Developer Guide

Contents

Using Console Version of Custom Library Tool

Follow the steps below to build a custom dynamic library using console version of the Custom Library Tool:
  1. Define a list of Intel IPP functions that the Intel IPP Custom Library Tools should export to your custom dynamic library. See the example text file below:
  2. Run
    python main.py
    with the following parameters:
    -c
    ,
    --console
    Launches the console version of the tool (the GUI version is used by default).
    -g
    ,
    --generate
    Enables the script generation mode (the build mode is used by default).
    -n <
    name
    >
    ,
    --name <
    name
    >
    Output library name.
    -p <
    path
    >
    ,
    --path <
    path
    >
    Path to the output directory.
    -root <
    root_path
    >
    Path to Intel IPP or Intel IPP Cryptography package root directory
    -f <
    function
    >
    ,
    --function <
    function
    >
    Name of a function to be included into your custom dynamic library.
    -ff <
    functions_file
    >
    ,
    --functions_file <
    functions_file
    >
    Path to a file with a list of functions to be included into your final dynamic library (the
    -f
    or
    --function
    flag can be used to add functions on the command line).
    -arch={ia32|intel64}
    Enables all actions for the IA-32 or the Intel ® 64 architecture (Intel ® 64 architecture is used by default).
    -mt
    ,
    --multi-threaded
    Enables multi-threaded libraries (single-threaded libraries are used by default).
    -tl={tbb|openmp}
    Sets Intel TBB or OpenMP* as the threading layer.
    -d , --custom_dispatcher <
    cpu_set
    >
    Sets the exact list of CPUs that must be supported by custom dynamic library and generates a C-file with the custom dispatcher.
    -h
    ,
    --help
    Prints command help.
    For example:
    # Generate build scripts in console mode # with the output dynamic library name “my_custom_dll.dll” # with functions defined in the “functions.txt” file # optimized only for processors with # Intel® Advanced Vector Extensions 512 (Intel® AVX-512) # using multi-threaded IA-32 Intel IPP libraries python main.py -c -g –n my_custom_dll -p “C:\my_project” -ff “C:\my_project\functions.txt” -d avx512bw -arch=ia32 -mt

Product and Performance Information

1

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