Intel oneAPI Runtime Common Libraries

Warnings

• gotcha This library does not provide a direct Python API. It is a collection of native runtime libraries for executables compiled with Intel oneAPI compilers. Expecting Python-level imports or functionality will lead to confusion.
  Fix: Understand its purpose as a runtime dependency for compiled code, not a Python-callable library. Its Python package exists for dependency management within Python ecosystems that leverage Intel-optimized native binaries.
• gotcha Proper environment variable setup (e.g., `LD_LIBRARY_PATH` on Linux, `PATH` on Windows) is crucial for applications to find these runtime libraries at execution. Failing to set these can result in 'library not found' errors, even if the package is installed via pip.
  Fix: Ensure that the directories containing the Intel runtime libraries are correctly added to the system's dynamic linker search paths (`LD_LIBRARY_PATH` or `PATH`) before running applications that depend on them. Intel oneAPI's `setvars.sh` (or `.bat`) scripts typically handle this.
• breaking Breaking changes in the underlying Intel compilers or oneAPI toolkits (e.g., deprecated compiler flags, removal of support for older hardware like Intel Xeon Phi) can indirectly affect executables that depend on `intel-cmplr-lib-rt`. Users might encounter issues if their applications were built with older compiler versions or target deprecated hardware features.
  Fix: Refer to the release notes for the specific Intel oneAPI compiler used to build the dependent application. Recompile applications with newer compilers and adjust flags or code as necessary to align with current best practices and supported features.
• gotcha This package is often a component of a larger Intel oneAPI installation. For a complete and correctly configured runtime environment, it should typically be installed alongside `intel-cmplr-lic-rt` and one or more compiler-specific runtime packages (e.g., `intel-sycl-rt`). Missing these companion packages can lead to incomplete runtime environments.
  Fix: Ensure all necessary Intel oneAPI runtime components are installed. When using `pip`, consider the dependency requirements of higher-level packages (like `intel-extension-for-pytorch`) which might explicitly list these associated runtime dependencies.

Install

• pip install intel-cmplr-lib-rt Install via pip

Quickstart

The `intel-cmplr-lib-rt` library does not offer a direct Python API for application development. It serves as a runtime dependency for executables, particularly those compiled with Intel's C++ or DPC++ compilers. Its inclusion in a Python environment usually signifies that a higher-level Python package depends on the underlying Intel runtime components. The `quickstart.code` block above illustrates how it might be used in a C++ compilation and execution flow, emphasizing the critical role of environment variables for linking and runtime library discovery.

# The intel-cmplr-lib-rt package does not expose a direct Python API for end-user applications.
# It provides runtime components primarily for executables compiled with Intel C++ or DPC++ compilers.
# Its presence in a Python environment is typically as a dependency for other Python packages
# that wrap or utilize Intel-compiled native code (e.g., Intel Extension for PyTorch).
# There are no standard Python import statements or direct usage patterns for this library.
# For C++ applications, linking and runtime environment setup are crucial:
# For compilation (example, requires Intel C++ compiler environment):
#  export LIBRARY_PATH=/path/to/intel/oneapi/compiler/latest/lib:$LIBRARY_PATH
#  icpx -o my_app my_app.cpp -L/path/to/intel/oneapi/compiler/latest/lib -lintel-cmplr-lib-rt
# For execution, ensure runtime libraries are discoverable:
#  export LD_LIBRARY_PATH=/path/to/intel/oneapi/compiler/latest/lib:$LD_LIBRARY_PATH
#  ./my_app