Intel® Inspector User Guide for Windows* OS

User Guide

  • 2022.0
  • 12/06/2021
  • Public Content
Contents

Command Line Output

Output from the
Intel Inspector
inspxe-cl
 command includes the following:

Result Directory

A
collect
 or
collect-with
 action generates a result directory to contain the analysis result.
By default, the result directory name is
r@@@{at}
 where:
  • @@@
     - next available number for similarly named result directories in the current working directory. The series begins with 000.
  • {at}
     - analysis type code. This code represents the analysis type used when the result was created.
For example:
r001ti3
, where:
  • 001
     - next available number
  • ti3
     - the selected analysis type (
    ti
     stands for "threading error analysis" and
    3
     indicates that this uses the level 3 setting, which is the highest setting. The
    ti3
     analysis type is known as the
    Locate Deadlocks and Data Races
     setting.)
The result directory name may be assigned according to the default naming conventions, or you may use the
result-dir
 option to specify the result directory name during collection. Your specified result directory name can include up to five @ counters, so the result directory can be specified in scripts without having to update the result directory value each time.
By default, result directories are created in the current working directory. If you want to specify an alternative location, use the
user-data-dir
 option to specify the parent directory for result directories. If you use this option during collection, remember to include it for any other actions you might perform on these results, such as generating reports.
You cannot put multiple results into the same result directory. If you specify the same result directory for multiple analysis runs, an error is returned. Use the auto-increment counter (@@@) to work around this restriction.

Result File

During a
collect
 or
collect-with
 action, the
inspxe-cl
 command tool creates a result directory in which it stores the
*.inspxe
 result file. In addition to analysis data, the result file stores data about the analysis environment, including start time, host machine name, product ID, CPU count, and analysis type used for collection.
The result filename is derived from the result directory name. For example: If the result directory name is
r001ti3
, the result filename is
r001ti3.inspxe
.

Summary Output, Log and Errors

At the end of an analysis run, the
inspxe-cl
 command tool also writes the following files to the result directory:
  • inspxe-cl.txt
    : Summary that includes the total number of problems found and types of problems found.
  • inspxe-cl.log
    : Information useful to the Intel support team if the
    inspxe-cl
     command encounters internal errors
  • The summary is also written to
    stdout
    .
  • Internal errors are written to
    stderr
    .

Exit Codes

The
inspxe-cl
 command returns the following exit codes:
0
Success and no new problems detected
1
Usage error
2
Internal error
4
Application returned a non-zero exit code
8
At least one new problem detected
12
Application returned a non-zero exit code and at least one new problem detected
Use the
return-app-exitcode
 action-option if you want to return the application exit code instead of the
inspxe-cl
 exit code.

Suppression Files

Suppression files can be created and used in the GUI or on the command line.
  • Suppression files that suppress all problems in the code (suppress-all files) can only be created from the command line.
  • Use the command line to convert third-party suppression files (Purify, Valgrind) to a format that can be used by
    Intel Inspector
    .
  • Once created, you can use suppression files when performing analysis or generating reports.

Reports

After collecting a result, you can use the
report
 action to generate one or more types of reports from the result:
  • status
    : Brief statement of the total number of new problems found, and a breakdown by state.
  • problems
    : Detailed list of all new problem sets found, including their location in the source code.
  • observations
    : Detailed list of all code locations in the source code used to form new problem sets.
  • summary
    : This report is generated automatically in .txt form, but can also be generated on demand. Brief statement of the total number of new problems found, and a breakdown by problem type.
By default, the report is output in text format to
stdout
.
  • To write a report to a file, use the
    report-output
     option when generating the report.
  • To specify CSV or XML format, use the
    format
     option.
  • To control how a report displays, you can use the
    sort-asc
    ,
    sort-desc
     and
    filter-include
     options.
  • To view a report from a result directory other than the default, use the
    result-dir
     option.

Product and Performance Information

1

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.