4paradigm / 4paradigm/k8s-vgpu-scheduler
Vgpu的限制问题
Open
- Dominant language
- Go
- Stars
- 595
- Forks
- 100
- PR merge metrics
- No merged PRs in 30d
Description
6月前更新的libvgpu.so。可以工作，在pytorch上工作正常，超出显存大小会正常报错。但是在tensorflow上不正常，显存限制不正常，可以超出切分的大小而不报错。
Contributor guide
No contributing guide indexed for this repository
Research direction
The issue involves the libvgpu.so library and its interaction with TensorFlow. Start by examining the vGPU device plugin code, particularly the memory management and error handling for TensorFlow. Look for tests or examples of memory limit enforcement. The goal is to understand why TensorFlow does not respect the memory limit and to make it fail appropriately when exceeding the allocated vGPU memory.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- kubernetes, pytorch, tensorflow
- Domain
- ai-infra-agents, cloud, devops, infrastructure
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100