A curated registry of resources for NVIDIA Base Command Manager (BCM) — the standard for provisioning, managing, and monitoring HPC and AI clusters at scale, including the Base View web console.
The goal is to be the definitive place to discover and install resources that help administrators and users deploy, operate, and optimize BCM clusters: from cmsh provisioning and workload managers to GPU configuration, Kubernetes, Jupyter, Spark, monitoring, and support.
Inclusion criteria: Resources must be directly related to Base Command Manager or its closely coupled ecosystem (Slurm, PBS, Kubernetes, Jupyter, Spark, environment modules, MPI, GPU tooling on BCM). See CONTRIBUTING.md for the full quality bar.
- Official Documentation and Training
- Installation and Provisioning
- Workload Managers
- GPU Configuration
- Monitoring and Observability
- MPI and Parallel Computing
- Containers and Kubernetes
- Jupyter and Interactive Apps
- Troubleshooting and Support
- Community Tools
- Related Awesome Lists
- Notes
- Contributing
- License
Core product documentation, manuals, release notes, and training for NVIDIA Base Command Manager.
- NVIDIA Base Command Manager
Official— Product home for BCM, the HPC and AI cluster management platform, including Base View. - NVIDIA Base Command Manager Documentation
Official— Docs hub with manuals for BCM 10 and 11, release notes, and deployment guides. - Base Command Manager 11 Administrator Manual
Official— Comprehensive administrator reference for installing, configuring, and managing BCM clusters. - Base Command Manager 11 Installation Manual
Official— Step-by-step installation and quickstart guide for bare-metal BCM clusters. - Base Command Manager 11 User Manual
Official— End-user guide covering modules, MPI, Slurm, PBS, and job submission. - Base Command Manager 11 Developer Manual
Official— Python API, monitoring data producers, and CMDaemon REST API reference. - Base Command Manager 10 Release Notes
Official— Release notes for the BCM 10 major release line. - NVIDIA Base Command Manager 11 Release Notes
Official— Release notes and lifecycle information for the current BCM 11 major line. - BCM 11.33.0 Release Notes
Official— Latest BCM 11 release (May 2026) adding BaseOS 7.5.0, Kgateway/Gateway API, Harbor v2.14, Run:ai via cm-runai-setup, Slurm 25.11. - Base Command Manager Training
Official— NVIDIA training and certification resources for cluster administrators.
Resources for installing BCM, provisioning nodes, managing software images, and configuring cluster hardware.
- Base Command Manager Installation Manual
Official— BCM head-node installation, network setup, and node provisioning.- Run:
cm-setuporcm-wlm-setupafter head-node install.
- Run:
- Provision a Node with NVIDIA BaseOS in BCM
Official— Guide to creating and deploying NVIDIA BaseOS images in BCM. - DGX BasePOD Deployment Guide
Official— BCM-based deployment for DGX BasePOD clusters. - DGX SuperPOD Administration Guide
Official— BCM administration procedures for DGX SuperPOD systems. - cmsh Provisioning
Official— Usecmshto identify, clone, and provision nodes via PXE or node-installer.- Example:
cmsh -c "device newnodes"to discover pending nodes.
- Example:
- Software Image Management
Official— Create, snapshot, update, and assign software images to node categories.- Example:
cmsh -c "image; use default-image; snapshot"to snapshot before changes.
- Example:
Resources for configuring and using Slurm, PBS Professional, OpenPBS, and LSF with BCM.
- Slurm Workload Manager in BCM
Official— User guide for loading Slurm modules, compiling, and submitting jobs. - cm-wlm-setup
Official— TUI wizard for installing and configuring Slurm, PBS, or LSF on BCM.- Run:
cm-wlm-setupon the head node.
- Run:
- PBS Professional and OpenPBS in BCM
Official— PBS job scripts, directives, and submission workflows on BCM. - Slurm Workload Management in NVIDIA Mission Control
Official— Mission Control Slurm integration, IMEX, and topology configuration on BCM. - NVIDIA Mission Control Workload Manager Installation
Official— Install and validate Slurm on GB200/GB300 NVL72 clusters with BCM. - Slurm on DGX BasePOD
Official— Slurm deployment guide for DGX BasePOD with BCM.
Resources for configuring NVIDIA GPUs, MIG, CUDA MPS, DCGM, and GPU workload scheduling in BCM.
- Configuring GPU Settings in BCM
Official— GPU units, GPU settings, andgpusettingssubmode in cmsh. - MIG Configuration in BCM
Official— Multi-Instance GPU profiles and autodetection for Slurmgres.conf.- Run:
cm-mig-manageor set MIG profiles in cmsh.
- Run:
- CUDA Multi-Process Service with cm-wlm-setup
Official— Configure NVIDIA MPS for GPU sharing through the workload-manager setup wizard. - NVIDIA Data Center GPU Manager
Official— DCGM for GPU health monitoring in BCM software images. - NVIDIA Container Toolkit
Official— Container runtime for GPU workloads used by BCM Docker and Kubernetes setups.
Resources for BCM monitoring, metrics, health checks, triggers, Prometheus/Grafana, and diagnostics.
- BCM Monitoring Infrastructure
Official— Metrics, health checks, triggers, actions, andmonitoringmode in cmsh. - BCM Prometheus Exporter
Official— Export BCM metrics to Prometheus for Grafana dashboards.- Query:
curl -sk https://localhost:8081/exporter
- Query:
- Observability Stack Configuration in Mission Control
Official— Wire BCM Prometheus endpoints into a Kubernetes observability stack. - Monitoring Data Producers
Official— Developer guide for custom metrics, health checks, and Prometheus producers. - cm-diagnose
Official— Built-in diagnostic tool to collect cluster state and logs for support.- Run:
cm-diagnoseorcm-diagnose --supporton the head node.
- Run:
Resources for MPI, environment modules, OpenMP, and parallel job execution on BCM clusters.
- Environment Modules in BCM
Official— Usingmodule load,module avail, andmodule liston BCM. - MPI Libraries in BCM
Official— MPICH, MVAPICH, Open MPI, and Intel MPI integration via modules. - Open MPI with CUDA Awareness
Official— GPU-aware Open MPI for CUDA-accelerated parallel applications. - OpenMP and MPI Examples
Official— Sample job scripts for hybrid OpenMP/MPI workloads on BCM. - Machine Learning Manual
Official— ML package modules, TensorFlow, and MPI dependencies on BCM.
Resources for Docker, Kubernetes, Cluster API, GPU Operator, and Spark on BCM.
- BCM Containerization Manual
Official— Docker, Kubernetes, and container management on BCM. - cm-kubernetes-setup
Official— TUI wizard to deploy Kubernetes on BCM nodes.- Run:
cm-kubernetes-setupon the head node.
- Run:
- Cluster API in BCM
Official— Declarative Kubernetes cluster provisioning withcm-kubernetes-capi-setup. - NVIDIA GPU Operator on BCM Kubernetes
Official— Deploy the GPU Operator duringcm-kubernetes-setup. - Run:ai on Base Command Manager
Official— BCM 11.33+ usescm-runai-setupto deploy and configure Run:ai on a BCM-managed Kubernetes cluster. The legacy 2.22 BCM install path is superseded. - Slinky Slurm Operator
Community— Kubernetes operator for full Slurm lifecycle, used in large GPU clusters. - NVIDIA KAI Scheduler
Community— CNCF Sandbox Kubernetes-native GPU scheduler underpinning Run:ai. - NVIDIA AI Cluster Runtime (AICR)
Community— Validated, version-locked Kubernetes recipes for GPU clusters (alpha). - Spark on Kubernetes in BCM
Official— Kubernetes Spark Operator and Spark-based Jupyter kernels on BCM.
Resources for JupyterHub, JupyterLab, Jupyter kernels, VNC, and interactive workload integration.
- cm-jupyter-setup
Official— CLI wizard to deploy JupyterHub on BCM.- Run:
cm-jupyter-setupon the head node.
- Run:
- Using Jupyter in BCM
Official— User guide for Jupyter notebooks, kernels, and extensions on BCM. - Jupyter Kernels for Slurm and PBS
Official— Configure Jupyter Enterprise Gateway kernels backed by H workload managers. - Jupyter VNC Extension
Official— Launch remote desktop sessions from Jupyter on BCM. - Jupyter WLM Magic Extension
Official— Magic commands to submit and manage jobs from Jupyter notebooks.
Resources for diagnosing BCM issues, node boot problems, logs, and getting support.
- Troubleshooting the Node Boot Process
Official— Node-installer logs, provisioning logging, and boot diagnostics. - cm-diagnose
Official— Collect cluster diagnostics and support bundles. - NVIDIA Enterprise Support
Official— Open support cases for Base Command Manager subscriptions. - NVIDIA Developer Forums - Base Command Manager
Community— Community Q&A, troubleshooting, and best practices. - BCM Cluster on Demand OCI PyPI Package
Official— Python CLI to spin up BCM clusters in Oracle Cloud Infrastructure.- Install:
pip install cm-cluster-on-demand-oci
- Install:
Community and automation resources that extend or interact with BCM.
- BCM Developer Manual Python API Examples
Official— Example scripts under/cm/local/examples/cmd/pythoncmfor automation.- Run:
module load python3 && python /cm/local/examples/cmd/pythoncm/status.py
- Run:
- Bright Cluster Installer Ansible Collection
Community— Ansible collection and examples for deploying BCM head nodes.- Install:
ansible-galaxy collection install brightcomputing.installer100
- Install:
- BCM Ansible Playbooks
Community— Community Ansible playbooks for NVIDIA Base Command Manager. - CMDaemon REST API
Official— JSON REST interface for integrating external tools with BCM. - Setting Up IBM Storage Scale with BCM
Community— Reference guide for IBM Storage Scale integration with BCM clusters.
- Awesome NVIDIA Agentic Skills and MCP Servers — Agentic skills and MCP servers for NVIDIA products.
- Awesome GPU Cloud — GPU cloud platforms and HPC cloud resources.
- Awesome MCP Servers — Model Context Protocol servers and integrations.
- Awesome Oracle Agentic Skills and MCP Servers — Agentic skills and MCP servers for Oracle Cloud and databases.
- Version numbering: BCM versions follow a
MAJOR.YY.MMpattern (e.g., 10.25.03, 11.25.08). Links point to the latest BCM 10/11 documentation where available; some deep sections still reference Bright Cluster Manager naming from before the NVIDIA rebrand. - Manual sections: Many capabilities are documented as chapters in the Administrator, Installation, User, Containerization, and Developer manuals rather than as standalone web pages.
- Community entries: Community-maintained tools are listed for convenience; verify compatibility with your BCM version before use in production.
- Updates welcome: If you find a new BCM resource, see CONTRIBUTING.md and open a pull request.
Read CONTRIBUTING.md for the quality bar, entry format, and PR process.
This list is released into the public domain under CC0-1.0.
Enterprise AI Atlas is maintained by Vibe Coding Agency. We prototype and ship agentic systems, MCP servers, and enterprise AI integrations for teams that need working software fast — without hiring a full AI engineering team.
Free guide: The Non-Technical Founder's Guide to Agentic AI — what agents and MCP servers are, and how to get a system built.