MindStudio Ops Profiler Development Environment Setup, Compilation, and UT Methods

1. Background Knowledge Required

For details about the code framework and core process, see the msOpProf Architecture Design Specifications.

2. Development Environment Setup

For hardware environment requirements, see Ascend Product Models.
Set up the environment by referring to the Operator Tool Development Environment Setup Guide.
The CMake version must be between 3.20.2 and 3.31.10 (inclusive).
Run the pip install numpy command to install the NumPy Python dependency library.
The generated .run package depends on pigz, which is typically provided by the system. Run the pigz --version command to check whether pigz has been installed.

3. Building and Packaging

There are two methods, each with its own advantages and disadvantages:

Method	Application Scenario	Advantages	Disadvantages
One-click script	Initial build and CI/CD pipeline	Zero configuration, one-step setup	Steps cannot be executed independently.
Step-by-step script	Routine development and incremental build	Flexibility and efficiency	Multiple steps are required.

3.1 Method 1: One-click Script

python build.py

3.2 Method 2: Step-by-Step Script

3.2.1 Downloading Dependencies

python download_dependencies.py

3.2.2 Building and Packaging

3.2.2.1 Starting Build

Run the following commands to start the build:

mkdir build
cd build
cmake ../cmake   # If you only need to build the project and do not need to generate a .runfile, run cmake .. instead.
make -j$(nproc)  # -j indicates the number of parallel build jobs, which can be specified as required. If nproc is unavailable, manually enter a number (for example, -j8).

NOTE

Debug version build
To perform GDB or VS Code graphical breakpoint debugging, build the debug version. The procedure is as follows:
Add -DCMAKE_BUILD_TYPE=Debug when running the preceding CMake command, for example, cmake ../cmake -DCMAKE_BUILD_TYPE=Debug.

If the generation time of the mindstudio-opprof_xxx_xxx.run file in the output directory is updated to the current build time, the building and packaging are successful.

3.2.2.2 Build Result Description

The build result is generated in the output directory:

output/
|-- bin                                                  # Executable binary file, which can directly call the tuning function
|-- filelist.csv                                         # Configuration file for packaging the .run package
|-- lib64                                                # Various dynamic and static libraries
|-- mindstudio-opprof/                                   # Directory for packing the .run package
|-- mindstudio-opprof_xxx_xxx.run                        # Installation package
|-- parser_install.sh                                    # Public script for parsing .run package installation
|-- version.info                                         # Version information

3.2.3 Cleanup/Rebuild

Delete the build directory and perform 3.2.2.1.

rm -rf build

4. Running UT

4.1 Method 1: One-Click Script

python build.py test

4.2 Method 2: Step-by-Step Script

4.2.1 Downloading Dependencies

python download_dependencies.py test

4.2.2 Running UT

NOTE

CMake entry description for UT
The UT build uses CMakeLists.txt in the root directory (that is, cmake .. instead of cmake ../cmake) of the project. Only the test and dependency are built, and the .run packaging process is not included.

mkdir build_ut
cd build_ut
cmake .. -DBUILD_TESTS=ON
make -j$(nproc) # -j indicates the number of parallel build jobs, which can be specified as required. If nproc is unavailable, manually enter a number (for example, -j8).
./test/ut/msopt_test --gtest_output=xml:test_detail.xml

If the number of executed test cases is the same as the number of passed test cases, the test is successful. The output is similar to the following:

[==========] 526 tests from 75 test cases ran. (24571 ms total)
[  PASSED  ] 526 tests.

4.2.3 Cleanup/Rebuild

Delete the build directory and perform 4.2.2.

rm -rf build_ut

5. FAQ

5.1 Why Is No .run Package Generated When I Run the make Command During Building?

It is possible that cmake .. is used when running the cmake command. The cmake command is described as follows:

cmake ..: Only the current project is built. The make install command installs the project to the output/ directory, but does not call makeself. Therefore, no Ascend-mindstudio-opprof-xxx.run file is generated.
cmake ../cmake: The "super build" of cmake/CMakeLists.txt is used. The project is built and installed first, and then parser.py and makeself are executed to generate the .run file in the output/ directory.