Huawei Computing said its AsyncFlow asynchronous streaming RL decouples rollout (inference) and trainer via a TransferQueue data hub and CheckpointEngine weight engine, addressing long-tail rollout stalls that leave hardware idle and raise costs. In tests on Ascend Atlas 900 A3 SuperPoD liquid-cooled and Atlas 800 A3 A3 air-cooled supernodes, training throughput on long-sequence workloads (prompt 2k → response 16k) rose from 59.3 to 226.8 — a 3.81x increase — while maintaining aligned convergenc

Huawei Computing said its AsyncFlow asynchronous streaming RL decouples rollout (inference) and trainer via a TransferQueue data hub and CheckpointEngine weight engine, addressing long-tail rollout stalls that leave hardware idle and raise costs. In tests on Ascend Atlas 900 A3 SuperPoD liquid-cooled and Atlas 800 A3 A3 air-cooled supernodes, training throughput on long-sequence workloads (prompt 2k → response 16k) rose from 59.3 to 226.8 — a 3.81x increase — while maintaining aligned convergence accuracy. AsyncFlow currently supports the verl framework.