Heterogeneous AI Computing Virtualization Middleware(Project under CNCF)

ascend cambricon cncf device-plugin gpu-acceleration gpu-management gpu-shareable gpu-virtualization hygon iluvatar kubernetes kubernetes-gpu-cluster metax mthreads nvidia vgpu vgpu-hypervisor
2 Open Issues Need Help Last updated: Jun 25, 2025

Open Issues Need Help

View All on GitHub

AI Summary: Refactor the `pkg/util/nodelock` package in the HAMi Kubernetes scheduler to replace the `Update` verb with `Patch` for updating node annotations, ensuring no behavioral changes and passing existing tests. This involves researching the feasibility of using `Patch`, refactoring the code, and updating the relevant Kubernetes cluster role definition.

Complexity: 3/5
kind/enhancement good first issue

Heterogeneous AI Computing Virtualization Middleware(Project under CNCF)

Go
#ascend#cambricon#cncf#device-plugin#gpu-acceleration#gpu-management#gpu-shareable#gpu-virtualization#hygon#iluvatar#kubernetes#kubernetes-gpu-cluster#metax#mthreads#nvidia#vgpu#vgpu-hypervisor

AI Summary: Refactor the `readFromConfigFile` function in `pkg/device/nvidia/device.go` to improve the handling of configuration overwriting. Specifically, address inconsistencies between `NvidiaConfig` and `DevicePluginConfigs.Nodeconfig` fields, ensuring that all relevant configuration options are explicitly defined and handled in a clear and maintainable way. This involves potentially adding missing fields to `Nodeconfig` and modifying the logic to correctly handle all configuration options.

Complexity: 4/5
kind/enhancement good first issue

Heterogeneous AI Computing Virtualization Middleware(Project under CNCF)

Go
#ascend#cambricon#cncf#device-plugin#gpu-acceleration#gpu-management#gpu-shareable#gpu-virtualization#hygon#iluvatar#kubernetes#kubernetes-gpu-cluster#metax#mthreads#nvidia#vgpu#vgpu-hypervisor