HAMi
HAMi copied to clipboard
demo's error is bash: symbol lookup error: /usr/local/vgpu/libvgpu.so: undefined symbol: cuMemAllocAsync
hi, I need help, when I use demo :
apiVersion: v1
kind: Pod
metadata:
name: gpu-pod
spec:
containers:
- name: ubuntu-container
image: ubuntu:18.04
command: ["bash", "-c", "sleep 86400"]
resources:
limits:
nvidia.com/gpu: 3 # 请求2个vGPUs
nvidia.com/gpumem: 3000 # 每个vGPU申请3000m显存 (可选,整数类型)
nvidia.com/gpucores: 20 # 每个vGPU的算力为30%实际显卡的算力 (可选,整数类型)
it's error for:
bash: symbol lookup error: /usr/local/vgpu/libvgpu.so: undefined symbol: cuMemAllocAsync
which CUDA Version and driver version are you using
hi,CUDA version is 11.0, driver version is 450.80.02
please update to the latest version using "helm repo update" and try again
please update to the latest version using "helm repo update" and try again
thank you,I'll try it.