---
id: tcmlib
version: "1.5.0"
license: Intel Simplified Software License
license_treatment: unclear
maintenance: active
---

# tcmlib — Thread Composability Manager

License: unclear · Maintenance: active · Downloads: 2.8M/mo

## What it is and what it does

tcmlib is a resource coordinator for Intel's parallel computing ecosystems. It sits between Intel oneAPI Threading Building Blocks (oneTBB) and Intel OpenMP, two popular threading runtimes that can coexist in the same process. When both are active, they can oversubscribe the CPU—spawning more threads than physical cores, which degrades performance. tcmlib prevents this by managing how many threads each runtime is allowed to create, ensuring they cooperate rather than compete.

The package is designed for developers using both oneTBB and OpenMP in scientific computing, data processing, or high-performance applications. It requires no explicit API calls—importing it activates coordination automatically. The library is platform-specific (x86_64 Linux and Windows only) and carries an Intel proprietary license. With no external Python dependencies and active maintenance, it integrates cleanly into existing projects, but its exact Python version support and system-level prerequisites are not documented.

Use it for:

- Optimize performance in scientific applications that use both oneTBB and OpenMP for parallelism.
- Prevent thread oversubscription in hybrid parallel workloads mixing multiple Intel threading libraries.
- Improve CPU utilization and reduce context-switching overhead in multi-threaded data processing pipelines.
- Coordinate resource allocation in high-performance computing environments with strict CPU budgets.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Manages CPU resource coordination between Intel oneAPI Threading Building Blocks and Intel OpenMP to prevent oversubscription when both runtimes run in the same process.

Yes, if you use both Intel oneAPI Threading Building Blocks and Intel OpenMP in the same process and want to avoid oversubscription. Install friction is moderate (platform-specific wheels), maintenance is active, and there are no known vulnerabilities. The main uncertainty is Python version compatibility and system prerequisites—verify those before committing. The proprietary license is standard for Intel tools but warrants review for commercial redistribution.

## Install

pip install tcmlib
uv add tcmlib
poetry add tcmlib

## Installing tcmlib

Before you install: Medium install friction due to platform-specific wheels (manylinux_2_28_x86_64, win_amd64). Active maintenance with a recent release (112 days old). No runtime dependencies to manage.

License in practice: Licensed under Intel Simplified Software License (proprietary, not SPDX-identified). License treatment is unclear—review Intel's terms before integrating into commercial or redistributed software.

Quickstart:

pip install tcmlib==1.5.0
import tcmlib
# tcmlib coordinates threading resources automatically when imported

Requires x86_64 architecture (manylinux_2_28 or Windows); Python version support is unspecified—verify compatibility before use.

Verify before relying:

- Exact Python version requirements and compatibility range not specified in metadata.
- Whether tcmlib requires Intel TBB and OpenMP to be pre-installed on the system.
- Performance impact and tuning options for resource coordination.
- Behavior and guarantees when only one threading runtime is present.

## Package facts

- License: Intel Simplified Software License (unclear)
- Python support: unspecified
- Install friction: medium
- Maintenance: active
- Downloads: 2.8M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

thread composability manager, oneapi threading coordination, openmp oversubscription prevention, cpu resource management threading, intel tbb openmp integration, thread pool coordination, runtime resource balancing, threading-coordination, intel-oneapi, performance-tuning

[View on SkillFed](https://skillfed.io/packages/tcmlib) · [View on PyPI](https://pypi.org/project/tcmlib/)