Intel® oneAPI Collective Communications Library Release Notes

Overview

The Intel® oneAPI Collective Communications Library (oneCCL) enables developers and researchers to more quickly train newer and deeper models. This is done by using optimized communication patterns to distribute model training across multiple nodes.

The library is designed for easy integration into deep learning (DL) frameworks, whether you are implementing them from scratch or customizing existing ones.

• Built on top of lower-level communication middleware - MPI and OFI (libfabrics) which transparently support many interconnects, such as Cornelis Networks, InfiniBand, and Ethernet.
• Optimized for high performance on Intel® CPUs and GPUs.
• Allows the tradeoff of compute for communication performance to drive scalability of communication patterns.
• Enables efficient implementations of collectives that are heavily used for neural network training, including allreduce, and allgather.

2022.0 Release 

What's New 2022.0 - oneAPI 2026.0

• Improved NCCL compatibility by setting NCCL like C API by default
• Intel® Arc™ Pro B-Series support delivers optimized scale up performance leveraging low latency protocol
• SPMD support for Allgather, Allreduce, Alltoall, ReduceScatter, Broadcast, pt2pt and Group API for scale up on Intel® Arc™ Pro B-Series
• Added support for user defined reduction operations for scale out on Intel® Data Center GPU Max Series
• Added reduction operations for scale up on Intel® Arc™ Pro B-Series
• Improved profiling information to assess imbalance across communicating processes by tracing tools
• Added onecclCommWindowRegister, onecclCommWindowDeregister, onecclMemAlloc, onecclMemFree APIs
• Introduced support for SYCL graph for scale up

Major Features Supported

• Service functionality
  • Interoperability with SYCL*:
  • Construction of oneCCL communicator object based on SYCL context and SYCL device
  • Construction of oneCCL stream object based on SYCL queue
  • Construction of oneCCL event object based on SYCL event
  • Retrieving of SYCL event from oneCCL event associated with oneCCL collective operation
  • Passing SYCL buffer as source/destination parameter of oneCCL collective operation

Notice of Deprecation

Deprecation of C++ API

• In oneCCL version 2021.17 included with the 2025.3 oneAPI release, oneCCL will add support for a new C API that closely follows the NVIDIA Collective Communications Libary (NCCL)* API standard. The existing C++ API will remain available and will remain the default API for the 2021.17 release. Details explaining how an application may link against and use the new API will be shared in this release.

See the oneCCL C API RFC document to view the proposed API and provide any feedback. 

• In oneCCL version 2022.0 included with the 2026.0 oneAPI release, oneCCL will use the new NCCL* like C API by default. This is a breaking change. The legacy C++ API will remain available, and details explaining how an application may link against and use the legacy API will be included in the 2022.0 release.

Applications cannot use both the C and C++ APIs simultaneously.

Support for the legacy C++ API shall remain in the release until future notice. The schedule for legacy API removal will be announced here.

System Requirements

Hardware Requirements

• CPU

  • Intel® Core™ Processor Family
  • Intel® Xeon® Processor Family
  • Intel® Xeon® Scalable Performance Processor Family
  • Intel® Xeon® 6 Processor Family
• GPU
  • Intel® Arc™ Pro B-Series B60 and B70 SKUs
  • Intel® Data Center GPU Max Series

Software Requirements

• Operating Systems:

  • Ubuntu* LTS 24.04, 22.04 
  • SUSE* Linux Enterprise Server* 15 SP4 SP5 SP6 SP7 
  • Red Hat* Enterprise Linux*  8, 9, 10
  • Rocky Linux 9
  • Fedora 41, 42
  • Debian 11, 12
  • Amazon Linux 2025, 2023
  • WSL2 (SLES, Ubuntu)
• Compilers:
  • GNU*: C, C++ 4.8.5 or higher
  • Intel® C++ Compiler for Linux* OS 19.1 or higher
  • Intel® C/C++ Compiler Classic 2021.1
  • Required for GPU support: Intel® oneAPI DPC++/C++ Compiler
• Fabric Software:
  • Open Fabric Interface (OFI): source codes or provided with the Intel® MPI Library
    • Minimum: OFI 1.13.2
    • Recommended: the latest OFI "main" branch
    • Build and install instructions

• SYCL support:

  • Intel® oneAPI DPC++/C++ Compiler with Level Zero support. To install Level Zero, refer to the instructions in Intel® Graphics Compute Runtime for oneAPI Level Zero and OpenCL™ Driver Runtime repository or to the installation guide for oneAPI users.

• BF16 support:
  • AVX512F-based implementation requires GCC 4.9 or higher.
  • AVX512_BF16-based implementation requires GCC 10.0 or higher and GNU binutils 2.33 or higher.
• FP16 support:
  • FX16C and AVX512F-based implementations require GCC 4.9, Clang 9.0 or higher.
  • AVX512FP16-based implementation requires GCC 12.0, Clang 14.0, Intel 2021.4.0 or higher.

Notices and Disclaimers

Intel technologies may require enabled hardware, software or service activation.

No product or component can be absolutely secure.

Your costs and results may vary.

© Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.

No license (express or implied, by estoppel or otherwise) to any intellectual property rights is granted by this document.

The products described may contain design defects or errors known as errata which may cause the product to deviate from published specifications. Current characterized errata are available on request.

Intel disclaims all express and implied warranties, including without limitation, the implied warranties of merchantability, fitness for a particular purpose, and non-infringement, as well as any warranty arising from course of performance, course of dealing, or usage in trade.

​