Implement optimized communication patterns to distribute deep learning model training across multiple nodes. This component is part of the Intel® oneAPI Base Toolkit. 

For the most current functional and security features, update to the latest version as it becomes available.

Install with APT

sudo apt install intel-oneapi-ccl-devel

For the next steps, see the Get Started Guide.

Set Up Your Environment

1. Install Miniforge.
2. Create and activate a new conda environment, replacing <your-env-name> with your preferred name for the environment:

   conda create -n <your-env-name>
   conda activate <your-env-name>
   

Install with conda

conda install -c https://software.repos.intel.com/python/conda/ -c conda-forge <package-name>

The following packages are available for installation:

• For both CPU and GPU (requires SYCL):
  • oneccl includes runtime only, needed for running applications that require oneCCL
  • oneccl-devel includes libraries, headers, and tools required to build and compile applications that use oneCCL
• For CPU only (does not require SYCL):
  • oneccl-cpu includes runtime only;needed for running applications that require oneCCL
  • oneccl-cpu-devel includes libraries, headers, and tools required to build and compile applications that use oneCCL

For the next steps, see the Get Started Guide

​Prerequisite for First-Time Users

1. Create the DNF repository file in the /temp directory as a normal user.

   tee > /tmp/oneAPI.repo << EOF
   [oneAPI]
   name=Intel® oneAPI repository
   baseurl=https://yum.repos.intel.com/oneapi
   enabled=1
   gpgcheck=1
   repo_gpgcheck=1
   gpgkey=https://yum.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
   EOF

2. Move the newly created oneAPI.repo file to the YUM configuration directory.

   sudo mv /tmp/oneAPI.repo /etc/yum.repos.d
   

Install with DNF

sudo dnf install intel-oneapi-ccl-devel

For the next steps, see the Get Started Guide.

Set Up Your Environment

Create and activate a virtual environment, replacing <your-env-name> with your preferred name for the environment:

python3.10 -m venv <your-env-name>
source <your-env-name>/bin/activate

Install with pip

sudo pip install <package-name>

• For both CPU and GPU (requires SYCL):
  • oneccl includes runtime only, needed for running applications that require oneCCL
  • oneccl-devel includes libraries, headers, and tools required to build and compile applications that use oneCCL
• For CPU only (does not require SYCL):
  • oneccl-cpu includes runtime only;needed for running applications that require oneCCL
  • oneccl-cpu-devel includes libraries, headers, and tools required to build and compile applications that use oneCCL

For the next steps, see the Get Started Guide.

Install with Spack

spack install intel-oneapi-ccl

For more information, refer to Spack documentation.

For the next steps, see the Get Started Guide.

Prerequisites for First-Time Users

1. Create the YUM repository file in the /temp directory as a normal user.

tee > /tmp/oneAPI.repo << EOF
[oneAPI]
name=Intel® oneAPI repository
baseurl=https://yum.repos.intel.com/oneapi
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://yum.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB
EOF

 

2. Move the newly created oneAPI.repo file to the YUM configuration directory /etc/yum.repos.d:

sudo mv /tmp/oneAPI.repo /etc/yum.repos.d

Install with YUM

sudo yum install intel-oneapi-ccl-devel

For the next steps, see the Get Started Guide.

Prerequisites for First-Time Users

Add the Intel oneAPI repository public key using the following command:

sudo zypper addrepo https://yum.repos.intel.com/oneapi oneAPI

Install with Zypper

sudo zypper install intel-oneapi-ccl-devel

For the next steps, see the Get Started Guide.

Linux* Installation Instructions

For more information, see the Get Started Guide.

Additional Resources

System Requirements

Complete Installation Guide

Release Notes