Zoe
Zoe

vLLM

1 posts

·7 min read

Running LLM inference on K8s: auto-injecting RDMA permissions and GPU-NIC topology affinity

KubernetesRDMAInfiniBand