---
title: "NCCL for Multiple Sparks — Overview"
canonical: "https://build.nvidia.com/spark/nccl/overview.md"
---

# Basic idea

NCCL (NVIDIA Collective Communication Library) enables high-performance GPU-to-GPU communication
across multiple nodes. This walkthrough sets up NCCL for multi-node distributed training on
two, three, or four DGX Spark systems with Blackwell architecture. You'll configure networking,
build NCCL from source with Blackwell support, and validate communication between nodes.

# What you'll accomplish

You'll have a working multi-node NCCL environment that enables high-bandwidth GPU communication
across DGX Spark systems for distributed training workloads, with validated network performance
and proper GPU topology detection.

# What to know before starting

- Working with Linux network configuration and netplan
- Basic understanding of MPI (Message Passing Interface) concepts
- SSH key management and passwordless authentication setup

# Prerequisites

- Two, three, or four DGX Spark systems
- Completed the matching connection playbook for your node count:
- 2 Sparks — [Connect Two Sparks](https://build.nvidia.com/spark/connect-two-sparks/stacked-sparks)
- 3 Sparks — [Connect Three Sparks](https://build.nvidia.com/spark/connect-three-sparks/three-sparks-ring)
- 4 Sparks — [Connect Multiple Sparks through a Switch](https://build.nvidia.com/spark/multi-sparks-through-switch/multi-sparks)
- NVIDIA driver installed: `nvidia-smi`
- CUDA toolkit available: `nvcc --version`
- Root/sudo privileges: `sudo whoami`

# Time & risk

* **Duration**: 30 minutes for setup and validation
* **Risk level**: Medium - involves network configuration changes
* **Rollback**: The NCCL & NCCL Tests repositories can be deleted from DGX Spark
* **Last Updated:** 12/15/2025
* Use nccl latest version v2.30.7-1