary Serialization

When your application is serialized, much time is spent on communicating processes. This tutorial demonstrates a workflow that's applied to a sample program and shows you how to:

  • Find communication problems by analyzing the application with Intel® Trace Analyzer charts
  • Detect serialization in communications between processes
  • Resynchronize the processes to remove serialization
  • Review the application


Reduce Trace File Size

Large trace files are hard to manage. It takes a long time to generate them, a lot of disk space to store them, and they are difficult to analyze. For these reasons, you should reduce the trace file size whenever possible. This lesson applies a workflow to sample code and shows how to apply the following options to reduce the trace file size:

  • Create a configuration file
  • Run the application with the -trace-collectives and -trace-pt2pt options
  • Use the MPI_Pcontrol function


Free Technical Webinars

Access the latest training and how-tos that demonstrate how your software can take advantage of the latest processor technologies using Intel® Software Development Products.

Choose from on-demand offerings and future sessions. Get started here:

Analyze Hybrid Program Performance

Improve Intel® MPI Library Performance

Fast, Lightweight, and Scalable MPI Performance Analysis

Development Tools Webinars


Ready to Get Started?



英特尔的编译器针对非英特尔微处理器的优化程度可能与英特尔微处理器相同(或不同)。这些优化包括 SSE2、SSE3 和 SSSE3 指令集和其他优化。对于在非英特尔制造的微处理器上进行的优化,英特尔不对相应的可用性、功能或有效性提供担保。该产品中依赖于微处理器的优化仅适用于英特尔微处理器。某些非特定于英特尔微架构的优化保留用于英特尔微处理器。关于此通知涵盖的特定指令集的更多信息,请参阅适用产品的用户指南和参考指南。

通知版本 #20110804