v0.6.0

• migrate to golangci-v2 @tariq1890
• Bump github.com/stretchr/testify from 1.9.0 to 1.10.0 @dependabot
• Bump github.com/NVIDIA/go-nvml from 0.12.0-6 to 0.12.9-0 @dependabot
• Bump github.com/NVIDIA/go-nvlib from 0.5.0 to 0.7.3 @dependabot
• update go version to 1.24.4 @tariq1890

v0.5.0

What's Changed

• Bump github.com/NVIDIA/go-nvlib to v0.5.0 by @elezar in #42

Full Changelog: v0.4.2...v0.5.0

v0.4.2

What's Changed

• add CI check to validate go module integrity by @tariq1890 in #35
• Bump golangci/golangci-lint-action from 5 to 6 by @dependabot in #37
• Update github.com/NVIDIA/go-nvml to v0.12.0-6 by @elezar in #38

Full Changelog: v0.4.1...v0.4.2

v0.4.1

What's Changed

• Bump github.com/NVIDIA/go-nvml from 0.12.0-4 to 0.12.0-5 by @dependabot in #36

Full Changelog: v0.4.0...v0.4.1

v0.4.0

What's Changed

• Fix bug in NewStaticDGX2Policy by @elezar in #27
• Update go nvml interfaces by @elezar in #32

Full Changelog: v0.3.2...v0.4.0

v0.3.2

What's Changed

• Quick return in error case by @kerthcet in #20
• Fix devicelist filter by @elezar in #23

Full Changelog: v0.3.1...v0.3.2

v0.3.1

What's Changed

• Fix get P2PNVlinkType error by @kerthcet in #16
• Fix construction of linked devices by @elezar in #22

New Contributors

• @kerthcet made their first contribution in #16

Full Changelog: v0.3.0...v0.3.1

v0.3.0

What's Changed

• Add blossom-ci github action by @ArangoGutierrez in #10
• Revert "Add blossom-ci github action" by @ArangoGutierrez in #11
• Update makefile for repo by @elezar in #9
• Migrate to go nvml by @elezar in #13
• Bump golang version by @elezar in #15

New Contributors

• @ArangoGutierrez made their first contribution in #10

Full Changelog: v0.2.3...v0.3.0

Initial release

This is the initial release of the gpuallocator package.

The gpuallocator package provides a generic abstraction for performing GPU allocations independent of the larger system the gpuallocator is integrated with.

The abstractions provided by this package are not meant to do actual "allocation" of GPUs to any specific entity, but rather run the algorithm responsible for deciding which GPUs should be chosen for allocation based on the set of GPUs available in the system and the number of GPUs being requested.

Different policies can be hooked in to run different allocation algorithms depending on the specific needs of the system.