flashinfer.comm.init_tp_moe_finalize_rmsnorm

flashinfer.comm.init_tp_moe_finalize_rmsnorm(group=None) → bool

Collectively initialize the process-wide TP multicast workspace.

group is the exact tensor-parallel group. Its size must exceed one and evenly divide K_EXPERTS. The default process group is used when it is omitted. Availability is reduced across the group, so every rank returns the same verdict.