pub fn step_tp_graph_enabled() -> Result<bool, String>Expand description
CUDA-graph door for the shape-stable TP segments (first increment: the device-routed expert program — per-layer multi-device parents built from per-rank children, launched on the model engine’s stream; zero per-token node updates). Mechanism proven by tp_graph_probe. VALUE-IDENTICAL: the graphs replay exactly the eager kernel/copy sequence.