User Guide

  • 2020 Update 2
  • 07/28/2020
  • Public Content

Data Transfers per Function

Key:
-u [--transfers_per_function]
This chart provides information about data transferred by each MPI function (for all ranks). This information can be useful for identifying functions most used for data transfers.

Example

| Function summary for all ranks |----------------------------------------------------------------------- |           Function       Volume(MB)        Volume(%)        Transfers |-----------------------------------------------------------------------               Reduce          1385.98          66.6665            65128                Csend          692.997          33.3335            78073            Allreduce      0.000839233      4.03675e-05              132               Gather      0.000335693       1.6147e-05               44                Bcast      0.000312805      1.50461e-05               50                 Send       3.8147e-06      1.83489e-07                1              Barrier                0                0              314 |======================================================================= | TOTAL                       2078.98              100           143742

Chart Entries

Columns
Rows
Function
Function name
Volume(MB)
Amount of data sent and received by the function (in megabytes)
Volume(%)
Percentage of amount of data sent and received by the function
Transfers
Number of function calls
TOTAL
Total values for all columns

Product and Performance Information

1

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