Registry / devops / intel-cmplr-lib-ur

intel-cmplr-lib-ur

JSON →
library2026.0.0pypypiunverified

The Intel® oneAPI Unified Runtime Libraries (`intel-cmplr-lib-ur`) package contains shared common libraries essential for providing a unified interface to device-agnostic runtimes like DPC++ across various software platforms. It is a foundational component within the Intel oneAPI ecosystem, often installed alongside common licensing and compiler-specific runtimes. The current version is 2025.3.3, with a rapid release cadence, indicating frequent updates.

pip install intel-cmplr-lib-ur
INSTALL
IMPORT
SIG · INTEL-CMPLR-LIB-UR
I
intel-cmplr-lib-ur
devopspythonv2026.0.0
Install
2.5s avg
Import
Disk
16MB
Pass rate
5/ 10
Env Coverage5 / 10
glibc
3.93.13
musl
3.93.13
Install & Compatibility
Where this runs
tested against v2026.0.0 · pip install
no network on importno background threads
Install × environment matrix
Each cell = how many times install + import succeeded across repeated harness runs. Partial = flaky.
glibc = Debian/Ubuntu slim · musl = Alpine Linux
musl
py 3.103.920 runs
build_error
glibc
py 3.103.920 runs
installs and imports cleanly · install 2.5s · import 0.000s · 18MB
16MB installed
● package 16MB
Code
Verified usage

Verified import paths — ran on the pinned version, not inferred.

None (backend library)
This library primarily provides underlying shared C/C++ libraries. Direct end-user Python imports (e.g., `import intel_cmplr_lib_ur`) are not common as it serves as a backend dependency for other oneAPI Python bindings and compilers (e.g., `dpctl`).
This package is a runtime library, not typically imported directly by end-user Python code. Its functionality is exposed through other higher-level oneAPI Python libraries.

The `intel-cmplr-lib-ur` package is a core runtime library, not typically directly interacted with in Python. Its primary function is to enable other oneAPI-optimized Python libraries, such as `dpnp` (Data Parallel NumPy) and `dpctl` (Data Parallel Control Library), to access and utilize Intel hardware (CPUs, GPUs). The quickstart demonstrates how to use `dpnp` and `dpctl`, which rely on the underlying runtime components provided by `intel-cmplr-lib-ur`, to perform array operations on a selected SYCL-enabled device.

# This package provides underlying runtime libraries. # It's primarily a dependency for other oneAPI Python libraries or C/C++ applications. # For example, to use oneAPI accelerated NumPy-like arrays via dpnp (which depends on underlying oneAPI runtimes): # Ensure intel-cmplr-lib-ur is installed (usually via 'pip install intel-cmplr-lib-ur') # Then, install higher-level Python bindings like dpnp or dpctl: # pip install dpnp dpctl import dpnp as np import dpctl # Select a SYCL-enabled device (e.g., GPU if available, otherwise CPU) try: queue = dpctl.SyclQueue('gpu') print("Running on GPU.") except dpctl.SyclQueueError: queue = dpctl.SyclQueue('cpu') print("Running on CPU.") a = np.asarray([1.0, 2.0, 3.0], sycl_queue=queue) b = np.asarray([4.0, 5.0, 6.0], sycl_queue=queue) c = a + b print(f"Result of a + b on {queue.device.name}: {c}")
Debug
Known issues
breakingThe package (and related oneAPI components) has been flagged for security concerns, specifically the 'Presence of code relocations indicates that the code segment might temporarily, at one point, become both writable and executable. That violates security policies adopted by most modern Linux distributions.' This could allow an attacker to overwrite code with a malicious program during this brief period.
fix
Review the software component behaviors for anomalies, consider exploratory testing in a sandbox, and replace with a more widely used alternative if possible, or vet thoroughly before use.
affects: All versions (identified in 2025.0.5 analysis, likely applies broadly)
gotchaInstalling patch updates for stand-alone performance libraries (e.g., IPP or MKL) over an Intel oneAPI Base Toolkit installation can unintentionally remove compilers (like dpcpp and Fortran) and other performance libraries, especially on Linux and Windows.
fix
For a smoother upgrade, it is advised to install the complete Base/HPC toolkit packages instead of individual component patches. If installing patches, ensure required compiler patch packages (dpcpp and/or fortran) are installed in advance. On Windows, install the TBB package patch before other library patches.
affects: oneAPI Base Toolkit 2025.3.0 and newer
gotchaDependency conflicts can arise between oneAPI runtime components. For instance, `intel-sycl-rt` might require a specific older version of `intel-cmplr-lib-ur`, which can lead to installation failures in complex environments like PyTorch builds.
fix
Carefully manage version compatibility across all oneAPI components. When encountering conflicts, consult Intel's official documentation or release notes for compatible component versions within a specific toolkit release. Consider using a consistent oneAPI toolkit installation rather than mixing individual component versions.
affects: All versions within the oneAPI ecosystem (e.g., 2025.0.2 vs 2025.1.0 seen in conflict)
Upgrade
Version history
2026.0.0latest on PyPI · released Apr 24, 2026
Audit
Dependencies
oneAPI toolkits (e.g., Base Toolkit)requiredThis package provides underlying runtime libraries primarily consumed by other oneAPI components, including higher-level Python libraries like `dpctl`, `intel-sycl-rt`, and C/C++ applications. It often requires other oneAPI components like a common licensing package and compiler-specific runtimes for full functionality.
Agent activity
4 hits · last 30 days
node
4
Resources
intel-cmplr-lib-ur — pip install intel-cmplr-lib-ur · libregistry