BotBeat
...
← Back

> ▌

CanonicalCanonical
PRODUCT LAUNCHCanonical2026-02-28

LXD 6.7 Released with AMD GPU Passthrough Support for Containers

Key Takeaways

  • ▸LXD 6.7 introduces AMD GPU passthrough support for containers via the new AMD CDI interface and gpu_cdi_amd extension
  • ▸The release includes major QEMU and EDK2 version updates that improve overall VM GPU passthrough functionality
  • ▸New features include storage pool database recovery for clusters, forced instance deletion API, and x86_64-v3 architecture support
Source:
Hacker Newshttps://www.phoronix.com/news/LXD-6.7-Released↗

Summary

Canonical has released LXD 6.7, introducing AMD GPU passthrough support for containers through the AMD CDI (Container Device Interface) toolkit. This significant update enables users to pass AMD GPUs directly to LXD containers using the new gpu_cdi_amd extension, which is bundled into the Snap package. The implementation allows straightforward configuration through simple command-line operations, bringing GPU acceleration capabilities to containerized workloads on AMD hardware.

Beyond AMD GPU support, LXD 6.7 includes general improvements to VM GPU passthrough capabilities with updated QEMU and EDK2 versions. The release also enhances the initial access experience for the LXD UI and introduces storage pool database recovery support for clusters, which should improve reliability in multi-node deployments.

Additional features in this release include forced instance deletion through the API, optimized instance state field retrieval for better performance, and support for x86_64-v3 architecture variant images. These improvements collectively strengthen LXD's position as a comprehensive system container and virtual machine manager, particularly for Ubuntu Linux environments where it sees widespread adoption in enterprise and cloud infrastructure deployments.

  • AMD GPU passthrough can be configured with simple commands using the LXD CLI

Editorial Opinion

This release represents a significant step toward GPU acceleration parity between NVIDIA and AMD hardware in containerized environments. While NVIDIA has long dominated the container GPU passthrough space, Canonical's implementation of AMD CDI support democratizes access to GPU-accelerated workloads for organizations standardized on AMD hardware. The timing is particularly relevant as AMD continues gaining market share in both datacenter and AI/ML applications, where containerized GPU workloads are increasingly critical.

MLOps & InfrastructureAI HardwareProduct LaunchOpen Source

Comments

Suggested

Google / AlphabetGoogle / Alphabet
RESEARCH

Deep Dive: Optimizing Sharded Matrix Multiplication on TPU with Pallas

2026-04-05
GitHubGitHub
PRODUCT LAUNCH

GitHub Launches Squad: Open Source Multi-Agent AI Framework to Simplify Complex Workflows

2026-04-05
NVIDIANVIDIA
RESEARCH

Nvidia Pivots to Optical Interconnects as Copper Hits Physical Limits, Plans 1,000+ GPU Systems by 2028

2026-04-05
← Back to news
© 2026 BotBeat
AboutPrivacy PolicyTerms of ServiceContact Us