class ClusterLaunchControlExample(tilus.Script):
    def __call__(self, ...):
        # try to cancel a pending thread block not dispatch to any SMs for execution
        self.clc.try_cancel(
            cancel_response[producer_stage],   # the shared memory slice int32[4] to store the response
            mbarrier=consumer_mbarriers[producer_stage],   # mbarrier to track the completion
            multicast=True   # whether to multicast the result and mbarrier to other blocks in the block cluster
        )

        # wait the mbarrier
        ...

        # query the result of try_cancel
        # is_valid: whether a block is canceled
        # blockIdx: the block index of the canceled block, when is_valid = True
        is_valid, blockIdx = self.clc.query_response(cancel_response[consumer_stage])