Read NVLink and GPU-to-adapter distance from nvidia-smi topo
The release binary is built without cgo, so without NVML, and NVML was the only source of NVLink peers: on every NVLink machine the portable agent reported none, and tensor-parallel groups were chosen by GPU index. Discovery now reads `nvidia-smi topo -m` after either path, for the NVLink peers (kept from NVML where it gave them), each card's NUMA node, and how far each network adapter is from it - PIX, PXB, PHB, NODE or SYS - which travels to the control plane as nic_paths for placement to use. Both the NICn-with-legend layout of newer drivers and the adapter-named columns of older ones are read. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
This commit is contained in:
1 parent
22bb098247
commit
bf2d171e89
8 files changed
+363
-30
No files matched your search
@@ -38,6 +38,10 @@ type GPU struct {
|
||||
// NVLinkPeers are the indices of GPUs reachable over NVLink. Within a node
|
||||
// a TP group should stay inside one NVLink domain.
|
||||
NVLinkPeers []int `json:"nvlink_peers,omitempty"`
|
||||
// NICPaths is how far each network adapter is from this GPU, by HCA name:
|
||||
// PIX, PXB, PHB, NODE or SYS, as nvidia-smi topo reports it. What decides
|
||||
// which adapter a GPU's share of a multi-node job should go through.
|
||||
NICPaths map[string]string `json:"nic_paths,omitempty"`
|
||||
// Health carries anything that should keep the GPU out of placement.
|
||||
Health GPUHealth `json:"health"`
|
||||
}
|
||||
|
||||
Reference in new issue
Block a user