What changed
NVIDIA released Topograph as an Apache-2.0 toolkit that discovers cloud or on-premises network topology, normalizes it, and publishes scheduler-ready Kubernetes labels, Node Feature Discovery resources, Slinky ConfigMaps, Slurm configuration or JSON. It includes Helm and native-package paths plus simulation utilities for testing without production hardware.
Why it matters
Distributed AI jobs are sensitive to GPU and fabric locality, yet topology data differs across clouds, InfiniBand, Spectrum-X and NVLink domains. A common discovery layer can make placement policy portable, but wrong or stale provider data can steer large workloads into an expensive failure mode.
What to do
Pin a Topograph revision and simulate the target provider first. In a non-production cluster, compare generated topology with an independently collected inventory, exercise scale and node-change events, inspect scheduler placement, health and metrics, and prove rollback before writing labels or Slurm configuration in production.
Caveat
The support matrix reflects upstream main as of 16 September 2026 and varies by provider, engine and configuration. NFD output requires an alpha feature gate, some provider-to-engine combinations are unsupported, and NVIDIA's efficiency narrative was not independently benchmarked here.