==================================================Ascend ============================= test session starts ============================== platform linux -- Python 3.9.19, pytest-6.2.5, py-1.11.0, pluggy-1.5.0 rootdir: /home/jenkins/mindspore/testcases/testcases/tests/st/infer/ops/test_internal_ops, configfile: ../../../../../../../../sault/virtual_test/virtualenv_007/sault/config/pytest.ini plugins: mock-3.14.0, hydra-core-1.3.2, forked-1.6.0, anyio-4.9.0, xdist-1.32.0 collected 8 items test_paged_attention.py [WARNING] ME(168852:281473710628656,MainProcess):2026-01-29-17:37:35.138.110 [mindspore/context.py:1334] For 'context.set_context', the parameter 'device_target' will be deprecated and removed in a future version. Please use the api mindspore.set_device() instead. TotalTime = 0.315229, [21] [bootstrap]: 0.00072008 [type_inference]: 0.236668 [event_method]: 1.467e-05 [auto_monad]: 0.0001131 [graph_reusing]: 4.67998e-06 [inline]: 2.96001e-06 [add_attr]: 0.00840916, [1] [add_attr_with_inline]: 0.0083946, [1] [Cycle 1]: 9.802e-05, [2] [tag_attr]: 2.012e-05 [meta_addattr_fg_expand]: 4.20999e-06 [parallel-infer-symbol]: 4.10998e-06 [pre_auto_parallel]: 4.691e-05 [insert-virtual-dataset]: 2.66e-06 [parallel-infer-symbol-second]: 9.50007e-07 [dataset_repeat_opt]: 2.71999e-06 [pipeline_split]: 2.02001e-06 [optimize]: 0.068314, [53] [py_interpret_to_execute]: 2.861e-05 [rewriter_before_opt_a]: 8.523e-05 [opt_a]: 0.065398, [2] [Cycle 1]: 0.0643722, [45] [expand_dump_flag]: 2.76999e-06 [switch_simplify]: 3.289e-05 [loop_unroll]: 1.92e-05 [a_1]: 0.0624223 [with_stream_mark]: 4.363e-05 [recompute_prepare]: 2.295e-05 [updatestate_depend_eliminate]: 7.25e-06 [updatestate_assign_eliminate]: 5.04e-06 [updatestate_loads_eliminate]: 4.58999e-06 [parameter_eliminate]: 3.24001e-06 [a_2]: 0.00014842 [accelerated_algorithm]: 1.211e-05 [shard]: 2.94001e-06 [meta_shard_fg_expand]: 4.71997e-06 [shard_inline]: 1.01e-05 [merge_send_recv]: 2.379e-05 [auto_parallel]: 1.394e-05 [parallel]: 5.628e-05 [flash_sp]: 2.376e-05 [merge_comm]: 5.50001e-06 [allreduce_fusion]: 4.66002e-06 [matmul_add_comm_reduction]: 1.187e-05 [allreduce_slice_to_reducescatter]: 8.89995e-07 [virtual_shard_identity]: 1.372e-05 [virtual_dataset]: 9.97999e-06 [get_grad_eliminate_]: 9.67999e-06 [virtual_output]: 9.64999e-06 [merge_forward]: 6.17999e-06 [cell_reuse_recompute_pass]: 3.02002e-06 [offload_activation]: 1.193e-05 [cell_reuse_handle_not_recompute_node_pass]: 2.586e-05 [merge_recompute_call_nodes]: 1.84e-06 [before_grad]: 1.453e-05 [set_forward_comm_id_for_comm_node_pass]: 5.29e-06 [meta_fg_expand]: 4.05e-06 [flash_sp_send_recv_attached]: 2.82002e-06 [receive_attached]: 1.093e-05 [after_resolve]: 2.114e-05 [a_after_grad]: 1.565e-05 [renormalize]: 0.00082273 [add_forward_monad_depend]: 7.41999e-06 [auto_monad_grad]: 2.52001e-06 [auto_monad_eliminator]: 2.101e-05 [cse]: 8.483e-05 [a_3]: 7.899e-05 [Cycle 2]: 0.00099371, [45] [expand_dump_flag]: 3.09001e-06 [switch_simplify]: 1.566e-05 [loop_unroll]: 1.034e-05 [a_1]: 0.00023325 [with_stream_mark]: 2.335e-05 [recompute_prepare]: 1.075e-05 [updatestate_depend_eliminate]: 5.69e-06 [updatestate_assign_eliminate]: 4.63001e-06 [updatestate_loads_eliminate]: 4.27e-06 [parameter_eliminate]: 1.83002e-06 [a_2]: 0.00012418 [accelerated_algorithm]: 1.104e-05 [shard]: 2.32999e-06 [meta_shard_fg_expand]: 2.96001e-06 [shard_inline]: 9.57999e-06 [merge_send_recv]: 9.39e-06 [auto_parallel]: 1.006e-05 [parallel]: 7.15e-06 [flash_sp]: 4.07e-06 [merge_comm]: 5.44e-06 [allreduce_fusion]: 4.62998e-06 [matmul_add_comm_reduction]: 1.153e-05 [allreduce_slice_to_reducescatter]: 9.50007e-07 [virtual_shard_identity]: 1.256e-05 [virtual_dataset]: 9.97999e-06 [get_grad_eliminate_]: 9.30001e-06 [virtual_output]: 9.59e-06 [merge_forward]: 5.55001e-06 [cell_reuse_recompute_pass]: 1.74998e-06 [offload_activation]: 1.207e-05 [cell_reuse_handle_not_recompute_node_pass]: 2.145e-05 [merge_recompute_call_nodes]: 1.88002e-06 [before_grad]: 1.322e-05 [set_forward_comm_id_for_comm_node_pass]: 5.80002e-06 [meta_fg_expand]: 3.17002e-06 [flash_sp_send_recv_attached]: 1.47999e-06 [receive_attached]: 2.17001e-06 [after_resolve]: 1.993e-05 [a_after_grad]: 1.471e-05 [renormalize]: 8.9989e-08 [add_forward_monad_depend]: 2.63003e-06 [auto_monad_grad]: 1.50999e-06 [auto_monad_eliminator]: 1.294e-05 [cse]: 3.798e-05 [a_3]: 6.174e-05 [py_interpret_to_execute_after_opt_a]: 2.025e-05 [slice_cell_reuse_recomputed_activation]: 2.19999e-06 [rewriter_after_opt_a]: 8.072e-05 [convert_after_rewriter]: 2.25e-05 [order_py_execute_after_rewriter]: 7.8e-06 [mutable_eliminate]: 0.0007478 [opt_b]: 0.00032913, [1] [Cycle 1]: 0.00031825, [7] [b_1]: 0.00020313 [b_2]: 1.182e-05 [updatestate_depend_eliminate]: 1.122e-05 [updatestate_assign_eliminate]: 4.08001e-06 [updatestate_loads_eliminate]: 4.18001e-06 [renormalize]: 8.10018e-07 [cse]: 4.75e-05 [optimize_parallel_all_gather_comm]: 2.168e-05 [overlap_param_gather]: 5.57001e-06 [cconv]: 3.519e-05 [loop_unroll]: 0.00047692 [opt_after_cconv]: 0.00015631, [1] [Cycle 1]: 0.00014908, [7] [c_1]: 5.096e-05 [parameter_eliminate]: 5.16998e-06 [updatestate_depend_eliminate]: 9.29e-06 [updatestate_assign_eliminate]: 3.95998e-06 [updatestate_loads_eliminate]: 4.09002e-06 [cse]: 3.924e-05 [renormalize]: 9.40025e-07 [remove_dup_value]: 6.853e-05 [tuple_transform]: 0.00010998, [1] [Cycle 1]: 0.00010457, [4] [d_1]: 7.215e-05 [none_parameter_eliminate]: 1.99e-06 [renormalize]: 1.69995e-07 [switch_simplify]: 1.127e-05 [partial_unused_args_eliminate]: 1.77001e-06 [add_recomputation]: 7.074e-05 [cse_after_recomputation]: 3.508e-05, [1] [Cycle 1]: 3.039e-05, [1] [cse]: 2.401e-05 [environ_conv]: 2.213e-05 [swap_dp_allreduce_reducescatter]: 7.61001e-06 [bias_add_comm_swap]: 3.8e-06 [label_micro_interleaved_index]: 5.39e-06 [label_fine_grained_interleaved_index]: 2.54001e-06 [merge_cast_opt]: 1.45001e-06 [slice_recompute_activation]: 2.41e-06 [micro_interleaved_order_control]: 2.17001e-06 [assign_add_opt]: 1.40999e-06 [ForceFp32Comm]: 9.20001e-07 [remove_cast_before_assign_add]: 1.05999e-06 [full_micro_interleaved_order_control]: 2.07999e-06 [reorder_send_recv_between_fp_bp]: 2.62001e-06 [comm_op_add_attrs]: 9.99979e-07 [add_comm_op_reuse_tag]: 1.30999e-06 [interleave_split_concat_branches]: 1.13001e-06 [interleave_parallel_branches]: 9.80013e-07 [overlap_opt_shard_in_pipeline]: 2.416e-05 [overlap_opt_shard_grad_in_pipeline]: 1.67999e-06 [control_data_broadcast_order]: 1.858e-05 [grouped_pairwise_exchange_alltoall]: 1.50001e-06 [offloading_packed_experts]: 5.09e-06 [overlap_recompute_and_grad_model_parallel]: 6.01e-06 [overlap_grad_matmul_and_grad_allreduce]: 1.25001e-06 [overlap_recompute_allgather_and_fa_grad]: 1.32e-06 [overlap_recompute_comm]: 2.73e-06 [overlap_grad_ring_attention]: 5.21002e-06 [overlap_grad_flash_sp]: 2.703e-05 [begin_end_overlap_inline]: 4.80009e-07 [split_matmul_comm_elemetwise]: 2.12999e-06 [split_layernorm_comm]: 1.72001e-06 [handle_group_info]: 9.49978e-07 [symbol_engine_optimizer]: 0.00020806, [1] [Cycle 1]: 0.000203, [6] [build]: 8.814e-05 [elim_shapecalc]: 1.747e-05 [elim_not_effective]: 2.615e-05 [opt_reshape]: 1.121e-05 [fold_const_symbol]: 2.104e-05 [renormalize]: 5.19998e-07 [detach_backward]: 2.21998e-06 [pipeline_parallel_scheduler]: 1.60999e-06 [auto_monad_reorder]: 3.108e-05 [get_jit_bprop_graph]: 2.19001e-06 [rewriter_after_jit_bprop_graph]: 4.94998e-06 [opt_after_jit_grad]: 0.00063407 [validate]: 7.182e-05 Sums bootstrap : 0.000720s : 0.24% type_inference : 0.236668s : 77.42% event_method : 0.000015s : 0.00% auto_monad : 0.000113s : 0.04% graph_reusing : 0.000005s : 0.00% inline : 0.000003s : 0.00% add_attr.add_attr_with_inline.tag_attr : 0.000020s : 0.01% add_attr.add_attr_with_inline.meta_addattr_fg_expand : 0.000004s : 0.00% parallel-infer-symbol : 0.000004s : 0.00% pre_auto_parallel : 0.000047s : 0.02% insert-virtual-dataset : 0.000003s : 0.00% parallel-infer-symbol-second : 0.000001s : 0.00% dataset_repeat_opt : 0.000003s : 0.00% pipeline_split : 0.000002s : 0.00% optimize.py_interpret_to_execute : 0.000029s : 0.01% optimize.rewriter_before_opt_a : 0.000085s : 0.03% optimize.opt_a.expand_dump_flag : 0.000006s : 0.00% optimize.opt_a.switch_simplify : 0.000049s : 0.02% optimize.opt_a.loop_unroll : 0.000030s : 0.01% optimize.opt_a.a_1 : 0.062656s : 20.50% optimize.opt_a.with_stream_mark : 0.000067s : 0.02% optimize.opt_a.recompute_prepare : 0.000034s : 0.01% optimize.opt_a.updatestate_depend_eliminate : 0.000013s : 0.00% optimize.opt_a.updatestate_assign_eliminate : 0.000010s : 0.00% optimize.opt_a.updatestate_loads_eliminate : 0.000009s : 0.00% optimize.opt_a.parameter_eliminate : 0.000005s : 0.00% optimize.opt_a.a_2 : 0.000273s : 0.09% optimize.opt_a.accelerated_algorithm : 0.000023s : 0.01% optimize.opt_a.shard : 0.000005s : 0.00% optimize.opt_a.meta_shard_fg_expand : 0.000008s : 0.00% optimize.opt_a.shard_inline : 0.000020s : 0.01% optimize.opt_a.merge_send_recv : 0.000033s : 0.01% optimize.opt_a.auto_parallel : 0.000024s : 0.01% optimize.opt_a.parallel : 0.000063s : 0.02% optimize.opt_a.flash_sp : 0.000028s : 0.01% optimize.opt_a.merge_comm : 0.000011s : 0.00% optimize.opt_a.allreduce_fusion : 0.000009s : 0.00% optimize.opt_a.matmul_add_comm_reduction : 0.000023s : 0.01% optimize.opt_a.allreduce_slice_to_reducescatter : 0.000002s : 0.00% optimize.opt_a.virtual_shard_identity : 0.000026s : 0.01% optimize.opt_a.virtual_dataset : 0.000020s : 0.01% optimize.opt_a.get_grad_eliminate_ : 0.000019s : 0.01% optimize.opt_a.virtual_output : 0.000019s : 0.01% optimize.opt_a.merge_forward : 0.000012s : 0.00% optimize.opt_a.cell_reuse_recompute_pass : 0.000005s : 0.00% optimize.opt_a.offload_activation : 0.000024s : 0.01% optimize.opt_a.cell_reuse_handle_not_recompute_node_pass : 0.000047s : 0.02% optimize.opt_a.merge_recompute_call_nodes : 0.000004s : 0.00% optimize.opt_a.before_grad : 0.000028s : 0.01% optimize.opt_a.set_forward_comm_id_for_comm_node_pass : 0.000011s : 0.00% optimize.opt_a.meta_fg_expand : 0.000007s : 0.00% optimize.opt_a.flash_sp_send_recv_attached : 0.000004s : 0.00% optimize.opt_a.receive_attached : 0.000013s : 0.00% optimize.opt_a.after_resolve : 0.000041s : 0.01% optimize.opt_a.a_after_grad : 0.000030s : 0.01% optimize.opt_a.renormalize : 0.000823s : 0.27% optimize.opt_a.add_forward_monad_depend : 0.000010s : 0.00% optimize.opt_a.auto_monad_grad : 0.000004s : 0.00% optimize.opt_a.auto_monad_eliminator : 0.000034s : 0.01% optimize.opt_a.cse : 0.000123s : 0.04% optimize.opt_a.a_3 : 0.000141s : 0.05% optimize.py_interpret_to_execute_after_opt_a : 0.000020s : 0.01% optimize.slice_cell_reuse_recomputed_activation : 0.000002s : 0.00% optimize.rewriter_after_opt_a : 0.000081s : 0.03% optimize.convert_after_rewriter : 0.000022s : 0.01% optimize.order_py_execute_after_rewriter : 0.000008s : 0.00% optimize.mutable_eliminate : 0.000748s : 0.24% optimize.opt_b.b_1 : 0.000203s : 0.07% optimize.opt_b.b_2 : 0.000012s : 0.00% optimize.opt_b.updatestate_depend_eliminate : 0.000011s : 0.00% optimize.opt_b.updatestate_assign_eliminate : 0.000004s : 0.00% optimize.opt_b.updatestate_loads_eliminate : 0.000004s : 0.00% optimize.opt_b.renormalize : 0.000001s : 0.00% optimize.opt_b.cse : 0.000047s : 0.02% optimize.optimize_parallel_all_gather_comm : 0.000022s : 0.01% optimize.overlap_param_gather : 0.000006s : 0.00% optimize.cconv : 0.000035s : 0.01% optimize.loop_unroll : 0.000477s : 0.16% optimize.opt_after_cconv.c_1 : 0.000051s : 0.02% optimize.opt_after_cconv.parameter_eliminate : 0.000005s : 0.00% optimize.opt_after_cconv.updatestate_depend_eliminate : 0.000009s : 0.00% optimize.opt_after_cconv.updatestate_assign_eliminate : 0.000004s : 0.00% optimize.opt_after_cconv.updatestate_loads_eliminate : 0.000004s : 0.00% optimize.opt_after_cconv.cse : 0.000039s : 0.01% optimize.opt_after_cconv.renormalize : 0.000001s : 0.00% optimize.remove_dup_value : 0.000069s : 0.02% optimize.tuple_transform.d_1 : 0.000072s : 0.02% optimize.tuple_transform.none_parameter_eliminate : 0.000002s : 0.00% optimize.tuple_transform.renormalize : 0.000000s : 0.00% optimize.tuple_transform.switch_simplify : 0.000011s : 0.00% optimize.partial_unused_args_eliminate : 0.000002s : 0.00% optimize.add_recomputation : 0.000071s : 0.02% optimize.cse_after_recomputation.cse : 0.000024s : 0.01% optimize.environ_conv : 0.000022s : 0.01% optimize.swap_dp_allreduce_reducescatter : 0.000008s : 0.00% optimize.bias_add_comm_swap : 0.000004s : 0.00% optimize.label_micro_interleaved_index : 0.000005s : 0.00% optimize.label_fine_grained_interleaved_index : 0.000003s : 0.00% optimize.merge_cast_opt : 0.000001s : 0.00% optimize.slice_recompute_activation : 0.000002s : 0.00% optimize.micro_interleaved_order_control : 0.000002s : 0.00% optimize.assign_add_opt : 0.000001s : 0.00% optimize.ForceFp32Comm : 0.000001s : 0.00% optimize.remove_cast_before_assign_add : 0.000001s : 0.00% optimize.full_micro_interleaved_order_control : 0.000002s : 0.00% optimize.reorder_send_recv_between_fp_bp : 0.000003s : 0.00% optimize.comm_op_add_attrs : 0.000001s : 0.00% optimize.add_comm_op_reuse_tag : 0.000001s : 0.00% optimize.interleave_split_concat_branches : 0.000001s : 0.00% optimize.interleave_parallel_branches : 0.000001s : 0.00% optimize.overlap_opt_shard_in_pipeline : 0.000024s : 0.01% optimize.overlap_opt_shard_grad_in_pipeline : 0.000002s : 0.00% optimize.control_data_broadcast_order : 0.000019s : 0.01% optimize.grouped_pairwise_exchange_alltoall : 0.000002s : 0.00% optimize.offloading_packed_experts : 0.000005s : 0.00% optimize.overlap_recompute_and_grad_model_parallel : 0.000006s : 0.00% optimize.overlap_grad_matmul_and_grad_allreduce : 0.000001s : 0.00% optimize.overlap_recompute_allgather_and_fa_grad : 0.000001s : 0.00% optimize.overlap_recompute_comm : 0.000003s : 0.00% optimize.overlap_grad_ring_attention : 0.000005s : 0.00% optimize.overlap_grad_flash_sp : 0.000027s : 0.01% optimize.begin_end_overlap_inline : 0.000000s : 0.00% optimize.split_matmul_comm_elemetwise : 0.000002s : 0.00% optimize.split_layernorm_comm : 0.000002s : 0.00% optimize.handle_group_info : 0.000001s : 0.00% optimize.symbol_engine_optimizer.build : 0.000088s : 0.03% optimize.symbol_engine_optimizer.elim_shapecalc : 0.000017s : 0.01% optimize.symbol_engine_optimizer.elim_not_effective : 0.000026s : 0.01% optimize.symbol_engine_optimizer.opt_reshape : 0.000011s : 0.00% optimize.symbol_engine_optimizer.fold_const_symbol : 0.000021s : 0.01% optimize.symbol_engine_optimizer.renormalize : 0.000001s : 0.00% detach_backward : 0.000002s : 0.00% pipeline_parallel_scheduler : 0.000002s : 0.00% auto_monad_reorder : 0.000031s : 0.01% get_jit_bprop_graph : 0.000002s : 0.00% rewriter_after_jit_bprop_graph : 0.000005s : 0.00% opt_after_jit_grad : 0.000634s : 0.21% validate : 0.000072s : 0.02% Time group info: ------[substitution.] 0.000157 33 6.22% : 0.000010s : 2: substitution.elim_not_effective 5.34% : 0.000008s : 2: substitution.fold_const_symbol 5.09% : 0.000008s : 8: substitution.graph_param_transform 70.72% : 0.000111s : 1: substitution.inline 2.58% : 0.000004s : 4: substitution.j_node_and_user_rematch 3.82% : 0.000006s : 4: substitution.remove_not_recompute_node 6.22% : 0.000010s : 12: substitution.replace_old_param ------[type_inference.] 0.236573 2 99.72% : 0.235921s : 1: type_inference.infer 0.28% : 0.000652s : 1: type_inference.specialize ------[replace.] 0.000019 1 100.00% : 0.000019s : 1: replace.inline ------[match.] 0.000110 1 100.00% : 0.000110s : 1: match.inline ------[predicate.] 0.062187 1879 0.00% : 0.000002s : 17: predicate.accumulaten_eliminater 0.00% : 0.000002s : 8: predicate.ad_related_special_op_eliminate 0.00% : 0.000002s : 16: predicate.addn_check_dump 0.00% : 0.000002s : 17: predicate.addn_zero_filter 0.00% : 0.000002s : 17: predicate.adjust_all_reduce_mul_add 0.01% : 0.000006s : 33: predicate.arithmetic_simplify 0.00% : 0.000002s : 17: predicate.cast_eliminate 0.00% : 0.000002s : 16: predicate.check_bprop_eliminate 0.00% : 0.000002s : 16: predicate.compare_switch_simplify 0.00% : 0.000001s : 8: predicate.const_output_eliminate 0.00% : 0.000002s : 16: predicate.depend_value_elim 0.00% : 0.000002s : 17: predicate.dict_get_item_const_eliminator 0.00% : 0.000002s : 17: predicate.dict_get_item_eliminator 0.00% : 0.000002s : 17: predicate.dict_set_item_eliminator 0.01% : 0.000003s : 16: predicate.dumpgradient_eliminate 0.00% : 0.000001s : 8: predicate.elim_not_effective 0.00% : 0.000002s : 8: predicate.elim_shapecalc_of_broadcastargs 0.00% : 0.000003s : 25: predicate.environ_add_const_eliminate 0.00% : 0.000003s : 25: predicate.environ_get_add_eliminate 0.00% : 0.000003s : 25: predicate.environ_get_depend_swap 0.01% : 0.000004s : 41: predicate.environ_get_eliminate 0.00% : 0.000003s : 25: predicate.environ_get_set_eliminate 0.00% : 0.000002s : 18: predicate.exchange_switch_depend_value 0.01% : 0.000005s : 18: predicate.float_depend_g_call 0.00% : 0.000002s : 16: predicate.float_environ_get_switch 0.00% : 0.000003s : 24: predicate.float_tuple_getitem_switch 0.00% : 0.000001s : 8: predicate.fold_const_symbol 0.00% : 0.000002s : 16: predicate.get_grad_eliminate 0.00% : 0.000001s : 8: predicate.graph_param_transform 0.00% : 0.000002s : 16: predicate.incorporate_call 0.00% : 0.000001s : 16: predicate.incorporate_call_switch 0.02% : 0.000012s : 83: predicate.inline 0.00% : 0.000002s : 16: predicate.inline_without_move 0.00% : 0.000001s : 16: predicate.j_node_and_user_rematch 0.00% : 0.000002s : 16: predicate.less_batch_normalization 0.01% : 0.000004s : 33: predicate.list_to_tuple_eliminator_ 0.01% : 0.000006s : 50: predicate.load_eliminater 0.00% : 0.000003s : 8: predicate.loop_unroll_after_grad 0.01% : 0.000004s : 25: predicate.loop_unroll_before_grad 0.01% : 0.000005s : 33: predicate.make_slice_get_slice_eliminator 0.00% : 0.000002s : 16: predicate.merge_addn 0.00% : 0.000002s : 16: predicate.micro_step_allgather_replace 0.00% : 0.000002s : 16: predicate.mini_step_allgather_replace 0.00% : 0.000002s : 17: predicate.minmaximum_grad 0.01% : 0.000004s : 8: predicate.mutable_eliminate 0.00% : 0.000001s : 8: predicate.opt_reshape 0.00% : 0.000001s : 8: predicate.parallel_virtual_node 0.00% : 0.000003s : 18: predicate.partial_defer_inline 0.00% : 0.000003s : 25: predicate.partial_eliminate 0.00% : 0.000002s : 17: predicate.print_const_string_wrapper 0.00% : 0.000002s : 16: predicate.reduce_all_const_elim 0.00% : 0.000003s : 17: predicate.reduce_eliminate 0.01% : 0.000006s : 50: predicate.redundant_stop_gradient_eliminater 0.00% : 0.000002s : 16: predicate.remove_not_recompute_node 0.01% : 0.000004s : 33: predicate.replace_applicator 0.00% : 0.000002s : 16: predicate.replace_old_param 0.00% : 0.000001s : 8: predicate.reset_defer_inline 0.00% : 0.000002s : 17: predicate.reshape_eliminate 0.00% : 0.000002s : 16: predicate.row_tensor_add_zeros_like 0.00% : 0.000001s : 8: predicate.row_tensor_eliminate 0.00% : 0.000002s : 16: predicate.same_eliminate 0.00% : 0.000002s : 16: predicate.set_cell_output_no_recompute 0.00% : 0.000003s : 16: predicate.shard_identity_eliminate 0.00% : 0.000002s : 16: predicate.special_op_eliminate 0.00% : 0.000002s : 16: predicate.specialize_transform 0.01% : 0.000004s : 16: predicate.split_environ_get_set_with_tuple_value 0.00% : 0.000003s : 16: predicate.stack_unstack_eliminate 0.00% : 0.000001s : 8: predicate.switch_call_monad_eliminater 99.61% : 0.061947s : 18: predicate.switch_defer_inline 0.01% : 0.000004s : 34: predicate.switch_layer_defer_inline 0.02% : 0.000010s : 67: predicate.switch_simplify 0.00% : 0.000002s : 17: predicate.tile_eliminate 0.00% : 0.000002s : 17: predicate.transpose_eliminate 0.01% : 0.000004s : 33: predicate.tuple_list_convert_item_index_to_positive 0.01% : 0.000003s : 33: predicate.tuple_list_get_item_const_eliminator 0.01% : 0.000004s : 33: predicate.tuple_list_get_item_depend_reorder 0.01% : 0.000008s : 49: predicate.tuple_list_get_item_eliminator 0.01% : 0.000003s : 33: predicate.tuple_list_get_set_item_eliminator 0.01% : 0.000006s : 49: predicate.tuple_list_set_item_eliminator 0.01% : 0.000004s : 33: predicate.tuple_to_list_eliminator_ 0.01% : 0.000005s : 50: predicate.updatestate_pure_node_eliminater 0.01% : 0.000007s : 66: predicate.updatestate_useless_node_eliminater 0.00% : 0.000001s : 8: predicate.value_based_eliminate 0.00% : 0.000002s : 16: predicate.virtual_dataset_eliminate 0.00% : 0.000002s : 16: predicate.virtual_output_eliminate 0.00% : 0.000001s : 8: predicate.virtual_view_grad_eliminate 0.00% : 0.000002s : 8: predicate.zero_like_fill_zero ------[func_graph_cloner_run.] 0.000580 6 48.69% : 0.000283s : 3: func_graph_cloner_run.FuncGraphClonerGraph 51.31% : 0.000298s : 3: func_graph_cloner_run.FuncGraphSpecializer ------[meta_graph.] 0.000000 0 ------[manager.] 0.000000 0 ------[pynative] 0.000000 0 ------[others.] 0.456414 192 0.00% : 0.000004s : 1: ForceFp32Comm 1.84% : 0.008415s : 1: add_attr 1.84% : 0.008400s : 1: add_attr_with_inline 0.00% : 0.000004s : 1: add_comm_op_reuse_tag 0.02% : 0.000076s : 1: add_recomputation 0.00% : 0.000004s : 1: assign_add_opt 0.03% : 0.000119s : 1: auto_monad 0.01% : 0.000037s : 1: auto_monad_reorder 0.00% : 0.000003s : 1: begin_end_overlap_inline 0.00% : 0.000007s : 1: bias_add_comm_swap 0.16% : 0.000750s : 1: bootstrap 0.01% : 0.000039s : 1: cconv 0.00% : 0.000004s : 1: comm_op_add_attrs 0.00% : 0.000022s : 1: control_data_broadcast_order 0.01% : 0.000027s : 1: convert_after_rewriter 0.01% : 0.000038s : 1: cse_after_recomputation 0.00% : 0.000007s : 1: dataset_repeat_opt 0.00% : 0.000006s : 1: detach_backward 0.01% : 0.000027s : 1: environ_conv 0.00% : 0.000021s : 1: event_method 0.00% : 0.000005s : 1: full_micro_interleaved_order_control 0.00% : 0.000005s : 1: get_jit_bprop_graph 0.00% : 0.000008s : 1: graph_reusing 0.00% : 0.000004s : 1: grouped_pairwise_exchange_alltoall 0.00% : 0.000004s : 1: handle_group_info 0.00% : 0.000006s : 1: inline 0.00% : 0.000006s : 1: insert-virtual-dataset 0.00% : 0.000004s : 1: interleave_parallel_branches 0.00% : 0.000004s : 1: interleave_split_concat_branches 0.00% : 0.000006s : 1: label_fine_grained_interleaved_index 0.00% : 0.000008s : 1: label_micro_interleaved_index 0.11% : 0.000486s : 1: loop_unroll 0.00% : 0.000005s : 1: merge_cast_opt 0.00% : 0.000005s : 1: micro_interleaved_order_control 0.17% : 0.000759s : 1: mutable_eliminate 0.00% : 0.000008s : 1: offloading_packed_experts 0.00% : 0.000019s : 1: opt.transform.loop_unroll_optimizer 0.01% : 0.000026s : 1: opt.transform.mutable_eliminate 13.88% : 0.063344s : 78: opt.transform.opt_a 0.01% : 0.000050s : 1: opt.transform.opt_after_cconv 0.01% : 0.000041s : 1: opt.transform.opt_after_jit_grad 0.04% : 0.000187s : 28: opt.transform.opt_b 0.02% : 0.000081s : 2: opt.transform.opt_trans_graph 0.02% : 0.000072s : 4: opt.transform.symbol_engine_opt 14.33% : 0.065402s : 1: opt_a 0.04% : 0.000160s : 1: opt_after_cconv 0.14% : 0.000646s : 1: opt_after_jit_grad 0.07% : 0.000333s : 1: opt_b 14.97% : 0.068319s : 1: optimize 0.01% : 0.000025s : 1: optimize_parallel_all_gather_comm 0.00% : 0.000011s : 1: order_py_execute_after_rewriter 0.01% : 0.000031s : 1: overlap_grad_flash_sp 0.00% : 0.000004s : 1: overlap_grad_matmul_and_grad_allreduce 0.00% : 0.000008s : 1: overlap_grad_ring_attention 0.00% : 0.000006s : 1: overlap_opt_shard_grad_in_pipeline 0.01% : 0.000028s : 1: overlap_opt_shard_in_pipeline 0.00% : 0.000009s : 1: overlap_param_gather 0.00% : 0.000004s : 1: overlap_recompute_allgather_and_fa_grad 0.00% : 0.000009s : 1: overlap_recompute_and_grad_model_parallel 0.00% : 0.000005s : 1: overlap_recompute_comm 0.00% : 0.000008s : 1: parallel-infer-symbol 0.00% : 0.000004s : 1: parallel-infer-symbol-second 0.00% : 0.000005s : 1: partial_unused_args_eliminate 0.00% : 0.000006s : 1: pipeline_parallel_scheduler 0.00% : 0.000005s : 1: pipeline_split 0.01% : 0.000052s : 1: pre_auto_parallel 0.01% : 0.000033s : 1: py_interpret_to_execute 0.01% : 0.000025s : 1: py_interpret_to_execute_after_opt_a 0.00% : 0.000004s : 1: remove_cast_before_assign_add 0.02% : 0.000074s : 1: remove_dup_value 0.09% : 0.000423s : 1: renormalize.infer 0.08% : 0.000388s : 1: renormalize.specialize 0.00% : 0.000006s : 1: reorder_send_recv_between_fp_bp 0.00% : 0.000009s : 1: rewriter_after_jit_bprop_graph 0.02% : 0.000086s : 1: rewriter_after_opt_a 0.02% : 0.000091s : 1: rewriter_before_opt_a 0.00% : 0.000006s : 1: slice_cell_reuse_recomputed_activation 0.00% : 0.000005s : 1: slice_recompute_activation 0.00% : 0.000004s : 1: split_layernorm_comm 0.00% : 0.000005s : 1: split_matmul_comm_elemetwise 0.00% : 0.000011s : 1: swap_dp_allreduce_reducescatter 0.05% : 0.000211s : 1: symbol_engine_optimizer 0.02% : 0.000113s : 1: tuple_transform 51.86% : 0.236693s : 1: type_inference mki_log delete old file:/home/jenkins/ascend/log/atb/atb_61224_20260129171906.log . [hook] pytest_runtest_teardown:test_paged_attention_mla_combine_cache_norm[0-float16-16] tests/st/infer/ops/test_internal_ops/test_paged_attention.py::test_paged_attention_mla_combine_cache_norm[0-float16-16],max_mem:270.0M [WARNING] ME(168852:281473710628656,MainProcess):2026-01-29-17:38:04.162.635 [mindspore/context.py:1334] For 'context.set_context', the parameter 'device_target' will be deprecated and removed in a future version. Please use the api mindspore.set_device() instead. TotalTime = 0.0860616, [21] [bootstrap]: 0.00091235 [type_inference]: 0.025108 [event_method]: 1.901e-05 [auto_monad]: 6.633e-05 [graph_reusing]: 4.37998e-06 [inline]: 2.84001e-06 [add_attr]: 0.00975702, [1] [add_attr_with_inline]: 0.00973966, [1] [Cycle 1]: 8.654e-05, [2] [tag_attr]: 2.479e-05 [meta_addattr_fg_expand]: 4.67e-06 [parallel-infer-symbol]: 3.93001e-06 [pre_auto_parallel]: 5.278e-05 [insert-virtual-dataset]: 2.44001e-06 [parallel-infer-symbol-second]: 6.69999e-07 [dataset_repeat_opt]: 2.01998e-06 [pipeline_split]: 1.76e-06 [optimize]: 0.0488645, [53] [py_interpret_to_execute]: 5.449e-05 [rewriter_before_opt_a]: 0.0001089 [opt_a]: 0.0138417, [2] [Cycle 1]: 0.0112207, [45] [expand_dump_flag]: 2.98998e-06 [switch_simplify]: 3.533e-05 [loop_unroll]: 1.856e-05 [a_1]: 0.00047657 [with_stream_mark]: 2.882e-05 [recompute_prepare]: 1.429e-05 [updatestate_depend_eliminate]: 6.39999e-06 [updatestate_assign_eliminate]: 4.98001e-06 [updatestate_loads_eliminate]: 4.82e-06 [parameter_eliminate]: 2.03997e-06 [a_2]: 0.00013647 [accelerated_algorithm]: 1.229e-05 [shard]: 3.09999e-06 [meta_shard_fg_expand]: 4.26001e-06 [shard_inline]: 1.496e-05 [merge_send_recv]: 1.257e-05 [auto_parallel]: 1.095e-05 [parallel]: 3.034e-05 [flash_sp]: 1.25e-05 [merge_comm]: 5.29998e-06 [allreduce_fusion]: 4.92999e-06 [matmul_add_comm_reduction]: 1.268e-05 [allreduce_slice_to_reducescatter]: 9.50007e-07 [virtual_shard_identity]: 1.366e-05 [virtual_dataset]: 1.053e-05 [get_grad_eliminate_]: 1.024e-05 [virtual_output]: 1.371e-05 [merge_forward]: 6.02999e-06 [cell_reuse_recompute_pass]: 2.11e-06 [offload_activation]: 1.332e-05 [cell_reuse_handle_not_recompute_node_pass]: 2.175e-05 [merge_recompute_call_nodes]: 1.55999e-06 [before_grad]: 8.237e-05 [set_forward_comm_id_for_comm_node_pass]: 7.01999e-06 [meta_fg_expand]: 5.42999e-06 [flash_sp_send_recv_attached]: 3.95e-06 [receive_attached]: 2.79001e-06 [after_resolve]: 3.111e-05 [a_after_grad]: 2.29e-05 [renormalize]: 0.00951718 [add_forward_monad_depend]: 1.128e-05 [auto_monad_grad]: 3.22002e-06 [auto_monad_eliminator]: 2.811e-05 [cse]: 8.382e-05 [a_3]: 9.75e-05 [Cycle 2]: 0.00259691, [45] [expand_dump_flag]: 2.63e-06 [switch_simplify]: 1.29e-05 [loop_unroll]: 1.071e-05 [a_1]: 0.00160158 [with_stream_mark]: 3.212e-05 [recompute_prepare]: 1.535e-05 [updatestate_depend_eliminate]: 7.3e-06 [updatestate_assign_eliminate]: 4.95999e-06 [updatestate_loads_eliminate]: 5.41002e-06 [parameter_eliminate]: 2.80997e-06 [a_2]: 0.00013108 [accelerated_algorithm]: 1.227e-05 [shard]: 3.01001e-06 [meta_shard_fg_expand]: 2.83998e-06 [shard_inline]: 2.991e-05 [merge_send_recv]: 1.183e-05 [auto_parallel]: 1.179e-05 [parallel]: 9.60001e-06 [flash_sp]: 4.99998e-06 [merge_comm]: 5.56e-06 [allreduce_fusion]: 5.02999e-06 [matmul_add_comm_reduction]: 1.715e-05 [allreduce_slice_to_reducescatter]: 9.30013e-07 [virtual_shard_identity]: 1.226e-05 [virtual_dataset]: 1.095e-05 [get_grad_eliminate_]: 1.052e-05 [virtual_output]: 9.88002e-06 [merge_forward]: 6.76999e-06 [cell_reuse_recompute_pass]: 3.28e-06 [offload_activation]: 1.622e-05 [cell_reuse_handle_not_recompute_node_pass]: 2.707e-05 [merge_recompute_call_nodes]: 1.77999e-06 [before_grad]: 1.691e-05 [set_forward_comm_id_for_comm_node_pass]: 7.15e-06 [meta_fg_expand]: 4.69998e-06 [flash_sp_send_recv_attached]: 1.62999e-06 [receive_attached]: 2.41998e-06 [after_resolve]: 2.888e-05 [a_after_grad]: 1.946e-05 [renormalize]: 6.99947e-08 [add_forward_monad_depend]: 4.03999e-06 [auto_monad_grad]: 2.63e-06 [auto_monad_eliminator]: 1.621e-05 [cse]: 6.667e-05 [a_3]: 6.541e-05 [py_interpret_to_execute_after_opt_a]: 2.742e-05 [slice_cell_reuse_recomputed_activation]: 1.99e-06 [rewriter_after_opt_a]: 6.448e-05 [convert_after_rewriter]: 9.77001e-06 [order_py_execute_after_rewriter]: 6.76e-06 [mutable_eliminate]: 0.0206758 [opt_b]: 0.00116055, [1] [Cycle 1]: 0.00113064, [7] [b_1]: 0.00086267 [b_2]: 2.134e-05 [updatestate_depend_eliminate]: 1.898e-05 [updatestate_assign_eliminate]: 5.20999e-06 [updatestate_loads_eliminate]: 5.38002e-06 [renormalize]: 1.32e-06 [cse]: 0.00011289 [optimize_parallel_all_gather_comm]: 3.198e-05 [overlap_param_gather]: 2.50002e-06 [cconv]: 4.487e-05 [loop_unroll]: 0.00087427 [opt_after_cconv]: 0.00021734, [1] [Cycle 1]: 0.000207, [7] [c_1]: 5.854e-05 [parameter_eliminate]: 7.88999e-06 [updatestate_depend_eliminate]: 1.305e-05 [updatestate_assign_eliminate]: 4.75999e-06 [updatestate_loads_eliminate]: 4.28001e-06 [cse]: 7.378e-05 [renormalize]: 7.59988e-07 [remove_dup_value]: 8.112e-05 [tuple_transform]: 0.0001389, [1] [Cycle 1]: 0.00013258, [4] [d_1]: 8.998e-05 [none_parameter_eliminate]: 2.04e-06 [renormalize]: 1.30007e-07 [switch_simplify]: 1.238e-05 [partial_unused_args_eliminate]: 2.96001e-06 [add_recomputation]: 8.239e-05 [cse_after_recomputation]: 4.657e-05, [1] [Cycle 1]: 4.143e-05, [1] [cse]: 3.094e-05 [environ_conv]: 9.86998e-06 [swap_dp_allreduce_reducescatter]: 7.06999e-06 [bias_add_comm_swap]: 4.10998e-06 [label_micro_interleaved_index]: 6.63e-06 [label_fine_grained_interleaved_index]: 3.14001e-06 [merge_cast_opt]: 1.62001e-06 [slice_recompute_activation]: 2.36e-06 [micro_interleaved_order_control]: 2.51998e-06 [assign_add_opt]: 1.74998e-06 [ForceFp32Comm]: 8.59989e-07 [remove_cast_before_assign_add]: 1.40001e-06 [full_micro_interleaved_order_control]: 2.26e-06 [reorder_send_recv_between_fp_bp]: 2.69999e-06 [comm_op_add_attrs]: 1.10001e-06 [add_comm_op_reuse_tag]: 1.33002e-06 [interleave_split_concat_branches]: 1.97999e-06 [interleave_parallel_branches]: 1.18001e-06 [overlap_opt_shard_in_pipeline]: 1.81e-06 [overlap_opt_shard_grad_in_pipeline]: 1.81e-06 [control_data_broadcast_order]: 1.951e-05 [grouped_pairwise_exchange_alltoall]: 2.02001e-06 [offloading_packed_experts]: 5.50001e-06 [overlap_recompute_and_grad_model_parallel]: 8.36002e-06 [overlap_grad_matmul_and_grad_allreduce]: 1.29e-06 [overlap_recompute_allgather_and_fa_grad]: 1.54e-06 [overlap_recompute_comm]: 5.62001e-06 [overlap_grad_ring_attention]: 5.62999e-06 [overlap_grad_flash_sp]: 2.574e-05 [begin_end_overlap_inline]: 5.09986e-07 [split_matmul_comm_elemetwise]: 2.37999e-06 [split_layernorm_comm]: 1.84998e-06 [handle_group_info]: 1.07998e-06 [symbol_engine_optimizer]: 0.00023183, [1] [Cycle 1]: 0.00022598, [6] [build]: 9.115e-05 [elim_shapecalc]: 2.576e-05 [elim_not_effective]: 2.153e-05 [opt_reshape]: 1.886e-05 [fold_const_symbol]: 2.153e-05 [renormalize]: 2.60014e-07 [detach_backward]: 2.34999e-06 [pipeline_parallel_scheduler]: 1.42e-06 [auto_monad_reorder]: 3.76e-05 [get_jit_bprop_graph]: 2.33002e-06 [rewriter_after_jit_bprop_graph]: 6.54001e-06 [opt_after_jit_grad]: 0.00091847 [validate]: 7.037e-05 Sums bootstrap : 0.000912s : 1.42% type_inference : 0.025108s : 39.18% event_method : 0.000019s : 0.03% auto_monad : 0.000066s : 0.10% graph_reusing : 0.000004s : 0.01% inline : 0.000003s : 0.00% add_attr.add_attr_with_inline.tag_attr : 0.000025s : 0.04% add_attr.add_attr_with_inline.meta_addattr_fg_expand : 0.000005s : 0.01% parallel-infer-symbol : 0.000004s : 0.01% pre_auto_parallel : 0.000053s : 0.08% insert-virtual-dataset : 0.000002s : 0.00% parallel-infer-symbol-second : 0.000001s : 0.00% dataset_repeat_opt : 0.000002s : 0.00% pipeline_split : 0.000002s : 0.00% optimize.py_interpret_to_execute : 0.000054s : 0.09% optimize.rewriter_before_opt_a : 0.000109s : 0.17% optimize.opt_a.expand_dump_flag : 0.000006s : 0.01% optimize.opt_a.switch_simplify : 0.000048s : 0.08% optimize.opt_a.loop_unroll : 0.000029s : 0.05% optimize.opt_a.a_1 : 0.002078s : 3.24% optimize.opt_a.with_stream_mark : 0.000061s : 0.10% optimize.opt_a.recompute_prepare : 0.000030s : 0.05% optimize.opt_a.updatestate_depend_eliminate : 0.000014s : 0.02% optimize.opt_a.updatestate_assign_eliminate : 0.000010s : 0.02% optimize.opt_a.updatestate_loads_eliminate : 0.000010s : 0.02% optimize.opt_a.parameter_eliminate : 0.000005s : 0.01% optimize.opt_a.a_2 : 0.000268s : 0.42% optimize.opt_a.accelerated_algorithm : 0.000025s : 0.04% optimize.opt_a.shard : 0.000006s : 0.01% optimize.opt_a.meta_shard_fg_expand : 0.000007s : 0.01% optimize.opt_a.shard_inline : 0.000045s : 0.07% optimize.opt_a.merge_send_recv : 0.000024s : 0.04% optimize.opt_a.auto_parallel : 0.000023s : 0.04% optimize.opt_a.parallel : 0.000040s : 0.06% optimize.opt_a.flash_sp : 0.000017s : 0.03% optimize.opt_a.merge_comm : 0.000011s : 0.02% optimize.opt_a.allreduce_fusion : 0.000010s : 0.02% optimize.opt_a.matmul_add_comm_reduction : 0.000030s : 0.05% optimize.opt_a.allreduce_slice_to_reducescatter : 0.000002s : 0.00% optimize.opt_a.virtual_shard_identity : 0.000026s : 0.04% optimize.opt_a.virtual_dataset : 0.000021s : 0.03% optimize.opt_a.get_grad_eliminate_ : 0.000021s : 0.03% optimize.opt_a.virtual_output : 0.000024s : 0.04% optimize.opt_a.merge_forward : 0.000013s : 0.02% optimize.opt_a.cell_reuse_recompute_pass : 0.000005s : 0.01% optimize.opt_a.offload_activation : 0.000030s : 0.05% optimize.opt_a.cell_reuse_handle_not_recompute_node_pass : 0.000049s : 0.08% optimize.opt_a.merge_recompute_call_nodes : 0.000003s : 0.01% optimize.opt_a.before_grad : 0.000099s : 0.15% optimize.opt_a.set_forward_comm_id_for_comm_node_pass : 0.000014s : 0.02% optimize.opt_a.meta_fg_expand : 0.000010s : 0.02% optimize.opt_a.flash_sp_send_recv_attached : 0.000006s : 0.01% optimize.opt_a.receive_attached : 0.000005s : 0.01% optimize.opt_a.after_resolve : 0.000060s : 0.09% optimize.opt_a.a_after_grad : 0.000042s : 0.07% optimize.opt_a.renormalize : 0.009517s : 14.85% optimize.opt_a.add_forward_monad_depend : 0.000015s : 0.02% optimize.opt_a.auto_monad_grad : 0.000006s : 0.01% optimize.opt_a.auto_monad_eliminator : 0.000044s : 0.07% optimize.opt_a.cse : 0.000150s : 0.23% optimize.opt_a.a_3 : 0.000163s : 0.25% optimize.py_interpret_to_execute_after_opt_a : 0.000027s : 0.04% optimize.slice_cell_reuse_recomputed_activation : 0.000002s : 0.00% optimize.rewriter_after_opt_a : 0.000064s : 0.10% optimize.convert_after_rewriter : 0.000010s : 0.02% optimize.order_py_execute_after_rewriter : 0.000007s : 0.01% optimize.mutable_eliminate : 0.020676s : 32.27% optimize.opt_b.b_1 : 0.000863s : 1.35% optimize.opt_b.b_2 : 0.000021s : 0.03% optimize.opt_b.updatestate_depend_eliminate : 0.000019s : 0.03% optimize.opt_b.updatestate_assign_eliminate : 0.000005s : 0.01% optimize.opt_b.updatestate_loads_eliminate : 0.000005s : 0.01% optimize.opt_b.renormalize : 0.000001s : 0.00% optimize.opt_b.cse : 0.000113s : 0.18% optimize.optimize_parallel_all_gather_comm : 0.000032s : 0.05% optimize.overlap_param_gather : 0.000003s : 0.00% optimize.cconv : 0.000045s : 0.07% optimize.loop_unroll : 0.000874s : 1.36% optimize.opt_after_cconv.c_1 : 0.000059s : 0.09% optimize.opt_after_cconv.parameter_eliminate : 0.000008s : 0.01% optimize.opt_after_cconv.updatestate_depend_eliminate : 0.000013s : 0.02% optimize.opt_after_cconv.updatestate_assign_eliminate : 0.000005s : 0.01% optimize.opt_after_cconv.updatestate_loads_eliminate : 0.000004s : 0.01% optimize.opt_after_cconv.cse : 0.000074s : 0.12% optimize.opt_after_cconv.renormalize : 0.000001s : 0.00% optimize.remove_dup_value : 0.000081s : 0.13% optimize.tuple_transform.d_1 : 0.000090s : 0.14% optimize.tuple_transform.none_parameter_eliminate : 0.000002s : 0.00% optimize.tuple_transform.renormalize : 0.000000s : 0.00% optimize.tuple_transform.switch_simplify : 0.000012s : 0.02% optimize.partial_unused_args_eliminate : 0.000003s : 0.00% optimize.add_recomputation : 0.000082s : 0.13% optimize.cse_after_recomputation.cse : 0.000031s : 0.05% optimize.environ_conv : 0.000010s : 0.02% optimize.swap_dp_allreduce_reducescatter : 0.000007s : 0.01% optimize.bias_add_comm_swap : 0.000004s : 0.01% optimize.label_micro_interleaved_index : 0.000007s : 0.01% optimize.label_fine_grained_interleaved_index : 0.000003s : 0.00% optimize.merge_cast_opt : 0.000002s : 0.00% optimize.slice_recompute_activation : 0.000002s : 0.00% optimize.micro_interleaved_order_control : 0.000003s : 0.00% optimize.assign_add_opt : 0.000002s : 0.00% optimize.ForceFp32Comm : 0.000001s : 0.00% optimize.remove_cast_before_assign_add : 0.000001s : 0.00% optimize.full_micro_interleaved_order_control : 0.000002s : 0.00% optimize.reorder_send_recv_between_fp_bp : 0.000003s : 0.00% optimize.comm_op_add_attrs : 0.000001s : 0.00% optimize.add_comm_op_reuse_tag : 0.000001s : 0.00% optimize.interleave_split_concat_branches : 0.000002s : 0.00% optimize.interleave_parallel_branches : 0.000001s : 0.00% optimize.overlap_opt_shard_in_pipeline : 0.000002s : 0.00% optimize.overlap_opt_shard_grad_in_pipeline : 0.000002s : 0.00% optimize.control_data_broadcast_order : 0.000020s : 0.03% optimize.grouped_pairwise_exchange_alltoall : 0.000002s : 0.00% optimize.offloading_packed_experts : 0.000006s : 0.01% optimize.overlap_recompute_and_grad_model_parallel : 0.000008s : 0.01% optimize.overlap_grad_matmul_and_grad_allreduce : 0.000001s : 0.00% optimize.overlap_recompute_allgather_and_fa_grad : 0.000002s : 0.00% optimize.overlap_recompute_comm : 0.000006s : 0.01% optimize.overlap_grad_ring_attention : 0.000006s : 0.01% optimize.overlap_grad_flash_sp : 0.000026s : 0.04% optimize.begin_end_overlap_inline : 0.000001s : 0.00% optimize.split_matmul_comm_elemetwise : 0.000002s : 0.00% optimize.split_layernorm_comm : 0.000002s : 0.00% optimize.handle_group_info : 0.000001s : 0.00% optimize.symbol_engine_optimizer.build : 0.000091s : 0.14% optimize.symbol_engine_optimizer.elim_shapecalc : 0.000026s : 0.04% optimize.symbol_engine_optimizer.elim_not_effective : 0.000022s : 0.03% optimize.symbol_engine_optimizer.opt_reshape : 0.000019s : 0.03% optimize.symbol_engine_optimizer.fold_const_symbol : 0.000022s : 0.03% optimize.symbol_engine_optimizer.renormalize : 0.000000s : 0.00% detach_backward : 0.000002s : 0.00% pipeline_parallel_scheduler : 0.000001s : 0.00% auto_monad_reorder : 0.000038s : 0.06% get_jit_bprop_graph : 0.000002s : 0.00% rewriter_after_jit_bprop_graph : 0.000007s : 0.01% opt_after_jit_grad : 0.000918s : 1.43% validate : 0.000070s : 0.11% Time group info: ------[substitution.] 0.000208 33 1.77% : 0.000004s : 2: substitution.elim_not_effective 4.14% : 0.000009s : 2: substitution.fold_const_symbol 4.13% : 0.000009s : 8: substitution.graph_param_transform 64.85% : 0.000135s : 1: substitution.inline 17.03% : 0.000035s : 4: substitution.j_node_and_user_rematch 2.48% : 0.000005s : 4: substitution.remove_not_recompute_node 5.59% : 0.000012s : 12: substitution.replace_old_param ------[type_inference.] 0.024978 2 95.30% : 0.023805s : 1: type_inference.infer 4.70% : 0.001174s : 1: type_inference.specialize ------[replace.] 0.000025 1 100.00% : 0.000025s : 1: replace.inline ------[match.] 0.000134 1 100.00% : 0.000134s : 1: match.inline ------[predicate.] 0.000807 1879 0.30% : 0.000002s : 17: predicate.accumulaten_eliminater 0.47% : 0.000004s : 8: predicate.ad_related_special_op_eliminate 0.22% : 0.000002s : 16: predicate.addn_check_dump 0.28% : 0.000002s : 17: predicate.addn_zero_filter 0.22% : 0.000002s : 17: predicate.adjust_all_reduce_mul_add 0.63% : 0.000005s : 33: predicate.arithmetic_simplify 0.29% : 0.000002s : 17: predicate.cast_eliminate 0.27% : 0.000002s : 16: predicate.check_bprop_eliminate 0.20% : 0.000002s : 16: predicate.compare_switch_simplify 0.07% : 0.000001s : 8: predicate.const_output_eliminate 0.21% : 0.000002s : 16: predicate.depend_value_elim 0.24% : 0.000002s : 17: predicate.dict_get_item_const_eliminator 0.28% : 0.000002s : 17: predicate.dict_get_item_eliminator 0.24% : 0.000002s : 17: predicate.dict_set_item_eliminator 0.40% : 0.000003s : 16: predicate.dumpgradient_eliminate 0.12% : 0.000001s : 8: predicate.elim_not_effective 0.19% : 0.000002s : 8: predicate.elim_shapecalc_of_broadcastargs 0.34% : 0.000003s : 25: predicate.environ_add_const_eliminate 0.32% : 0.000003s : 25: predicate.environ_get_add_eliminate 0.33% : 0.000003s : 25: predicate.environ_get_depend_swap 0.54% : 0.000004s : 41: predicate.environ_get_eliminate 0.32% : 0.000003s : 25: predicate.environ_get_set_eliminate 0.26% : 0.000002s : 18: predicate.exchange_switch_depend_value 0.50% : 0.000004s : 18: predicate.float_depend_g_call 0.21% : 0.000002s : 16: predicate.float_environ_get_switch 0.64% : 0.000005s : 24: predicate.float_tuple_getitem_switch 0.08% : 0.000001s : 8: predicate.fold_const_symbol 0.25% : 0.000002s : 16: predicate.get_grad_eliminate 0.12% : 0.000001s : 8: predicate.graph_param_transform 0.20% : 0.000002s : 16: predicate.incorporate_call 0.18% : 0.000001s : 16: predicate.incorporate_call_switch 1.65% : 0.000013s : 83: predicate.inline 0.26% : 0.000002s : 16: predicate.inline_without_move 0.18% : 0.000001s : 16: predicate.j_node_and_user_rematch 0.36% : 0.000003s : 16: predicate.less_batch_normalization 0.59% : 0.000005s : 33: predicate.list_to_tuple_eliminator_ 0.75% : 0.000006s : 50: predicate.load_eliminater 0.42% : 0.000003s : 8: predicate.loop_unroll_after_grad 0.37% : 0.000003s : 25: predicate.loop_unroll_before_grad 67.29% : 0.000543s : 33: predicate.make_slice_get_slice_eliminator 0.23% : 0.000002s : 16: predicate.merge_addn 0.25% : 0.000002s : 16: predicate.micro_step_allgather_replace 0.25% : 0.000002s : 16: predicate.mini_step_allgather_replace 0.20% : 0.000002s : 17: predicate.minmaximum_grad 1.01% : 0.000008s : 8: predicate.mutable_eliminate 0.14% : 0.000001s : 8: predicate.opt_reshape 0.17% : 0.000001s : 8: predicate.parallel_virtual_node 0.32% : 0.000003s : 18: predicate.partial_defer_inline 0.38% : 0.000003s : 25: predicate.partial_eliminate 0.26% : 0.000002s : 17: predicate.print_const_string_wrapper 0.24% : 0.000002s : 16: predicate.reduce_all_const_elim 0.42% : 0.000003s : 17: predicate.reduce_eliminate 0.74% : 0.000006s : 50: predicate.redundant_stop_gradient_eliminater 0.21% : 0.000002s : 16: predicate.remove_not_recompute_node 0.47% : 0.000004s : 33: predicate.replace_applicator 0.22% : 0.000002s : 16: predicate.replace_old_param 0.21% : 0.000002s : 8: predicate.reset_defer_inline 0.26% : 0.000002s : 17: predicate.reshape_eliminate 0.25% : 0.000002s : 16: predicate.row_tensor_add_zeros_like 0.18% : 0.000001s : 8: predicate.row_tensor_eliminate 0.30% : 0.000002s : 16: predicate.same_eliminate 0.18% : 0.000001s : 16: predicate.set_cell_output_no_recompute 0.30% : 0.000002s : 16: predicate.shard_identity_eliminate 0.35% : 0.000003s : 16: predicate.special_op_eliminate 0.31% : 0.000002s : 16: predicate.specialize_transform 0.34% : 0.000003s : 16: predicate.split_environ_get_set_with_tuple_value 0.32% : 0.000003s : 16: predicate.stack_unstack_eliminate 0.14% : 0.000001s : 8: predicate.switch_call_monad_eliminater 0.26% : 0.000002s : 18: predicate.switch_defer_inline 0.51% : 0.000004s : 34: predicate.switch_layer_defer_inline 1.09% : 0.000009s : 67: predicate.switch_simplify 0.26% : 0.000002s : 17: predicate.tile_eliminate 0.23% : 0.000002s : 17: predicate.transpose_eliminate 0.51% : 0.000004s : 33: predicate.tuple_list_convert_item_index_to_positive 0.56% : 0.000005s : 33: predicate.tuple_list_get_item_const_eliminator 0.52% : 0.000004s : 33: predicate.tuple_list_get_item_depend_reorder 1.18% : 0.000010s : 49: predicate.tuple_list_get_item_eliminator 0.49% : 0.000004s : 33: predicate.tuple_list_get_set_item_eliminator 0.87% : 0.000007s : 49: predicate.tuple_list_set_item_eliminator 0.52% : 0.000004s : 33: predicate.tuple_to_list_eliminator_ 0.63% : 0.000005s : 50: predicate.updatestate_pure_node_eliminater 1.07% : 0.000009s : 66: predicate.updatestate_useless_node_eliminater 0.20% : 0.000002s : 8: predicate.value_based_eliminate 0.27% : 0.000002s : 16: predicate.virtual_dataset_eliminate 0.26% : 0.000002s : 16: predicate.virtual_output_eliminate 0.13% : 0.000001s : 8: predicate.virtual_view_grad_eliminate 0.95% : 0.000008s : 8: predicate.zero_like_fill_zero ------[func_graph_cloner_run.] 0.001197 6 44.70% : 0.000535s : 3: func_graph_cloner_run.FuncGraphClonerGraph 55.30% : 0.000662s : 3: func_graph_cloner_run.FuncGraphSpecializer ------[meta_graph.] 0.000000 0 ------[manager.] 0.000000 0 ------[pynative] 0.000000 0 ------[others.] 0.158097 192 0.00% : 0.000004s : 1: ForceFp32Comm 6.18% : 0.009765s : 1: add_attr 6.16% : 0.009744s : 1: add_attr_with_inline 0.00% : 0.000007s : 1: add_comm_op_reuse_tag 0.06% : 0.000088s : 1: add_recomputation 0.00% : 0.000005s : 1: assign_add_opt 0.05% : 0.000072s : 1: auto_monad 0.03% : 0.000042s : 1: auto_monad_reorder 0.00% : 0.000003s : 1: begin_end_overlap_inline 0.00% : 0.000007s : 1: bias_add_comm_swap 0.60% : 0.000953s : 1: bootstrap 0.03% : 0.000050s : 1: cconv 0.00% : 0.000004s : 1: comm_op_add_attrs 0.02% : 0.000026s : 1: control_data_broadcast_order 0.01% : 0.000016s : 1: convert_after_rewriter 0.03% : 0.000050s : 1: cse_after_recomputation 0.00% : 0.000005s : 1: dataset_repeat_opt 0.00% : 0.000006s : 1: detach_backward 0.01% : 0.000013s : 1: environ_conv 0.02% : 0.000027s : 1: event_method 0.00% : 0.000006s : 1: full_micro_interleaved_order_control 0.00% : 0.000008s : 1: get_jit_bprop_graph 0.01% : 0.000008s : 1: graph_reusing 0.00% : 0.000006s : 1: grouped_pairwise_exchange_alltoall 0.00% : 0.000007s : 1: handle_group_info 0.00% : 0.000006s : 1: inline 0.01% : 0.000009s : 1: insert-virtual-dataset 0.00% : 0.000005s : 1: interleave_parallel_branches 0.00% : 0.000005s : 1: interleave_split_concat_branches 0.00% : 0.000007s : 1: label_fine_grained_interleaved_index 0.01% : 0.000010s : 1: label_micro_interleaved_index 0.57% : 0.000896s : 1: loop_unroll 0.00% : 0.000005s : 1: merge_cast_opt 0.00% : 0.000005s : 1: micro_interleaved_order_control 19.83% : 0.031355s : 1: mutable_eliminate 0.01% : 0.000012s : 1: offloading_packed_experts 0.02% : 0.000035s : 1: opt.transform.loop_unroll_optimizer 0.04% : 0.000056s : 1: opt.transform.mutable_eliminate 1.84% : 0.002914s : 78: opt.transform.opt_a 0.04% : 0.000057s : 1: opt.transform.opt_after_cconv 0.04% : 0.000068s : 1: opt.transform.opt_after_jit_grad 0.51% : 0.000812s : 28: opt.transform.opt_b 0.06% : 0.000099s : 2: opt.transform.opt_trans_graph 0.05% : 0.000083s : 4: opt.transform.symbol_engine_opt 8.76% : 0.013849s : 1: opt_a 0.14% : 0.000222s : 1: opt_after_cconv 0.59% : 0.000939s : 1: opt_after_jit_grad 0.74% : 0.001172s : 1: opt_b 30.91% : 0.048872s : 1: optimize 0.02% : 0.000037s : 1: optimize_parallel_all_gather_comm 0.01% : 0.000010s : 1: order_py_execute_after_rewriter 0.02% : 0.000032s : 1: overlap_grad_flash_sp 0.00% : 0.000005s : 1: overlap_grad_matmul_and_grad_allreduce 0.01% : 0.000013s : 1: overlap_grad_ring_attention 0.00% : 0.000006s : 1: overlap_opt_shard_grad_in_pipeline 0.00% : 0.000005s : 1: overlap_opt_shard_in_pipeline 0.01% : 0.000009s : 1: overlap_param_gather 0.00% : 0.000005s : 1: overlap_recompute_allgather_and_fa_grad 0.01% : 0.000012s : 1: overlap_recompute_and_grad_model_parallel 0.01% : 0.000009s : 1: overlap_recompute_comm 0.01% : 0.000008s : 1: parallel-infer-symbol 0.00% : 0.000006s : 1: parallel-infer-symbol-second 0.00% : 0.000007s : 1: partial_unused_args_eliminate 0.01% : 0.000008s : 1: pipeline_parallel_scheduler 0.00% : 0.000005s : 1: pipeline_split 0.04% : 0.000060s : 1: pre_auto_parallel 0.04% : 0.000062s : 1: py_interpret_to_execute 0.02% : 0.000032s : 1: py_interpret_to_execute_after_opt_a 0.01% : 0.000008s : 1: remove_cast_before_assign_add 0.06% : 0.000093s : 1: remove_dup_value 5.55% : 0.008777s : 1: renormalize.infer 0.45% : 0.000716s : 1: renormalize.specialize 0.00% : 0.000006s : 1: reorder_send_recv_between_fp_bp 0.01% : 0.000010s : 1: rewriter_after_jit_bprop_graph 0.04% : 0.000069s : 1: rewriter_after_opt_a 0.08% : 0.000120s : 1: rewriter_before_opt_a 0.00% : 0.000005s : 1: slice_cell_reuse_recomputed_activation 0.00% : 0.000005s : 1: slice_recompute_activation 0.00% : 0.000005s : 1: split_layernorm_comm 0.00% : 0.000006s : 1: split_matmul_comm_elemetwise 0.01% : 0.000014s : 1: swap_dp_allreduce_reducescatter 0.15% : 0.000235s : 1: symbol_engine_optimizer 0.09% : 0.000142s : 1: tuple_transform 15.90% : 0.025142s : 1: type_inference . [hook] pytest_runtest_teardown:test_paged_attention_mla_combine_cache_norm[0-float16-128] tests/st/infer/ops/test_internal_ops/test_paged_attention.py::test_paged_attention_mla_combine_cache_norm[0-float16-128],max_mem:270.0M [WARNING] ME(168852:281473710628656,MainProcess):2026-01-29-17:38:08.391.948 [mindspore/context.py:1334] For 'context.set_context', the parameter 'device_target' will be deprecated and removed in a future version. Please use the api mindspore.set_device() instead. TotalTime = 0.100465, [21] [bootstrap]: 0.00084185 [type_inference]: 0.00743623 [event_method]: 1.788e-05 [auto_monad]: 6.154e-05 [graph_reusing]: 5.02e-06 [inline]: 2.93e-06 [add_attr]: 0.0306574, [1] [add_attr_with_inline]: 0.0306395, [1] [Cycle 1]: 8.211e-05, [2] [tag_attr]: 2.582e-05 [meta_addattr_fg_expand]: 4.12003e-06 [parallel-infer-symbol]: 3.98999e-06 [pre_auto_parallel]: 4.589e-05 [insert-virtual-dataset]: 2.48e-06 [parallel-infer-symbol-second]: 5.79981e-07 [dataset_repeat_opt]: 1.94999e-06 [pipeline_split]: 1.53002e-06 [optimize]: 0.0602385, [53] [py_interpret_to_execute]: 5.886e-05 [rewriter_before_opt_a]: 9.971e-05 [opt_a]: 0.0567301, [2] [Cycle 1]: 0.0554599, [45] [expand_dump_flag]: 3.41999e-06 [switch_simplify]: 2.97e-05 [loop_unroll]: 1.846e-05 [a_1]: 0.052843 [with_stream_mark]: 5.483e-05 [recompute_prepare]: 2.701e-05 [updatestate_depend_eliminate]: 7.88001e-06 [updatestate_assign_eliminate]: 6.04999e-06 [updatestate_loads_eliminate]: 5.33002e-06 [parameter_eliminate]: 3.66999e-06 [a_2]: 0.00016463 [accelerated_algorithm]: 1.461e-05 [shard]: 3.59002e-06 [meta_shard_fg_expand]: 6.91001e-06 [shard_inline]: 1.04e-05 [merge_send_recv]: 1.286e-05 [auto_parallel]: 1.373e-05 [parallel]: 3.378e-05 [flash_sp]: 1.683e-05 [merge_comm]: 5.08002e-06 [allreduce_fusion]: 4.97999e-06 [matmul_add_comm_reduction]: 1.349e-05 [allreduce_slice_to_reducescatter]: 1.41002e-06 [virtual_shard_identity]: 1.565e-05 [virtual_dataset]: 1.222e-05 [get_grad_eliminate_]: 1.196e-05 [virtual_output]: 1.485e-05 [merge_forward]: 6.89999e-06 [cell_reuse_recompute_pass]: 3.58e-06 [offload_activation]: 1.508e-05 [cell_reuse_handle_not_recompute_node_pass]: 3.207e-05 [merge_recompute_call_nodes]: 1.50999e-06 [before_grad]: 1.808e-05 [set_forward_comm_id_for_comm_node_pass]: 6.46999e-06 [meta_fg_expand]: 4.80001e-06 [flash_sp_send_recv_attached]: 2.66e-06 [receive_attached]: 2.51998e-06 [after_resolve]: 2.712e-05 [a_after_grad]: 2.115e-05 [renormalize]: 0.00124187 [add_forward_monad_depend]: 1.313e-05 [auto_monad_grad]: 2.96999e-06 [auto_monad_eliminator]: 2.511e-05 [cse]: 8.639e-05 [a_3]: 9.118e-05 [Cycle 2]: 0.00125369, [45] [expand_dump_flag]: 3.43e-06 [switch_simplify]: 1.392e-05 [loop_unroll]: 1.073e-05 [a_1]: 0.00025026 [with_stream_mark]: 2.902e-05 [recompute_prepare]: 1.185e-05 [updatestate_depend_eliminate]: 8.32e-06 [updatestate_assign_eliminate]: 5.29998e-06 [updatestate_loads_eliminate]: 5.84e-06 [parameter_eliminate]: 1.99e-06 [a_2]: 0.00012828 [accelerated_algorithm]: 1.132e-05 [shard]: 4.53001e-06 [meta_shard_fg_expand]: 3.33e-06 [shard_inline]: 1.264e-05 [merge_send_recv]: 1.079e-05 [auto_parallel]: 1.244e-05 [parallel]: 1.201e-05 [flash_sp]: 4.63999e-06 [merge_comm]: 5.79999e-06 [allreduce_fusion]: 4.60999e-06 [matmul_add_comm_reduction]: 1.191e-05 [allreduce_slice_to_reducescatter]: 7.99977e-07 [virtual_shard_identity]: 1.24e-05 [virtual_dataset]: 1.044e-05 [get_grad_eliminate_]: 1.222e-05 [virtual_output]: 1.037e-05 [merge_forward]: 6.26e-06 [cell_reuse_recompute_pass]: 3.36999e-06 [offload_activation]: 1.571e-05 [cell_reuse_handle_not_recompute_node_pass]: 2.352e-05 [merge_recompute_call_nodes]: 1.81e-06 [before_grad]: 1.843e-05 [set_forward_comm_id_for_comm_node_pass]: 6.36e-06 [meta_fg_expand]: 4.38999e-06 [flash_sp_send_recv_attached]: 2.27001e-06 [receive_attached]: 2.04999e-06 [after_resolve]: 2.467e-05 [a_after_grad]: 1.622e-05 [renormalize]: 7.00238e-08 [add_forward_monad_depend]: 1.895e-05 [auto_monad_grad]: 2.44999e-06 [auto_monad_eliminator]: 1.669e-05 [cse]: 6.669e-05 [a_3]: 6.83e-05 [py_interpret_to_execute_after_opt_a]: 3.037e-05 [slice_cell_reuse_recomputed_activation]: 2.14e-06 [rewriter_after_opt_a]: 7.088e-05 [convert_after_rewriter]: 1.05e-05 [order_py_execute_after_rewriter]: 7.65998e-06 [mutable_eliminate]: 0.00087263 [opt_b]: 0.00040191, [1] [Cycle 1]: 0.00039093, [7] [b_1]: 0.00021794 [b_2]: 1.332e-05 [updatestate_depend_eliminate]: 1.269e-05 [updatestate_assign_eliminate]: 5.04e-06 [updatestate_loads_eliminate]: 5.05999e-06 [renormalize]: 7.09988e-07 [cse]: 7.536e-05 [optimize_parallel_all_gather_comm]: 2.554e-05 [overlap_param_gather]: 2.14e-06 [cconv]: 4.602e-05 [loop_unroll]: 0.0006156 [opt_after_cconv]: 0.00019209, [1] [Cycle 1]: 0.00018355, [7] [c_1]: 5.158e-05 [parameter_eliminate]: 5.60001e-06 [updatestate_depend_eliminate]: 8.90999e-06 [updatestate_assign_eliminate]: 4.41002e-06 [updatestate_loads_eliminate]: 4.87998e-06 [cse]: 5.897e-05 [renormalize]: 7.2e-07 [remove_dup_value]: 8.217e-05 [tuple_transform]: 0.00014606, [1] [Cycle 1]: 0.00013786, [4] [d_1]: 9.301e-05 [none_parameter_eliminate]: 2.51e-06 [renormalize]: 1.50001e-07 [switch_simplify]: 1.213e-05 [partial_unused_args_eliminate]: 1.84998e-06 [add_recomputation]: 7.964e-05 [cse_after_recomputation]: 4.343e-05, [1] [Cycle 1]: 3.831e-05, [1] [cse]: 2.876e-05 [environ_conv]: 1.003e-05 [swap_dp_allreduce_reducescatter]: 1.034e-05 [bias_add_comm_swap]: 4.13999e-06 [label_micro_interleaved_index]: 6.49001e-06 [label_fine_grained_interleaved_index]: 2.66e-06 [merge_cast_opt]: 1.99999e-06 [slice_recompute_activation]: 2.19999e-06 [micro_interleaved_order_control]: 2.85998e-06 [assign_add_opt]: 1.57001e-06 [ForceFp32Comm]: 8.59989e-07 [remove_cast_before_assign_add]: 1.07998e-06 [full_micro_interleaved_order_control]: 2.44001e-06 [reorder_send_recv_between_fp_bp]: 2.76e-06 [comm_op_add_attrs]: 1.35001e-06 [add_comm_op_reuse_tag]: 1.00001e-06 [interleave_split_concat_branches]: 2.07001e-06 [interleave_parallel_branches]: 1.22999e-06 [overlap_opt_shard_in_pipeline]: 1.89e-06 [overlap_opt_shard_grad_in_pipeline]: 1.79e-06 [control_data_broadcast_order]: 2.013e-05 [grouped_pairwise_exchange_alltoall]: 1.74e-06 [offloading_packed_experts]: 5.91998e-06 [overlap_recompute_and_grad_model_parallel]: 6.71999e-06 [overlap_grad_matmul_and_grad_allreduce]: 1.31002e-06 [overlap_recompute_allgather_and_fa_grad]: 1.60999e-06 [overlap_recompute_comm]: 2.42001e-06 [overlap_grad_ring_attention]: 6.07001e-06 [overlap_grad_flash_sp]: 2.68e-05 [begin_end_overlap_inline]: 5.60016e-07 [split_matmul_comm_elemetwise]: 2.17999e-06 [split_layernorm_comm]: 1.89999e-06 [handle_group_info]: 1.06002e-06 [symbol_engine_optimizer]: 0.00025619, [1] [Cycle 1]: 0.00025053, [6] [build]: 9.871e-05 [elim_shapecalc]: 2.675e-05 [elim_not_effective]: 2.727e-05 [opt_reshape]: 1.699e-05 [fold_const_symbol]: 2.426e-05 [renormalize]: 1.8999e-07 [detach_backward]: 2.68e-06 [pipeline_parallel_scheduler]: 1.66e-06 [auto_monad_reorder]: 3.484e-05 [get_jit_bprop_graph]: 1.91e-06 [rewriter_after_jit_bprop_graph]: 7.16001e-06 [opt_after_jit_grad]: 0.00084596 [validate]: 7.017e-05 Sums bootstrap : 0.000842s : 1.23% type_inference : 0.007436s : 10.90% event_method : 0.000018s : 0.03% auto_monad : 0.000062s : 0.09% graph_reusing : 0.000005s : 0.01% inline : 0.000003s : 0.00% add_attr.add_attr_with_inline.tag_attr : 0.000026s : 0.04% add_attr.add_attr_with_inline.meta_addattr_fg_expand : 0.000004s : 0.01% parallel-infer-symbol : 0.000004s : 0.01% pre_auto_parallel : 0.000046s : 0.07% insert-virtual-dataset : 0.000002s : 0.00% parallel-infer-symbol-second : 0.000001s : 0.00% dataset_repeat_opt : 0.000002s : 0.00% pipeline_split : 0.000002s : 0.00% optimize.py_interpret_to_execute : 0.000059s : 0.09% optimize.rewriter_before_opt_a : 0.000100s : 0.15% optimize.opt_a.expand_dump_flag : 0.000007s : 0.01% optimize.opt_a.switch_simplify : 0.000044s : 0.06% optimize.opt_a.loop_unroll : 0.000029s : 0.04% optimize.opt_a.a_1 : 0.053093s : 77.83% optimize.opt_a.with_stream_mark : 0.000084s : 0.12% optimize.opt_a.recompute_prepare : 0.000039s : 0.06% optimize.opt_a.updatestate_depend_eliminate : 0.000016s : 0.02% optimize.opt_a.updatestate_assign_eliminate : 0.000011s : 0.02% optimize.opt_a.updatestate_loads_eliminate : 0.000011s : 0.02% optimize.opt_a.parameter_eliminate : 0.000006s : 0.01% optimize.opt_a.a_2 : 0.000293s : 0.43% optimize.opt_a.accelerated_algorithm : 0.000026s : 0.04% optimize.opt_a.shard : 0.000008s : 0.01% optimize.opt_a.meta_shard_fg_expand : 0.000010s : 0.02% optimize.opt_a.shard_inline : 0.000023s : 0.03% optimize.opt_a.merge_send_recv : 0.000024s : 0.03% optimize.opt_a.auto_parallel : 0.000026s : 0.04% optimize.opt_a.parallel : 0.000046s : 0.07% optimize.opt_a.flash_sp : 0.000021s : 0.03% optimize.opt_a.merge_comm : 0.000011s : 0.02% optimize.opt_a.allreduce_fusion : 0.000010s : 0.01% optimize.opt_a.matmul_add_comm_reduction : 0.000025s : 0.04% optimize.opt_a.allreduce_slice_to_reducescatter : 0.000002s : 0.00% optimize.opt_a.virtual_shard_identity : 0.000028s : 0.04% optimize.opt_a.virtual_dataset : 0.000023s : 0.03% optimize.opt_a.get_grad_eliminate_ : 0.000024s : 0.04% optimize.opt_a.virtual_output : 0.000025s : 0.04% optimize.opt_a.merge_forward : 0.000013s : 0.02% optimize.opt_a.cell_reuse_recompute_pass : 0.000007s : 0.01% optimize.opt_a.offload_activation : 0.000031s : 0.05% optimize.opt_a.cell_reuse_handle_not_recompute_node_pass : 0.000056s : 0.08% optimize.opt_a.merge_recompute_call_nodes : 0.000003s : 0.00% optimize.opt_a.before_grad : 0.000037s : 0.05% optimize.opt_a.set_forward_comm_id_for_comm_node_pass : 0.000013s : 0.02% optimize.opt_a.meta_fg_expand : 0.000009s : 0.01% optimize.opt_a.flash_sp_send_recv_attached : 0.000005s : 0.01% optimize.opt_a.receive_attached : 0.000005s : 0.01% optimize.opt_a.after_resolve : 0.000052s : 0.08% optimize.opt_a.a_after_grad : 0.000037s : 0.05% optimize.opt_a.renormalize : 0.001242s : 1.82% optimize.opt_a.add_forward_monad_depend : 0.000032s : 0.05% optimize.opt_a.auto_monad_grad : 0.000005s : 0.01% optimize.opt_a.auto_monad_eliminator : 0.000042s : 0.06% optimize.opt_a.cse : 0.000153s : 0.22% optimize.opt_a.a_3 : 0.000159s : 0.23% optimize.py_interpret_to_execute_after_opt_a : 0.000030s : 0.04% optimize.slice_cell_reuse_recomputed_activation : 0.000002s : 0.00% optimize.rewriter_after_opt_a : 0.000071s : 0.10% optimize.convert_after_rewriter : 0.000010s : 0.02% optimize.order_py_execute_after_rewriter : 0.000008s : 0.01% optimize.mutable_eliminate : 0.000873s : 1.28% optimize.opt_b.b_1 : 0.000218s : 0.32% optimize.opt_b.b_2 : 0.000013s : 0.02% optimize.opt_b.updatestate_depend_eliminate : 0.000013s : 0.02% optimize.opt_b.updatestate_assign_eliminate : 0.000005s : 0.01% optimize.opt_b.updatestate_loads_eliminate : 0.000005s : 0.01% optimize.opt_b.renormalize : 0.000001s : 0.00% optimize.opt_b.cse : 0.000075s : 0.11% optimize.optimize_parallel_all_gather_comm : 0.000026s : 0.04% optimize.overlap_param_gather : 0.000002s : 0.00% optimize.cconv : 0.000046s : 0.07% optimize.loop_unroll : 0.000616s : 0.90% optimize.opt_after_cconv.c_1 : 0.000052s : 0.08% optimize.opt_after_cconv.parameter_eliminate : 0.000006s : 0.01% optimize.opt_after_cconv.updatestate_depend_eliminate : 0.000009s : 0.01% optimize.opt_after_cconv.updatestate_assign_eliminate : 0.000004s : 0.01% optimize.opt_after_cconv.updatestate_loads_eliminate : 0.000005s : 0.01% optimize.opt_after_cconv.cse : 0.000059s : 0.09% optimize.opt_after_cconv.renormalize : 0.000001s : 0.00% optimize.remove_dup_value : 0.000082s : 0.12% optimize.tuple_transform.d_1 : 0.000093s : 0.14% optimize.tuple_transform.none_parameter_eliminate : 0.000003s : 0.00% optimize.tuple_transform.renormalize : 0.000000s : 0.00% optimize.tuple_transform.switch_simplify : 0.000012s : 0.02% optimize.partial_unused_args_eliminate : 0.000002s : 0.00% optimize.add_recomputation : 0.000080s : 0.12% optimize.cse_after_recomputation.cse : 0.000029s : 0.04% optimize.environ_conv : 0.000010s : 0.01% optimize.swap_dp_allreduce_reducescatter : 0.000010s : 0.02% optimize.bias_add_comm_swap : 0.000004s : 0.01% optimize.label_micro_interleaved_index : 0.000006s : 0.01% optimize.label_fine_grained_interleaved_index : 0.000003s : 0.00% optimize.merge_cast_opt : 0.000002s : 0.00% optimize.slice_recompute_activation : 0.000002s : 0.00% optimize.micro_interleaved_order_control : 0.000003s : 0.00% optimize.assign_add_opt : 0.000002s : 0.00% optimize.ForceFp32Comm : 0.000001s : 0.00% optimize.remove_cast_before_assign_add : 0.000001s : 0.00% optimize.full_micro_interleaved_order_control : 0.000002s : 0.00% optimize.reorder_send_recv_between_fp_bp : 0.000003s : 0.00% optimize.comm_op_add_attrs : 0.000001s : 0.00% optimize.add_comm_op_reuse_tag : 0.000001s : 0.00% optimize.interleave_split_concat_branches : 0.000002s : 0.00% optimize.interleave_parallel_branches : 0.000001s : 0.00% optimize.overlap_opt_shard_in_pipeline : 0.000002s : 0.00% optimize.overlap_opt_shard_grad_in_pipeline : 0.000002s : 0.00% optimize.control_data_broadcast_order : 0.000020s : 0.03% optimize.grouped_pairwise_exchange_alltoall : 0.000002s : 0.00% optimize.offloading_packed_experts : 0.000006s : 0.01% optimize.overlap_recompute_and_grad_model_parallel : 0.000007s : 0.01% optimize.overlap_grad_matmul_and_grad_allreduce : 0.000001s : 0.00% optimize.overlap_recompute_allgather_and_fa_grad : 0.000002s : 0.00% optimize.overlap_recompute_comm : 0.000002s : 0.00% optimize.overlap_grad_ring_attention : 0.000006s : 0.01% optimize.overlap_grad_flash_sp : 0.000027s : 0.04% optimize.begin_end_overlap_inline : 0.000001s : 0.00% optimize.split_matmul_comm_elemetwise : 0.000002s : 0.00% optimize.split_layernorm_comm : 0.000002s : 0.00% optimize.handle_group_info : 0.000001s : 0.00% optimize.symbol_engine_optimizer.build : 0.000099s : 0.14% optimize.symbol_engine_optimizer.elim_shapecalc : 0.000027s : 0.04% optimize.symbol_engine_optimizer.elim_not_effective : 0.000027s : 0.04% optimize.symbol_engine_optimizer.opt_reshape : 0.000017s : 0.02% optimize.symbol_engine_optimizer.fold_const_symbol : 0.000024s : 0.04% optimize.symbol_engine_optimizer.renormalize : 0.000000s : 0.00% detach_backward : 0.000003s : 0.00% pipeline_parallel_scheduler : 0.000002s : 0.00% auto_monad_reorder : 0.000035s : 0.05% get_jit_bprop_graph : 0.000002s : 0.00% rewriter_after_jit_bprop_graph : 0.000007s : 0.01% opt_after_jit_grad : 0.000846s : 1.24% validate : 0.000070s : 0.10% Time group info: ------[substitution.] 0.000186 33 2.17% : 0.000004s : 2: substitution.elim_not_effective 4.51% : 0.000008s : 2: substitution.fold_const_symbol 4.91% : 0.000009s : 8: substitution.graph_param_transform 75.86% : 0.000141s : 1: substitution.inline 2.55% : 0.000005s : 4: substitution.j_node_and_user_rematch 3.62% : 0.000007s : 4: substitution.remove_not_recompute_node 6.37% : 0.000012s : 12: substitution.replace_old_param ------[type_inference.] 0.007333 2 86.55% : 0.006347s : 1: type_inference.infer 13.45% : 0.000987s : 1: type_inference.specialize ------[replace.] 0.000027 1 100.00% : 0.000027s : 1: replace.inline ------[match.] 0.000139 1 100.00% : 0.000139s : 1: match.inline ------[predicate.] 0.000255 1879 0.94% : 0.000002s : 17: predicate.accumulaten_eliminater 1.51% : 0.000004s : 8: predicate.ad_related_special_op_eliminate 0.73% : 0.000002s : 16: predicate.addn_check_dump 0.74% : 0.000002s : 17: predicate.addn_zero_filter 0.68% : 0.000002s : 17: predicate.adjust_all_reduce_mul_add 2.75% : 0.000007s : 33: predicate.arithmetic_simplify 0.82% : 0.000002s : 17: predicate.cast_eliminate 0.84% : 0.000002s : 16: predicate.check_bprop_eliminate 0.74% : 0.000002s : 16: predicate.compare_switch_simplify 0.25% : 0.000001s : 8: predicate.const_output_eliminate 0.72% : 0.000002s : 16: predicate.depend_value_elim 0.72% : 0.000002s : 17: predicate.dict_get_item_const_eliminator 0.96% : 0.000002s : 17: predicate.dict_get_item_eliminator 0.91% : 0.000002s : 17: predicate.dict_set_item_eliminator 1.75% : 0.000004s : 16: predicate.dumpgradient_eliminate 0.35% : 0.000001s : 8: predicate.elim_not_effective 0.72% : 0.000002s : 8: predicate.elim_shapecalc_of_broadcastargs 1.23% : 0.000003s : 25: predicate.environ_add_const_eliminate 0.99% : 0.000003s : 25: predicate.environ_get_add_eliminate 0.95% : 0.000002s : 25: predicate.environ_get_depend_swap 1.79% : 0.000005s : 41: predicate.environ_get_eliminate 1.00% : 0.000003s : 25: predicate.environ_get_set_eliminate 0.73% : 0.000002s : 18: predicate.exchange_switch_depend_value 1.92% : 0.000005s : 18: predicate.float_depend_g_call 0.76% : 0.000002s : 16: predicate.float_environ_get_switch 1.09% : 0.000003s : 24: predicate.float_tuple_getitem_switch 0.26% : 0.000001s : 8: predicate.fold_const_symbol 0.95% : 0.000002s : 16: predicate.get_grad_eliminate 0.38% : 0.000001s : 8: predicate.graph_param_transform 0.75% : 0.000002s : 16: predicate.incorporate_call 0.60% : 0.000002s : 16: predicate.incorporate_call_switch 5.13% : 0.000013s : 83: predicate.inline 0.78% : 0.000002s : 16: predicate.inline_without_move 0.52% : 0.000001s : 16: predicate.j_node_and_user_rematch 1.28% : 0.000003s : 16: predicate.less_batch_normalization 1.67% : 0.000004s : 33: predicate.list_to_tuple_eliminator_ 2.06% : 0.000005s : 50: predicate.load_eliminater 1.17% : 0.000003s : 8: predicate.loop_unroll_after_grad 1.18% : 0.000003s : 25: predicate.loop_unroll_before_grad 1.76% : 0.000005s : 33: predicate.make_slice_get_slice_eliminator 0.78% : 0.000002s : 16: predicate.merge_addn 0.76% : 0.000002s : 16: predicate.micro_step_allgather_replace 0.73% : 0.000002s : 16: predicate.mini_step_allgather_replace 0.71% : 0.000002s : 17: predicate.minmaximum_grad 1.75% : 0.000004s : 8: predicate.mutable_eliminate 0.45% : 0.000001s : 8: predicate.opt_reshape 0.53% : 0.000001s : 8: predicate.parallel_virtual_node 1.56% : 0.000004s : 18: predicate.partial_defer_inline 1.08% : 0.000003s : 25: predicate.partial_eliminate 0.74% : 0.000002s : 17: predicate.print_const_string_wrapper 0.73% : 0.000002s : 16: predicate.reduce_all_const_elim 1.02% : 0.000003s : 17: predicate.reduce_eliminate 2.05% : 0.000005s : 50: predicate.redundant_stop_gradient_eliminater 0.66% : 0.000002s : 16: predicate.remove_not_recompute_node 1.71% : 0.000004s : 33: predicate.replace_applicator 0.77% : 0.000002s : 16: predicate.replace_old_param 0.36% : 0.000001s : 8: predicate.reset_defer_inline 0.78% : 0.000002s : 17: predicate.reshape_eliminate 0.74% : 0.000002s : 16: predicate.row_tensor_add_zeros_like 0.77% : 0.000002s : 8: predicate.row_tensor_eliminate 1.10% : 0.000003s : 16: predicate.same_eliminate 0.62% : 0.000002s : 16: predicate.set_cell_output_no_recompute 0.95% : 0.000002s : 16: predicate.shard_identity_eliminate 0.82% : 0.000002s : 16: predicate.special_op_eliminate 0.81% : 0.000002s : 16: predicate.specialize_transform 1.30% : 0.000003s : 16: predicate.split_environ_get_set_with_tuple_value 1.07% : 0.000003s : 16: predicate.stack_unstack_eliminate 0.44% : 0.000001s : 8: predicate.switch_call_monad_eliminater 0.77% : 0.000002s : 18: predicate.switch_defer_inline 1.46% : 0.000004s : 34: predicate.switch_layer_defer_inline 3.68% : 0.000009s : 67: predicate.switch_simplify 0.71% : 0.000002s : 17: predicate.tile_eliminate 0.72% : 0.000002s : 17: predicate.transpose_eliminate 1.56% : 0.000004s : 33: predicate.tuple_list_convert_item_index_to_positive 1.67% : 0.000004s : 33: predicate.tuple_list_get_item_const_eliminator 1.50% : 0.000004s : 33: predicate.tuple_list_get_item_depend_reorder 3.51% : 0.000009s : 49: predicate.tuple_list_get_item_eliminator 1.46% : 0.000004s : 33: predicate.tuple_list_get_set_item_eliminator 2.51% : 0.000006s : 49: predicate.tuple_list_set_item_eliminator 1.52% : 0.000004s : 33: predicate.tuple_to_list_eliminator_ 1.91% : 0.000005s : 50: predicate.updatestate_pure_node_eliminater 2.79% : 0.000007s : 66: predicate.updatestate_useless_node_eliminater 0.47% : 0.000001s : 8: predicate.value_based_eliminate 0.92% : 0.000002s : 16: predicate.virtual_dataset_eliminate 0.82% : 0.000002s : 16: predicate.virtual_output_eliminate 0.38% : 0.000001s : 8: predicate.virtual_view_grad_eliminate 0.78% : 0.000002s : 8: predicate.zero_like_fill_zero ------[func_graph_cloner_run.] 0.001052 6 48.45% : 0.000510s : 3: func_graph_cloner_run.FuncGraphClonerGraph 51.55% : 0.000542s : 3: func_graph_cloner_run.FuncGraphSpecializer ------[meta_graph.] 0.000000 0 ------[manager.] 0.000000 0 ------[pynative] 0.000000 0 ------[others.] 0.246832 192 0.00% : 0.000006s : 1: ForceFp32Comm 12.42% : 0.030665s : 1: add_attr 12.42% : 0.030645s : 1: add_attr_with_inline 0.00% : 0.000007s : 1: add_comm_op_reuse_tag 0.03% : 0.000085s : 1: add_recomputation 0.00% : 0.000005s : 1: assign_add_opt 0.03% : 0.000066s : 1: auto_monad 0.02% : 0.000039s : 1: auto_monad_reorder 0.00% : 0.000004s : 1: begin_end_overlap_inline 0.00% : 0.000011s : 1: bias_add_comm_swap 0.36% : 0.000884s : 1: bootstrap 0.02% : 0.000051s : 1: cconv 0.00% : 0.000004s : 1: comm_op_add_attrs 0.01% : 0.000024s : 1: control_data_broadcast_order 0.01% : 0.000015s : 1: convert_after_rewriter 0.02% : 0.000047s : 1: cse_after_recomputation 0.00% : 0.000005s : 1: dataset_repeat_opt 0.00% : 0.000006s : 1: detach_backward 0.01% : 0.000014s : 1: environ_conv 0.01% : 0.000024s : 1: event_method 0.00% : 0.000008s : 1: full_micro_interleaved_order_control 0.00% : 0.000005s : 1: get_jit_bprop_graph 0.00% : 0.000008s : 1: graph_reusing 0.00% : 0.000008s : 1: grouped_pairwise_exchange_alltoall 0.00% : 0.000006s : 1: handle_group_info 0.00% : 0.000006s : 1: inline 0.00% : 0.000006s : 1: insert-virtual-dataset 0.00% : 0.000006s : 1: interleave_parallel_branches 0.00% : 0.000006s : 1: interleave_split_concat_branches 0.00% : 0.000006s : 1: label_fine_grained_interleaved_index 0.00% : 0.000011s : 1: label_micro_interleaved_index 0.25% : 0.000626s : 1: loop_unroll 0.00% : 0.000008s : 1: merge_cast_opt 0.00% : 0.000009s : 1: micro_interleaved_order_control 0.36% : 0.000888s : 1: mutable_eliminate 0.00% : 0.000009s : 1: offloading_packed_experts 0.01% : 0.000023s : 1: opt.transform.loop_unroll_optimizer 0.01% : 0.000029s : 1: opt.transform.mutable_eliminate 21.82% : 0.053869s : 78: opt.transform.opt_a 0.02% : 0.000050s : 1: opt.transform.opt_after_cconv 0.02% : 0.000057s : 1: opt.transform.opt_after_jit_grad 0.08% : 0.000198s : 28: opt.transform.opt_b 0.04% : 0.000102s : 2: opt.transform.opt_trans_graph 0.04% : 0.000091s : 4: opt.transform.symbol_engine_opt 22.99% : 0.056736s : 1: opt_a 0.08% : 0.000198s : 1: opt_after_cconv 0.35% : 0.000860s : 1: opt_after_jit_grad 0.17% : 0.000408s : 1: opt_b 24.41% : 0.060245s : 1: optimize 0.01% : 0.000032s : 1: optimize_parallel_all_gather_comm 0.00% : 0.000011s : 1: order_py_execute_after_rewriter 0.01% : 0.000033s : 1: overlap_grad_flash_sp 0.00% : 0.000008s : 1: overlap_grad_matmul_and_grad_allreduce 0.01% : 0.000013s : 1: overlap_grad_ring_attention 0.00% : 0.000008s : 1: overlap_opt_shard_grad_in_pipeline 0.00% : 0.000005s : 1: overlap_opt_shard_in_pipeline 0.00% : 0.000005s : 1: overlap_param_gather 0.00% : 0.000008s : 1: overlap_recompute_allgather_and_fa_grad 0.00% : 0.000010s : 1: overlap_recompute_and_grad_model_parallel 0.00% : 0.000006s : 1: overlap_recompute_comm 0.00% : 0.000008s : 1: parallel-infer-symbol 0.00% : 0.000004s : 1: parallel-infer-symbol-second 0.00% : 0.000005s : 1: partial_unused_args_eliminate 0.00% : 0.000006s : 1: pipeline_parallel_scheduler 0.00% : 0.000005s : 1: pipeline_split 0.02% : 0.000050s : 1: pre_auto_parallel 0.03% : 0.000066s : 1: py_interpret_to_execute 0.02% : 0.000037s : 1: py_interpret_to_execute_after_opt_a 0.00% : 0.000004s : 1: remove_cast_before_assign_add 0.04% : 0.000088s : 1: remove_dup_value 0.21% : 0.000525s : 1: renormalize.infer 0.28% : 0.000701s : 1: renormalize.specialize 0.00% : 0.000006s : 1: reorder_send_recv_between_fp_bp 0.00% : 0.000011s : 1: rewriter_after_jit_bprop_graph 0.03% : 0.000077s : 1: rewriter_after_opt_a 0.04% : 0.000105s : 1: rewriter_before_opt_a 0.00% : 0.000005s : 1: slice_cell_reuse_recomputed_activation 0.00% : 0.000005s : 1: slice_recompute_activation 0.00% : 0.000008s : 1: split_layernorm_comm 0.00% : 0.000006s : 1: split_matmul_comm_elemetwise 0.01% : 0.000014s : 1: swap_dp_allreduce_reducescatter 0.11% : 0.000260s : 1: symbol_engine_optimizer 0.06% : 0.000150s : 1: tuple_transform 3.02% : 0.007456s : 1: type_inference . [hook] pytest_runtest_teardown:test_paged_attention_mla_combine_cache_norm[0-float32-16] tests/st/infer/ops/test_internal_ops/test_paged_attention.py::test_paged_attention_mla_combine_cache_norm[0-float32-16],max_mem:270.0M [WARNING] ME(168852:281473710628656,MainProcess):2026-01-29-17:38:13.160.784 [mindspore/context.py:1334] For 'context.set_context', the parameter 'device_target' will be deprecated and removed in a future version. Please use the api mindspore.set_device() instead. TotalTime = 0.181378, [21] [bootstrap]: 0.00642119 [type_inference]: 0.0467706 [event_method]: 1.442e-05 [auto_monad]: 5.542e-05 [graph_reusing]: 5.40001e-06 [inline]: 2.99999e-06 [add_attr]: 0.120263, [1] [add_attr_with_inline]: 0.120245, [1] [Cycle 1]: 7.6e-05, [2] [tag_attr]: 2.575e-05 [meta_addattr_fg_expand]: 4.43001e-06 [parallel-infer-symbol]: 3.74002e-06 [pre_auto_parallel]: 4.384e-05 [insert-virtual-dataset]: 2.67001e-06 [parallel-infer-symbol-second]: 7.09988e-07 [dataset_repeat_opt]: 2.44001e-06 [pipeline_split]: 1.66e-06 [optimize]: 0.00678812, [53] [py_interpret_to_execute]: 3.804e-05 [rewriter_before_opt_a]: 7.514e-05 [opt_a]: 0.00367029, [2] [Cycle 1]: 0.00251643, [45] [expand_dump_flag]: 2.39999e-06 [switch_simplify]: 3.063e-05 [loop_unroll]: 1.796e-05 [a_1]: 0.00049746 [with_stream_mark]: 2.741e-05 [recompute_prepare]: 1.363e-05 [updatestate_depend_eliminate]: 5.91e-06 [updatestate_assign_eliminate]: 5.00001e-06 [updatestate_loads_eliminate]: 4.70001e-06 [parameter_eliminate]: 1.73002e-06 [a_2]: 0.00014448 [accelerated_algorithm]: 1.17e-05 [shard]: 2.53e-06 [meta_shard_fg_expand]: 2.32999e-06 [shard_inline]: 1.007e-05 [merge_send_recv]: 9.41e-06 [auto_parallel]: 8.03001e-06 [parallel]: 3.488e-05 [flash_sp]: 1.063e-05 [merge_comm]: 5.09998e-06 [allreduce_fusion]: 4.80999e-06 [matmul_add_comm_reduction]: 1.104e-05 [allreduce_slice_to_reducescatter]: 6.39993e-07 [virtual_shard_identity]: 1.224e-05 [virtual_dataset]: 1.522e-05 [get_grad_eliminate_]: 1.136e-05 [virtual_output]: 1.277e-05 [merge_forward]: 5.66e-06 [cell_reuse_recompute_pass]: 1.40001e-06 [offload_activation]: 1.183e-05 [cell_reuse_handle_not_recompute_node_pass]: 1.905e-05 [merge_recompute_call_nodes]: 1.37999e-06 [before_grad]: 1.569e-05 [set_forward_comm_id_for_comm_node_pass]: 5.23002e-06 [meta_fg_expand]: 3.73999e-06 [flash_sp_send_recv_attached]: 2.42001e-06 [receive_attached]: 2.79999e-06 [after_resolve]: 2.873e-05 [a_after_grad]: 1.831e-05 [renormalize]: 0.00103501 [add_forward_monad_depend]: 7.24001e-06 [auto_monad_grad]: 2.76999e-06 [auto_monad_eliminator]: 1.956e-05 [cse]: 7.516e-05 [a_3]: 8.481e-05 [Cycle 2]: 0.00114374, [45] [expand_dump_flag]: 1.81003e-06 [switch_simplify]: 1.391e-05 [loop_unroll]: 1.225e-05 [a_1]: 0.00031419 [with_stream_mark]: 1.856e-05 [recompute_prepare]: 1.448e-05 [updatestate_depend_eliminate]: 5.36002e-06 [updatestate_assign_eliminate]: 4.17e-06 [updatestate_loads_eliminate]: 4.21001e-06 [parameter_eliminate]: 1.24e-06 [a_2]: 0.00014957 [accelerated_algorithm]: 1.196e-05 [shard]: 1.66e-06 [meta_shard_fg_expand]: 2.07999e-06 [shard_inline]: 1.066e-05 [merge_send_recv]: 8.30999e-06 [auto_parallel]: 8.45001e-06 [parallel]: 5.68002e-06 [flash_sp]: 3.29001e-06 [merge_comm]: 5.10999e-06 [allreduce_fusion]: 4.94998e-06 [matmul_add_comm_reduction]: 8.67e-06 [allreduce_slice_to_reducescatter]: 7.09988e-07 [virtual_shard_identity]: 1.252e-05 [virtual_dataset]: 1.2e-05 [get_grad_eliminate_]: 1.17e-05 [virtual_output]: 1.127e-05 [merge_forward]: 5.12e-06 [cell_reuse_recompute_pass]: 1.84998e-06 [offload_activation]: 1.032e-05 [cell_reuse_handle_not_recompute_node_pass]: 2.251e-05 [merge_recompute_call_nodes]: 8.30012e-07 [before_grad]: 1.494e-05 [set_forward_comm_id_for_comm_node_pass]: 5.05001e-06 [meta_fg_expand]: 2.99001e-06 [flash_sp_send_recv_attached]: 1.00001e-06 [receive_attached]: 1.78002e-06 [after_resolve]: 2.072e-05 [a_after_grad]: 1.957e-05 [renormalize]: 9.00181e-08 [add_forward_monad_depend]: 1.66002e-06 [auto_monad_grad]: 1.27e-06 [auto_monad_eliminator]: 1.033e-05 [cse]: 6.031e-05 [a_3]: 7.473e-05 [py_interpret_to_execute_after_opt_a]: 1.706e-05 [slice_cell_reuse_recomputed_activation]: 2.12999e-06 [rewriter_after_opt_a]: 6.034e-05 [convert_after_rewriter]: 1.042e-05 [order_py_execute_after_rewriter]: 7.54002e-06 [mutable_eliminate]: 0.00075457 [opt_b]: 0.00041259, [1] [Cycle 1]: 0.00040439, [7] [b_1]: 0.00024079 [b_2]: 1.28e-05 [updatestate_depend_eliminate]: 8.40001e-06 [updatestate_assign_eliminate]: 4.62998e-06 [updatestate_loads_eliminate]: 4.43999e-06 [renormalize]: 1.07998e-06 [cse]: 8.725e-05 [optimize_parallel_all_gather_comm]: 2.574e-05 [overlap_param_gather]: 2.14999e-06 [cconv]: 3.165e-05 [loop_unroll]: 0.00054021 [opt_after_cconv]: 0.00019358, [1] [Cycle 1]: 0.0001841, [7] [c_1]: 5.716e-05 [parameter_eliminate]: 3.14999e-06 [updatestate_depend_eliminate]: 7.21001e-06 [updatestate_assign_eliminate]: 4.48001e-06 [updatestate_loads_eliminate]: 4.4e-06 [cse]: 5.235e-05 [renormalize]: 6.69999e-07 [remove_dup_value]: 6.533e-05 [tuple_transform]: 0.00016004, [1] [Cycle 1]: 0.00015469, [4] [d_1]: 9.596e-05 [none_parameter_eliminate]: 1.90001e-06 [renormalize]: 1.8999e-07 [switch_simplify]: 1.386e-05 [partial_unused_args_eliminate]: 2.37999e-06 [add_recomputation]: 7.054e-05 [cse_after_recomputation]: 4.374e-05, [1] [Cycle 1]: 3.842e-05, [1] [cse]: 3.165e-05 [environ_conv]: 7.8e-06 [swap_dp_allreduce_reducescatter]: 6.67002e-06 [bias_add_comm_swap]: 3.4e-06 [label_micro_interleaved_index]: 5.05001e-06 [label_fine_grained_interleaved_index]: 3.14999e-06 [merge_cast_opt]: 1.60999e-06 [slice_recompute_activation]: 2.23002e-06 [micro_interleaved_order_control]: 2.51998e-06 [assign_add_opt]: 1.37999e-06 [ForceFp32Comm]: 9.00007e-07 [remove_cast_before_assign_add]: 1.08001e-06 [full_micro_interleaved_order_control]: 2.14e-06 [reorder_send_recv_between_fp_bp]: 3.09999e-06 [comm_op_add_attrs]: 1.19e-06 [add_comm_op_reuse_tag]: 9.90025e-07 [interleave_split_concat_branches]: 1.20001e-06 [interleave_parallel_branches]: 1.04e-06 [overlap_opt_shard_in_pipeline]: 1.34e-06 [overlap_opt_shard_grad_in_pipeline]: 1.81998e-06 [control_data_broadcast_order]: 1.836e-05 [grouped_pairwise_exchange_alltoall]: 1.60999e-06 [offloading_packed_experts]: 6.02999e-06 [overlap_recompute_and_grad_model_parallel]: 6.15002e-06 [overlap_grad_matmul_and_grad_allreduce]: 1.13001e-06 [overlap_recompute_allgather_and_fa_grad]: 1.51002e-06 [overlap_recompute_comm]: 2.63e-06 [overlap_grad_ring_attention]: 4.93001e-06 [overlap_grad_flash_sp]: 2.611e-05 [begin_end_overlap_inline]: 4.89992e-07 [split_matmul_comm_elemetwise]: 1.86e-06 [split_layernorm_comm]: 1.67999e-06 [handle_group_info]: 1.09e-06 [symbol_engine_optimizer]: 0.0002247, [1] [Cycle 1]: 0.00021857, [6] [build]: 9.031e-05 [elim_shapecalc]: 2.468e-05 [elim_not_effective]: 2.57e-05 [opt_reshape]: 1.47e-05 [fold_const_symbol]: 2.133e-05 [renormalize]: 1.90019e-07 [detach_backward]: 2.53998e-06 [pipeline_parallel_scheduler]: 1.69e-06 [auto_monad_reorder]: 2.63e-05 [get_jit_bprop_graph]: 2.14999e-06 [rewriter_after_jit_bprop_graph]: 8.68001e-06 [opt_after_jit_grad]: 0.00063809 [validate]: 5.925e-05 Sums bootstrap : 0.006421s : 10.72% type_inference : 0.046771s : 78.08% event_method : 0.000014s : 0.02% auto_monad : 0.000055s : 0.09% graph_reusing : 0.000005s : 0.01% inline : 0.000003s : 0.01% add_attr.add_attr_with_inline.tag_attr : 0.000026s : 0.04% add_attr.add_attr_with_inline.meta_addattr_fg_expand : 0.000004s : 0.01% parallel-infer-symbol : 0.000004s : 0.01% pre_auto_parallel : 0.000044s : 0.07% insert-virtual-dataset : 0.000003s : 0.00% parallel-infer-symbol-second : 0.000001s : 0.00% dataset_repeat_opt : 0.000002s : 0.00% pipeline_split : 0.000002s : 0.00% optimize.py_interpret_to_execute : 0.000038s : 0.06% optimize.rewriter_before_opt_a : 0.000075s : 0.13% optimize.opt_a.expand_dump_flag : 0.000004s : 0.01% optimize.opt_a.switch_simplify : 0.000045s : 0.07% optimize.opt_a.loop_unroll : 0.000030s : 0.05% optimize.opt_a.a_1 : 0.000812s : 1.36% optimize.opt_a.with_stream_mark : 0.000046s : 0.08% optimize.opt_a.recompute_prepare : 0.000028s : 0.05% optimize.opt_a.updatestate_depend_eliminate : 0.000011s : 0.02% optimize.opt_a.updatestate_assign_eliminate : 0.000009s : 0.02% optimize.opt_a.updatestate_loads_eliminate : 0.000009s : 0.01% optimize.opt_a.parameter_eliminate : 0.000003s : 0.00% optimize.opt_a.a_2 : 0.000294s : 0.49% optimize.opt_a.accelerated_algorithm : 0.000024s : 0.04% optimize.opt_a.shard : 0.000004s : 0.01% optimize.opt_a.meta_shard_fg_expand : 0.000004s : 0.01% optimize.opt_a.shard_inline : 0.000021s : 0.03% optimize.opt_a.merge_send_recv : 0.000018s : 0.03% optimize.opt_a.auto_parallel : 0.000016s : 0.03% optimize.opt_a.parallel : 0.000041s : 0.07% optimize.opt_a.flash_sp : 0.000014s : 0.02% optimize.opt_a.merge_comm : 0.000010s : 0.02% optimize.opt_a.allreduce_fusion : 0.000010s : 0.02% optimize.opt_a.matmul_add_comm_reduction : 0.000020s : 0.03% optimize.opt_a.allreduce_slice_to_reducescatter : 0.000001s : 0.00% optimize.opt_a.virtual_shard_identity : 0.000025s : 0.04% optimize.opt_a.virtual_dataset : 0.000027s : 0.05% optimize.opt_a.get_grad_eliminate_ : 0.000023s : 0.04% optimize.opt_a.virtual_output : 0.000024s : 0.04% optimize.opt_a.merge_forward : 0.000011s : 0.02% optimize.opt_a.cell_reuse_recompute_pass : 0.000003s : 0.01% optimize.opt_a.offload_activation : 0.000022s : 0.04% optimize.opt_a.cell_reuse_handle_not_recompute_node_pass : 0.000042s : 0.07% optimize.opt_a.merge_recompute_call_nodes : 0.000002s : 0.00% optimize.opt_a.before_grad : 0.000031s : 0.05% optimize.opt_a.set_forward_comm_id_for_comm_node_pass : 0.000010s : 0.02% optimize.opt_a.meta_fg_expand : 0.000007s : 0.01% optimize.opt_a.flash_sp_send_recv_attached : 0.000003s : 0.01% optimize.opt_a.receive_attached : 0.000005s : 0.01% optimize.opt_a.after_resolve : 0.000049s : 0.08% optimize.opt_a.a_after_grad : 0.000038s : 0.06% optimize.opt_a.renormalize : 0.001035s : 1.73% optimize.opt_a.add_forward_monad_depend : 0.000009s : 0.01% optimize.opt_a.auto_monad_grad : 0.000004s : 0.01% optimize.opt_a.auto_monad_eliminator : 0.000030s : 0.05% optimize.opt_a.cse : 0.000135s : 0.23% optimize.opt_a.a_3 : 0.000160s : 0.27% optimize.py_interpret_to_execute_after_opt_a : 0.000017s : 0.03% optimize.slice_cell_reuse_recomputed_activation : 0.000002s : 0.00% optimize.rewriter_after_opt_a : 0.000060s : 0.10% optimize.convert_after_rewriter : 0.000010s : 0.02% optimize.order_py_execute_after_rewriter : 0.000008s : 0.01% optimize.mutable_eliminate : 0.000755s : 1.26% optimize.opt_b.b_1 : 0.000241s : 0.40% optimize.opt_b.b_2 : 0.000013s : 0.02% optimize.opt_b.updatestate_depend_eliminate : 0.000008s : 0.01% optimize.opt_b.updatestate_assign_eliminate : 0.000005s : 0.01% optimize.opt_b.updatestate_loads_eliminate : 0.000004s : 0.01% optimize.opt_b.renormalize : 0.000001s : 0.00% optimize.opt_b.cse : 0.000087s : 0.15% optimize.optimize_parallel_all_gather_comm : 0.000026s : 0.04% optimize.overlap_param_gather : 0.000002s : 0.00% optimize.cconv : 0.000032s : 0.05% optimize.loop_unroll : 0.000540s : 0.90% optimize.opt_after_cconv.c_1 : 0.000057s : 0.10% optimize.opt_after_cconv.parameter_eliminate : 0.000003s : 0.01% optimize.opt_after_cconv.updatestate_depend_eliminate : 0.000007s : 0.01% optimize.opt_after_cconv.updatestate_assign_eliminate : 0.000004s : 0.01% optimize.opt_after_cconv.updatestate_loads_eliminate : 0.000004s : 0.01% optimize.opt_after_cconv.cse : 0.000052s : 0.09% optimize.opt_after_cconv.renormalize : 0.000001s : 0.00% optimize.remove_dup_value : 0.000065s : 0.11% optimize.tuple_transform.d_1 : 0.000096s : 0.16% optimize.tuple_transform.none_parameter_eliminate : 0.000002s : 0.00% optimize.tuple_transform.renormalize : 0.000000s : 0.00% optimize.tuple_transform.switch_simplify : 0.000014s : 0.02% optimize.partial_unused_args_eliminate : 0.000002s : 0.00% optimize.add_recomputation : 0.000071s : 0.12% optimize.cse_after_recomputation.cse : 0.000032s : 0.05% optimize.environ_conv : 0.000008s : 0.01% optimize.swap_dp_allreduce_reducescatter : 0.000007s : 0.01% optimize.bias_add_comm_swap : 0.000003s : 0.01% optimize.label_micro_interleaved_index : 0.000005s : 0.01% optimize.label_fine_grained_interleaved_index : 0.000003s : 0.01% optimize.merge_cast_opt : 0.000002s : 0.00% optimize.slice_recompute_activation : 0.000002s : 0.00% optimize.micro_interleaved_order_control : 0.000003s : 0.00% optimize.assign_add_opt : 0.000001s : 0.00% optimize.ForceFp32Comm : 0.000001s : 0.00% optimize.remove_cast_before_assign_add : 0.000001s : 0.00% optimize.full_micro_interleaved_order_control : 0.000002s : 0.00% optimize.reorder_send_recv_between_fp_bp : 0.000003s : 0.01% optimize.comm_op_add_attrs : 0.000001s : 0.00% optimize.add_comm_op_reuse_tag : 0.000001s : 0.00% optimize.interleave_split_concat_branches : 0.000001s : 0.00% optimize.interleave_parallel_branches : 0.000001s : 0.00% optimize.overlap_opt_shard_in_pipeline : 0.000001s : 0.00% optimize.overlap_opt_shard_grad_in_pipeline : 0.000002s : 0.00% optimize.control_data_broadcast_order : 0.000018s : 0.03% optimize.grouped_pairwise_exchange_alltoall : 0.000002s : 0.00% optimize.offloading_packed_experts : 0.000006s : 0.01% optimize.overlap_recompute_and_grad_model_parallel : 0.000006s : 0.01% optimize.overlap_grad_matmul_and_grad_allreduce : 0.000001s : 0.00% optimize.overlap_recompute_allgather_and_fa_grad : 0.000002s : 0.00% optimize.overlap_recompute_comm : 0.000003s : 0.00% optimize.overlap_grad_ring_attention : 0.000005s : 0.01% optimize.overlap_grad_flash_sp : 0.000026s : 0.04% optimize.begin_end_overlap_inline : 0.000000s : 0.00% optimize.split_matmul_comm_elemetwise : 0.000002s : 0.00% optimize.split_layernorm_comm : 0.000002s : 0.00% optimize.handle_group_info : 0.000001s : 0.00% optimize.symbol_engine_optimizer.build : 0.000090s : 0.15% optimize.symbol_engine_optimizer.elim_shapecalc : 0.000025s : 0.04% optimize.symbol_engine_optimizer.elim_not_effective : 0.000026s : 0.04% optimize.symbol_engine_optimizer.opt_reshape : 0.000015s : 0.02% optimize.symbol_engine_optimizer.fold_const_symbol : 0.000021s : 0.04% optimize.symbol_engine_optimizer.renormalize : 0.000000s : 0.00% detach_backward : 0.000003s : 0.00% pipeline_parallel_scheduler : 0.000002s : 0.00% auto_monad_reorder : 0.000026s : 0.04% get_jit_bprop_graph : 0.000002s : 0.00% rewriter_after_jit_bprop_graph : 0.000009s : 0.01% opt_after_jit_grad : 0.000638s : 1.07% validate : 0.000059s : 0.10% Time group info: ------[substitution.] 0.000150 33 2.65% : 0.000004s : 2: substitution.elim_not_effective 5.04% : 0.000008s : 2: substitution.fold_const_symbol 4.90% : 0.000007s : 8: substitution.graph_param_transform 74.96% : 0.000112s : 1: substitution.inline 2.59% : 0.000004s : 4: substitution.j_node_and_user_rematch 3.89% : 0.000006s : 4: substitution.remove_not_recompute_node 5.97% : 0.000009s : 12: substitution.replace_old_param ------[type_inference.] 0.046669 2 98.24% : 0.045847s : 1: type_inference.infer 1.76% : 0.000822s : 1: type_inference.specialize ------[replace.] 0.000022 1 100.00% : 0.000022s : 1: replace.inline ------[match.] 0.000112 1 100.00% : 0.000112s : 1: match.inline ------[predicate.] 0.000231 1879 0.78% : 0.000002s : 17: predicate.accumulaten_eliminater 0.92% : 0.000002s : 8: predicate.ad_related_special_op_eliminate 0.69% : 0.000002s : 16: predicate.addn_check_dump 0.80% : 0.000002s : 17: predicate.addn_zero_filter 0.74% : 0.000002s : 17: predicate.adjust_all_reduce_mul_add 2.33% : 0.000005s : 33: predicate.arithmetic_simplify 1.53% : 0.000004s : 17: predicate.cast_eliminate 0.78% : 0.000002s : 16: predicate.check_bprop_eliminate 0.66% : 0.000002s : 16: predicate.compare_switch_simplify 0.27% : 0.000001s : 8: predicate.const_output_eliminate 0.74% : 0.000002s : 16: predicate.depend_value_elim 0.90% : 0.000002s : 17: predicate.dict_get_item_const_eliminator 0.94% : 0.000002s : 17: predicate.dict_get_item_eliminator 0.83% : 0.000002s : 17: predicate.dict_set_item_eliminator 1.53% : 0.000004s : 16: predicate.dumpgradient_eliminate 0.39% : 0.000001s : 8: predicate.elim_not_effective 0.48% : 0.000001s : 8: predicate.elim_shapecalc_of_broadcastargs 1.14% : 0.000003s : 25: predicate.environ_add_const_eliminate 1.15% : 0.000003s : 25: predicate.environ_get_add_eliminate 1.04% : 0.000002s : 25: predicate.environ_get_depend_swap 2.02% : 0.000005s : 41: predicate.environ_get_eliminate 1.15% : 0.000003s : 25: predicate.environ_get_set_eliminate 0.76% : 0.000002s : 18: predicate.exchange_switch_depend_value 1.56% : 0.000004s : 18: predicate.float_depend_g_call 0.90% : 0.000002s : 16: predicate.float_environ_get_switch 1.08% : 0.000002s : 24: predicate.float_tuple_getitem_switch 0.29% : 0.000001s : 8: predicate.fold_const_symbol 1.00% : 0.000002s : 16: predicate.get_grad_eliminate 0.33% : 0.000001s : 8: predicate.graph_param_transform 0.75% : 0.000002s : 16: predicate.incorporate_call 0.62% : 0.000001s : 16: predicate.incorporate_call_switch 5.13% : 0.000012s : 83: predicate.inline 0.92% : 0.000002s : 16: predicate.inline_without_move 0.51% : 0.000001s : 16: predicate.j_node_and_user_rematch 0.99% : 0.000002s : 16: predicate.less_batch_normalization 1.60% : 0.000004s : 33: predicate.list_to_tuple_eliminator_ 2.16% : 0.000005s : 50: predicate.load_eliminater 1.04% : 0.000002s : 8: predicate.loop_unroll_after_grad 1.42% : 0.000003s : 25: predicate.loop_unroll_before_grad 1.85% : 0.000004s : 33: predicate.make_slice_get_slice_eliminator 0.85% : 0.000002s : 16: predicate.merge_addn 0.80% : 0.000002s : 16: predicate.micro_step_allgather_replace 0.75% : 0.000002s : 16: predicate.mini_step_allgather_replace 0.74% : 0.000002s : 17: predicate.minmaximum_grad 1.43% : 0.000003s : 8: predicate.mutable_eliminate 0.43% : 0.000001s : 8: predicate.opt_reshape 0.46% : 0.000001s : 8: predicate.parallel_virtual_node 1.02% : 0.000002s : 18: predicate.partial_defer_inline 1.29% : 0.000003s : 25: predicate.partial_eliminate 0.73% : 0.000002s : 17: predicate.print_const_string_wrapper 0.81% : 0.000002s : 16: predicate.reduce_all_const_elim 1.08% : 0.000003s : 17: predicate.reduce_eliminate 2.19% : 0.000005s : 50: predicate.redundant_stop_gradient_eliminater 0.83% : 0.000002s : 16: predicate.remove_not_recompute_node 1.41% : 0.000003s : 33: predicate.replace_applicator 0.69% : 0.000002s : 16: predicate.replace_old_param 0.33% : 0.000001s : 8: predicate.reset_defer_inline 0.93% : 0.000002s : 17: predicate.reshape_eliminate 1.08% : 0.000003s : 16: predicate.row_tensor_add_zeros_like 0.46% : 0.000001s : 8: predicate.row_tensor_eliminate 1.00% : 0.000002s : 16: predicate.same_eliminate 0.65% : 0.000002s : 16: predicate.set_cell_output_no_recompute 0.97% : 0.000002s : 16: predicate.shard_identity_eliminate 0.82% : 0.000002s : 16: predicate.special_op_eliminate 0.80% : 0.000002s : 16: predicate.specialize_transform 1.05% : 0.000002s : 16: predicate.split_environ_get_set_with_tuple_value 0.99% : 0.000002s : 16: predicate.stack_unstack_eliminate 0.48% : 0.000001s : 8: predicate.switch_call_monad_eliminater 0.92% : 0.000002s : 18: predicate.switch_defer_inline 1.70% : 0.000004s : 34: predicate.switch_layer_defer_inline 3.86% : 0.000009s : 67: predicate.switch_simplify 0.93% : 0.000002s : 17: predicate.tile_eliminate 0.76% : 0.000002s : 17: predicate.transpose_eliminate 1.46% : 0.000003s : 33: predicate.tuple_list_convert_item_index_to_positive 1.72% : 0.000004s : 33: predicate.tuple_list_get_item_const_eliminator 1.97% : 0.000005s : 33: predicate.tuple_list_get_item_depend_reorder 2.85% : 0.000007s : 49: predicate.tuple_list_get_item_eliminator 1.61% : 0.000004s : 33: predicate.tuple_list_get_set_item_eliminator 2.46% : 0.000006s : 49: predicate.tuple_list_set_item_eliminator 1.68% : 0.000004s : 33: predicate.tuple_to_list_eliminator_ 2.10% : 0.000005s : 50: predicate.updatestate_pure_node_eliminater 3.10% : 0.000007s : 66: predicate.updatestate_useless_node_eliminater 0.42% : 0.000001s : 8: predicate.value_based_eliminate 0.84% : 0.000002s : 16: predicate.virtual_dataset_eliminate 0.88% : 0.000002s : 16: predicate.virtual_output_eliminate 0.42% : 0.000001s : 8: predicate.virtual_view_grad_eliminate 0.55% : 0.000001s : 8: predicate.zero_like_fill_zero ------[func_graph_cloner_run.] 0.002619 6 82.10% : 0.002150s : 3: func_graph_cloner_run.FuncGraphClonerGraph 17.90% : 0.000469s : 3: func_graph_cloner_run.FuncGraphSpecializer ------[meta_graph.] 0.000000 0 ------[manager.] 0.000000 0 ------[pynative] 0.000000 0 ------[others.] 0.311421 192 0.00% : 0.000003s : 1: ForceFp32Comm 38.62% : 0.120271s : 1: add_attr 38.61% : 0.120250s : 1: add_attr_with_inline 0.00% : 0.000004s : 1: add_comm_op_reuse_tag 0.02% : 0.000075s : 1: add_recomputation 0.00% : 0.000004s : 1: assign_add_opt 0.02% : 0.000061s : 1: auto_monad 0.01% : 0.000031s : 1: auto_monad_reorder 0.00% : 0.000003s : 1: begin_end_overlap_inline 0.00% : 0.000006s : 1: bias_add_comm_swap 2.10% : 0.006534s : 1: bootstrap 0.01% : 0.000036s : 1: cconv 0.00% : 0.000004s : 1: comm_op_add_attrs 0.01% : 0.000022s : 1: control_data_broadcast_order 0.00% : 0.000014s : 1: convert_after_rewriter 0.02% : 0.000047s : 1: cse_after_recomputation 0.00% : 0.000005s : 1: dataset_repeat_opt 0.00% : 0.000006s : 1: detach_backward 0.00% : 0.000011s : 1: environ_conv 0.01% : 0.000021s : 1: event_method 0.00% : 0.000005s : 1: full_micro_interleaved_order_control 0.00% : 0.000005s : 1: get_jit_bprop_graph 0.00% : 0.000009s : 1: graph_reusing 0.00% : 0.000004s : 1: grouped_pairwise_exchange_alltoall 0.00% : 0.000004s : 1: handle_group_info 0.00% : 0.000006s : 1: inline 0.00% : 0.000006s : 1: insert-virtual-dataset 0.00% : 0.000004s : 1: interleave_parallel_branches 0.00% : 0.000004s : 1: interleave_split_concat_branches 0.00% : 0.000007s : 1: label_fine_grained_interleaved_index 0.00% : 0.000008s : 1: label_micro_interleaved_index 0.18% : 0.000549s : 1: loop_unroll 0.00% : 0.000004s : 1: merge_cast_opt 0.00% : 0.000006s : 1: micro_interleaved_order_control 0.25% : 0.000764s : 1: mutable_eliminate 0.00% : 0.000009s : 1: offloading_packed_experts 0.01% : 0.000025s : 1: opt.transform.loop_unroll_optimizer 0.01% : 0.000024s : 1: opt.transform.mutable_eliminate 0.51% : 0.001578s : 78: opt.transform.opt_a 0.02% : 0.000056s : 1: opt.transform.opt_after_cconv 0.02% : 0.000049s : 1: opt.transform.opt_after_jit_grad 0.07% : 0.000220s : 28: opt.transform.opt_b 0.03% : 0.000107s : 2: opt.transform.opt_trans_graph 0.03% : 0.000082s : 4: opt.transform.symbol_engine_opt 1.18% : 0.003674s : 1: opt_a 0.06% : 0.000200s : 1: opt_after_cconv 0.21% : 0.000648s : 1: opt_after_jit_grad 0.13% : 0.000417s : 1: opt_b 2.18% : 0.006794s : 1: optimize 0.01% : 0.000031s : 1: optimize_parallel_all_gather_comm 0.00% : 0.000011s : 1: order_py_execute_after_rewriter 0.01% : 0.000030s : 1: overlap_grad_flash_sp 0.00% : 0.000004s : 1: overlap_grad_matmul_and_grad_allreduce 0.00% : 0.000009s : 1: overlap_grad_ring_attention 0.00% : 0.000006s : 1: overlap_opt_shard_grad_in_pipeline 0.00% : 0.000011s : 1: overlap_opt_shard_in_pipeline 0.00% : 0.000006s : 1: overlap_param_gather 0.00% : 0.000004s : 1: overlap_recompute_allgather_and_fa_grad 0.00% : 0.000010s : 1: overlap_recompute_and_grad_model_parallel 0.00% : 0.000006s : 1: overlap_recompute_comm 0.00% : 0.000008s : 1: parallel-infer-symbol 0.00% : 0.000004s : 1: parallel-infer-symbol-second 0.00% : 0.000006s : 1: partial_unused_args_eliminate 0.00% : 0.000012s : 1: pipeline_parallel_scheduler 0.00% : 0.000005s : 1: pipeline_split 0.02% : 0.000048s : 1: pre_auto_parallel 0.01% : 0.000042s : 1: py_interpret_to_execute 0.01% : 0.000021s : 1: py_interpret_to_execute_after_opt_a 0.00% : 0.000004s : 1: remove_cast_before_assign_add 0.02% : 0.000070s : 1: remove_dup_value 0.16% : 0.000507s : 1: renormalize.infer 0.17% : 0.000519s : 1: renormalize.specialize 0.00% : 0.000006s : 1: reorder_send_recv_between_fp_bp 0.00% : 0.000012s : 1: rewriter_after_jit_bprop_graph 0.02% : 0.000065s : 1: rewriter_after_opt_a 0.03% : 0.000079s : 1: rewriter_before_opt_a 0.00% : 0.000005s : 1: slice_cell_reuse_recomputed_activation 0.00% : 0.000005s : 1: slice_recompute_activation 0.00% : 0.000008s : 1: split_layernorm_comm 0.00% : 0.000005s : 1: split_matmul_comm_elemetwise 0.00% : 0.000010s : 1: swap_dp_allreduce_reducescatter 0.07% : 0.000227s : 1: symbol_engine_optimizer 0.05% : 0.000163s : 1: tuple_transform 15.03% : 0.046793s : 1: type_inference . [hook] pytest_runtest_teardown:test_paged_attention_mla_combine_cache_norm[0-float32-128] tests/st/infer/ops/test_internal_ops/test_paged_attention.py::test_paged_attention_mla_combine_cache_norm[0-float32-128],max_mem:270.0M [WARNING] ME(168852:281473710628656,MainProcess):2026-01-29-17:38:19.651.920 [mindspore/context.py:1334] For 'context.set_context', the parameter 'device_target' will be deprecated and removed in a future version. Please use the api mindspore.set_device() instead. TotalTime = 0.23123, [21] [bootstrap]: 0.00091733 [type_inference]: 0.00806458 [event_method]: 1.669e-05 [auto_monad]: 5.692e-05 [graph_reusing]: 5.52001e-06 [inline]: 2.37001e-06 [add_attr]: 0.00607393, [1] [add_attr_with_inline]: 0.00605546, [1] [Cycle 1]: 7.713e-05, [2] [tag_attr]: 2.369e-05 [meta_addattr_fg_expand]: 4.17e-06 [parallel-infer-symbol]: 3.88001e-06 [pre_auto_parallel]: 4.655e-05 [insert-virtual-dataset]: 2.61999e-06 [parallel-infer-symbol-second]: 7.39994e-07 [dataset_repeat_opt]: 2.11e-06 [pipeline_split]: 1.78002e-06 [optimize]: 0.214837, [53] [py_interpret_to_execute]: 3.621e-05 [rewriter_before_opt_a]: 7.275e-05 [opt_a]: 0.00470902, [2] [Cycle 1]: 0.00348816, [45] [expand_dump_flag]: 3.25e-06 [switch_simplify]: 4.073e-05 [loop_unroll]: 2.851e-05 [a_1]: 0.0008328 [with_stream_mark]: 2.803e-05 [recompute_prepare]: 3.026e-05 [updatestate_depend_eliminate]: 6.59999e-06 [updatestate_assign_eliminate]: 5.66e-06 [updatestate_loads_eliminate]: 5.29e-06 [parameter_eliminate]: 1.81998e-06 [a_2]: 0.00023511 [accelerated_algorithm]: 1.855e-05 [shard]: 3.23e-06 [meta_shard_fg_expand]: 3.84002e-06 [shard_inline]: 1.405e-05 [merge_send_recv]: 1.186e-05 [auto_parallel]: 1.099e-05 [parallel]: 3.278e-05 [flash_sp]: 1.274e-05 [merge_comm]: 5.39e-06 [allreduce_fusion]: 6.08002e-06 [matmul_add_comm_reduction]: 2.253e-05 [allreduce_slice_to_reducescatter]: 1.07e-06 [virtual_shard_identity]: 1.863e-05 [virtual_dataset]: 1.553e-05 [get_grad_eliminate_]: 1.557e-05 [virtual_output]: 1.26e-05 [merge_forward]: 5.96e-06 [cell_reuse_recompute_pass]: 1.70001e-06 [offload_activation]: 1.579e-05 [cell_reuse_handle_not_recompute_node_pass]: 2.592e-05 [merge_recompute_call_nodes]: 1.41002e-06 [before_grad]: 2.721e-05 [set_forward_comm_id_for_comm_node_pass]: 5.81e-06 [meta_fg_expand]: 4.53001e-06 [flash_sp_send_recv_attached]: 3.56001e-06 [receive_attached]: 2.71999e-06 [after_resolve]: 2.329e-05 [a_after_grad]: 2.652e-05 [renormalize]: 0.0012831 [add_forward_monad_depend]: 8.16002e-06 [auto_monad_grad]: 2.53e-06 [auto_monad_eliminator]: 2.065e-05 [cse]: 8.605e-05 [a_3]: 8.499e-05 [Cycle 2]: 0.00120371, [45] [expand_dump_flag]: 2.61999e-06 [switch_simplify]: 1.388e-05 [loop_unroll]: 1.183e-05 [a_1]: 0.0002446 [with_stream_mark]: 2.516e-05 [recompute_prepare]: 1.376e-05 [updatestate_depend_eliminate]: 8.32e-06 [updatestate_assign_eliminate]: 5.27001e-06 [updatestate_loads_eliminate]: 4.75999e-06 [parameter_eliminate]: 1.96e-06 [a_2]: 0.0001293 [accelerated_algorithm]: 1.403e-05 [shard]: 1.89e-06 [meta_shard_fg_expand]: 2.37999e-06 [shard_inline]: 9.91e-06 [merge_send_recv]: 8.03999e-06 [auto_parallel]: 1.203e-05 [parallel]: 1.414e-05 [flash_sp]: 4.4e-06 [merge_comm]: 5.29998e-06 [allreduce_fusion]: 4.65999e-06 [matmul_add_comm_reduction]: 1.534e-05 [allreduce_slice_to_reducescatter]: 6.59988e-07 [virtual_shard_identity]: 1.154e-05 [virtual_dataset]: 1.046e-05 [get_grad_eliminate_]: 1.035e-05 [virtual_output]: 1.075e-05 [merge_forward]: 5.25999e-06 [cell_reuse_recompute_pass]: 2.49999e-06 [offload_activation]: 1.069e-05 [cell_reuse_handle_not_recompute_node_pass]: 1.891e-05 [merge_recompute_call_nodes]: 1.37e-06 [before_grad]: 1.478e-05 [set_forward_comm_id_for_comm_node_pass]: 5.87999e-06 [meta_fg_expand]: 3.45e-06 [flash_sp_send_recv_attached]: 1.22999e-06 [receive_attached]: 2.06e-06 [after_resolve]: 2.1e-05 [a_after_grad]: 1.547e-05 [renormalize]: 8.00064e-08 [add_forward_monad_depend]: 1.54e-06 [auto_monad_grad]: 1.86e-06 [auto_monad_eliminator]: 1.064e-05 [cse]: 5.428e-05 [a_3]: 6.318e-05 [py_interpret_to_execute_after_opt_a]: 4.018e-05 [slice_cell_reuse_recomputed_activation]: 2.37999e-06 [rewriter_after_opt_a]: 5.776e-05 [convert_after_rewriter]: 9.08002e-06 [order_py_execute_after_rewriter]: 7.01001e-06 [mutable_eliminate]: 0.00094798 [opt_b]: 0.00058058, [1] [Cycle 1]: 0.00056971, [7] [b_1]: 0.00032001 [b_2]: 2.041e-05 [updatestate_depend_eliminate]: 1.595e-05 [updatestate_assign_eliminate]: 4.96002e-06 [updatestate_loads_eliminate]: 8.91997e-06 [renormalize]: 1.02998e-06 [cse]: 9.15e-05 [optimize_parallel_all_gather_comm]: 3.481e-05 [overlap_param_gather]: 9.49e-06 [cconv]: 3.764e-05 [loop_unroll]: 0.206892 [opt_after_cconv]: 0.00028356, [1] [Cycle 1]: 0.00026882, [7] [c_1]: 7.389e-05 [parameter_eliminate]: 7.46001e-06 [updatestate_depend_eliminate]: 1.701e-05 [updatestate_assign_eliminate]: 5.91998e-06 [updatestate_loads_eliminate]: 5.40001e-06 [cse]: 0.00010531 [renormalize]: 1.62001e-06 [remove_dup_value]: 8.408e-05 [tuple_transform]: 0.00015241, [1] [Cycle 1]: 0.00014628, [4] [d_1]: 0.00010843 [none_parameter_eliminate]: 2.04999e-06 [renormalize]: 2.80008e-07 [switch_simplify]: 1.247e-05 [partial_unused_args_eliminate]: 2.47001e-06 [add_recomputation]: 8.022e-05 [cse_after_recomputation]: 5.148e-05, [1] [Cycle 1]: 4.161e-05, [1] [cse]: 3.42e-05 [environ_conv]: 9.36e-06 [swap_dp_allreduce_reducescatter]: 1.6e-05 [bias_add_comm_swap]: 4.12e-06 [label_micro_interleaved_index]: 9.10001e-06 [label_fine_grained_interleaved_index]: 3.16999e-06 [merge_cast_opt]: 1.44998e-06 [slice_recompute_activation]: 2.19999e-06 [micro_interleaved_order_control]: 2.76999e-06 [assign_add_opt]: 1.99e-06 [ForceFp32Comm]: 9.09989e-07 [remove_cast_before_assign_add]: 1.26002e-06 [full_micro_interleaved_order_control]: 2.27001e-06 [reorder_send_recv_between_fp_bp]: 2.76e-06 [comm_op_add_attrs]: 1.12999e-06 [add_comm_op_reuse_tag]: 1.34e-06 [interleave_split_concat_branches]: 1.44e-06 [interleave_parallel_branches]: 1.05001e-06 [overlap_opt_shard_in_pipeline]: 1.32e-06 [overlap_opt_shard_grad_in_pipeline]: 2.07001e-06 [control_data_broadcast_order]: 2.121e-05 [grouped_pairwise_exchange_alltoall]: 1.71002e-06 [offloading_packed_experts]: 5.62999e-06 [overlap_recompute_and_grad_model_parallel]: 6.53e-06 [overlap_grad_matmul_and_grad_allreduce]: 1.25999e-06 [overlap_recompute_allgather_and_fa_grad]: 1.27999e-06 [overlap_recompute_comm]: 3.15998e-06 [overlap_grad_ring_attention]: 5.60001e-06 [overlap_grad_flash_sp]: 2.902e-05 [begin_end_overlap_inline]: 5.40022e-07 [split_matmul_comm_elemetwise]: 2.64999e-06 [split_layernorm_comm]: 1.72999e-06 [handle_group_info]: 1.04998e-06 [symbol_engine_optimizer]: 0.00027022, [1] [Cycle 1]: 0.00026398, [6] [build]: 0.00011451 [elim_shapecalc]: 3.057e-05 [elim_not_effective]: 2.386e-05 [opt_reshape]: 2.005e-05 [fold_const_symbol]: 2.537e-05 [renormalize]: 2.50002e-07 [detach_backward]: 2.23002e-06 [pipeline_parallel_scheduler]: 1.57001e-06 [auto_monad_reorder]: 2.983e-05 [get_jit_bprop_graph]: 2.76999e-06 [rewriter_after_jit_bprop_graph]: 6.53998e-06 [opt_after_jit_grad]: 0.00088584 [validate]: 6.942e-05 Sums bootstrap : 0.000917s : 0.41% type_inference : 0.008065s : 3.61% event_method : 0.000017s : 0.01% auto_monad : 0.000057s : 0.03% graph_reusing : 0.000006s : 0.00% inline : 0.000002s : 0.00% add_attr.add_attr_with_inline.tag_attr : 0.000024s : 0.01% add_attr.add_attr_with_inline.meta_addattr_fg_expand : 0.000004s : 0.00% parallel-infer-symbol : 0.000004s : 0.00% pre_auto_parallel : 0.000047s : 0.02% insert-virtual-dataset : 0.000003s : 0.00% parallel-infer-symbol-second : 0.000001s : 0.00% dataset_repeat_opt : 0.000002s : 0.00% pipeline_split : 0.000002s : 0.00% optimize.py_interpret_to_execute : 0.000036s : 0.02% optimize.rewriter_before_opt_a : 0.000073s : 0.03% optimize.opt_a.expand_dump_flag : 0.000006s : 0.00% optimize.opt_a.switch_simplify : 0.000055s : 0.02% optimize.opt_a.loop_unroll : 0.000040s : 0.02% optimize.opt_a.a_1 : 0.001077s : 0.48% optimize.opt_a.with_stream_mark : 0.000053s : 0.02% optimize.opt_a.recompute_prepare : 0.000044s : 0.02% optimize.opt_a.updatestate_depend_eliminate : 0.000015s : 0.01% optimize.opt_a.updatestate_assign_eliminate : 0.000011s : 0.00% optimize.opt_a.updatestate_loads_eliminate : 0.000010s : 0.00% optimize.opt_a.parameter_eliminate : 0.000004s : 0.00% optimize.opt_a.a_2 : 0.000364s : 0.16% optimize.opt_a.accelerated_algorithm : 0.000033s : 0.01% optimize.opt_a.shard : 0.000005s : 0.00% optimize.opt_a.meta_shard_fg_expand : 0.000006s : 0.00% optimize.opt_a.shard_inline : 0.000024s : 0.01% optimize.opt_a.merge_send_recv : 0.000020s : 0.01% optimize.opt_a.auto_parallel : 0.000023s : 0.01% optimize.opt_a.parallel : 0.000047s : 0.02% optimize.opt_a.flash_sp : 0.000017s : 0.01% optimize.opt_a.merge_comm : 0.000011s : 0.00% optimize.opt_a.allreduce_fusion : 0.000011s : 0.00% optimize.opt_a.matmul_add_comm_reduction : 0.000038s : 0.02% optimize.opt_a.allreduce_slice_to_reducescatter : 0.000002s : 0.00% optimize.opt_a.virtual_shard_identity : 0.000030s : 0.01% optimize.opt_a.virtual_dataset : 0.000026s : 0.01% optimize.opt_a.get_grad_eliminate_ : 0.000026s : 0.01% optimize.opt_a.virtual_output : 0.000023s : 0.01% optimize.opt_a.merge_forward : 0.000011s : 0.01% optimize.opt_a.cell_reuse_recompute_pass : 0.000004s : 0.00% optimize.opt_a.offload_activation : 0.000026s : 0.01% optimize.opt_a.cell_reuse_handle_not_recompute_node_pass : 0.000045s : 0.02% optimize.opt_a.merge_recompute_call_nodes : 0.000003s : 0.00% optimize.opt_a.before_grad : 0.000042s : 0.02% optimize.opt_a.set_forward_comm_id_for_comm_node_pass : 0.000012s : 0.01% optimize.opt_a.meta_fg_expand : 0.000008s : 0.00% optimize.opt_a.flash_sp_send_recv_attached : 0.000005s : 0.00% optimize.opt_a.receive_attached : 0.000005s : 0.00% optimize.opt_a.after_resolve : 0.000044s : 0.02% optimize.opt_a.a_after_grad : 0.000042s : 0.02% optimize.opt_a.renormalize : 0.001283s : 0.57% optimize.opt_a.add_forward_monad_depend : 0.000010s : 0.00% optimize.opt_a.auto_monad_grad : 0.000004s : 0.00% optimize.opt_a.auto_monad_eliminator : 0.000031s : 0.01% optimize.opt_a.cse : 0.000140s : 0.06% optimize.opt_a.a_3 : 0.000148s : 0.07% optimize.py_interpret_to_execute_after_opt_a : 0.000040s : 0.02% optimize.slice_cell_reuse_recomputed_activation : 0.000002s : 0.00% optimize.rewriter_after_opt_a : 0.000058s : 0.03% optimize.convert_after_rewriter : 0.000009s : 0.00% optimize.order_py_execute_after_rewriter : 0.000007s : 0.00% optimize.mutable_eliminate : 0.000948s : 0.42% optimize.opt_b.b_1 : 0.000320s : 0.14% optimize.opt_b.b_2 : 0.000020s : 0.01% optimize.opt_b.updatestate_depend_eliminate : 0.000016s : 0.01% optimize.opt_b.updatestate_assign_eliminate : 0.000005s : 0.00% optimize.opt_b.updatestate_loads_eliminate : 0.000009s : 0.00% optimize.opt_b.renormalize : 0.000001s : 0.00% optimize.opt_b.cse : 0.000092s : 0.04% optimize.optimize_parallel_all_gather_comm : 0.000035s : 0.02% optimize.overlap_param_gather : 0.000009s : 0.00% optimize.cconv : 0.000038s : 0.02% optimize.loop_unroll : 0.206892s : 92.55% optimize.opt_after_cconv.c_1 : 0.000074s : 0.03% optimize.opt_after_cconv.parameter_eliminate : 0.000007s : 0.00% optimize.opt_after_cconv.updatestate_depend_eliminate : 0.000017s : 0.01% optimize.opt_after_cconv.updatestate_assign_eliminate : 0.000006s : 0.00% optimize.opt_after_cconv.updatestate_loads_eliminate : 0.000005s : 0.00% optimize.opt_after_cconv.cse : 0.000105s : 0.05% optimize.opt_after_cconv.renormalize : 0.000002s : 0.00% optimize.remove_dup_value : 0.000084s : 0.04% optimize.tuple_transform.d_1 : 0.000108s : 0.05% optimize.tuple_transform.none_parameter_eliminate : 0.000002s : 0.00% optimize.tuple_transform.renormalize : 0.000000s : 0.00% optimize.tuple_transform.switch_simplify : 0.000012s : 0.01% optimize.partial_unused_args_eliminate : 0.000002s : 0.00% optimize.add_recomputation : 0.000080s : 0.04% optimize.cse_after_recomputation.cse : 0.000034s : 0.02% optimize.environ_conv : 0.000009s : 0.00% optimize.swap_dp_allreduce_reducescatter : 0.000016s : 0.01% optimize.bias_add_comm_swap : 0.000004s : 0.00% optimize.label_micro_interleaved_index : 0.000009s : 0.00% optimize.label_fine_grained_interleaved_index : 0.000003s : 0.00% optimize.merge_cast_opt : 0.000001s : 0.00% optimize.slice_recompute_activation : 0.000002s : 0.00% optimize.micro_interleaved_order_control : 0.000003s : 0.00% optimize.assign_add_opt : 0.000002s : 0.00% optimize.ForceFp32Comm : 0.000001s : 0.00% optimize.remove_cast_before_assign_add : 0.000001s : 0.00% optimize.full_micro_interleaved_order_control : 0.000002s : 0.00% optimize.reorder_send_recv_between_fp_bp : 0.000003s : 0.00% optimize.comm_op_add_attrs : 0.000001s : 0.00% optimize.add_comm_op_reuse_tag : 0.000001s : 0.00% optimize.interleave_split_concat_branches : 0.000001s : 0.00% optimize.interleave_parallel_branches : 0.000001s : 0.00% optimize.overlap_opt_shard_in_pipeline : 0.000001s : 0.00% optimize.overlap_opt_shard_grad_in_pipeline : 0.000002s : 0.00% optimize.control_data_broadcast_order : 0.000021s : 0.01% optimize.grouped_pairwise_exchange_alltoall : 0.000002s : 0.00% optimize.offloading_packed_experts : 0.000006s : 0.00% optimize.overlap_recompute_and_grad_model_parallel : 0.000007s : 0.00% optimize.overlap_grad_matmul_and_grad_allreduce : 0.000001s : 0.00% optimize.overlap_recompute_allgather_and_fa_grad : 0.000001s : 0.00% optimize.overlap_recompute_comm : 0.000003s : 0.00% optimize.overlap_grad_ring_attention : 0.000006s : 0.00% optimize.overlap_grad_flash_sp : 0.000029s : 0.01% optimize.begin_end_overlap_inline : 0.000001s : 0.00% optimize.split_matmul_comm_elemetwise : 0.000003s : 0.00% optimize.split_layernorm_comm : 0.000002s : 0.00% optimize.handle_group_info : 0.000001s : 0.00% optimize.symbol_engine_optimizer.build : 0.000115s : 0.05% optimize.symbol_engine_optimizer.elim_shapecalc : 0.000031s : 0.01% optimize.symbol_engine_optimizer.elim_not_effective : 0.000024s : 0.01% optimize.symbol_engine_optimizer.opt_reshape : 0.000020s : 0.01% optimize.symbol_engine_optimizer.fold_const_symbol : 0.000025s : 0.01% optimize.symbol_engine_optimizer.renormalize : 0.000000s : 0.00% detach_backward : 0.000002s : 0.00% pipeline_parallel_scheduler : 0.000002s : 0.00% auto_monad_reorder : 0.000030s : 0.01% get_jit_bprop_graph : 0.000003s : 0.00% rewriter_after_jit_bprop_graph : 0.000007s : 0.00% opt_after_jit_grad : 0.000886s : 0.40% validate : 0.000069s : 0.03% Time group info: ------[substitution.] 0.000175 33 2.18% : 0.000004s : 2: substitution.elim_not_effective 5.14% : 0.000009s : 2: substitution.fold_const_symbol 4.44% : 0.000008s : 8: substitution.graph_param_transform 77.28% : 0.000135s : 1: substitution.inline 2.82% : 0.000005s : 4: substitution.j_node_and_user_rematch 3.00% : 0.000005s : 4: substitution.remove_not_recompute_node 5.15% : 0.000009s : 12: substitution.replace_old_param ------[type_inference.] 0.007911 2 86.78% : 0.006865s : 1: type_inference.infer 13.22% : 0.001046s : 1: type_inference.specialize ------[replace.] 0.000024 1 100.00% : 0.000024s : 1: replace.inline ------[match.] 0.000134 1 100.00% : 0.000134s : 1: match.inline ------[predicate.] 0.000246 1879 0.88% : 0.000002s : 17: predicate.accumulaten_eliminater 0.89% : 0.000002s : 8: predicate.ad_related_special_op_eliminate 0.67% : 0.000002s : 16: predicate.addn_check_dump 0.76% : 0.000002s : 17: predicate.addn_zero_filter 0.81% : 0.000002s : 17: predicate.adjust_all_reduce_mul_add 2.05% : 0.000005s : 33: predicate.arithmetic_simplify 0.85% : 0.000002s : 17: predicate.cast_eliminate 0.75% : 0.000002s : 16: predicate.check_bprop_eliminate 0.65% : 0.000002s : 16: predicate.compare_switch_simplify 0.27% : 0.000001s : 8: predicate.const_output_eliminate 0.73% : 0.000002s : 16: predicate.depend_value_elim 0.77% : 0.000002s : 17: predicate.dict_get_item_const_eliminator 0.95% : 0.000002s : 17: predicate.dict_get_item_eliminator 0.72% : 0.000002s : 17: predicate.dict_set_item_eliminator 1.45% : 0.000004s : 16: predicate.dumpgradient_eliminate 0.32% : 0.000001s : 8: predicate.elim_not_effective 0.64% : 0.000002s : 8: predicate.elim_shapecalc_of_broadcastargs 1.14% : 0.000003s : 25: predicate.environ_add_const_eliminate 0.98% : 0.000002s : 25: predicate.environ_get_add_eliminate 1.01% : 0.000002s : 25: predicate.environ_get_depend_swap 1.76% : 0.000004s : 41: predicate.environ_get_eliminate 1.13% : 0.000003s : 25: predicate.environ_get_set_eliminate 0.78% : 0.000002s : 18: predicate.exchange_switch_depend_value 1.44% : 0.000004s : 18: predicate.float_depend_g_call 0.69% : 0.000002s : 16: predicate.float_environ_get_switch 1.09% : 0.000003s : 24: predicate.float_tuple_getitem_switch 0.27% : 0.000001s : 8: predicate.fold_const_symbol 0.89% : 0.000002s : 16: predicate.get_grad_eliminate 0.38% : 0.000001s : 8: predicate.graph_param_transform 0.69% : 0.000002s : 16: predicate.incorporate_call 0.60% : 0.000001s : 16: predicate.incorporate_call_switch 5.94% : 0.000015s : 83: predicate.inline 0.80% : 0.000002s : 16: predicate.inline_without_move 0.46% : 0.000001s : 16: predicate.j_node_and_user_rematch 1.22% : 0.000003s : 16: predicate.less_batch_normalization 1.81% : 0.000004s : 33: predicate.list_to_tuple_eliminator_ 2.17% : 0.000005s : 50: predicate.load_eliminater 3.03% : 0.000007s : 8: predicate.loop_unroll_after_grad 1.22% : 0.000003s : 25: predicate.loop_unroll_before_grad 2.12% : 0.000005s : 33: predicate.make_slice_get_slice_eliminator 0.82% : 0.000002s : 16: predicate.merge_addn 0.71% : 0.000002s : 16: predicate.micro_step_allgather_replace 0.69% : 0.000002s : 16: predicate.mini_step_allgather_replace 0.65% : 0.000002s : 17: predicate.minmaximum_grad 1.65% : 0.000004s : 8: predicate.mutable_eliminate 0.42% : 0.000001s : 8: predicate.opt_reshape 0.48% : 0.000001s : 8: predicate.parallel_virtual_node 1.14% : 0.000003s : 18: predicate.partial_defer_inline 1.15% : 0.000003s : 25: predicate.partial_eliminate 0.71% : 0.000002s : 17: predicate.print_const_string_wrapper 0.68% : 0.000002s : 16: predicate.reduce_all_const_elim 1.11% : 0.000003s : 17: predicate.reduce_eliminate 2.15% : 0.000005s : 50: predicate.redundant_stop_gradient_eliminater 0.79% : 0.000002s : 16: predicate.remove_not_recompute_node 1.38% : 0.000003s : 33: predicate.replace_applicator 0.72% : 0.000002s : 16: predicate.replace_old_param 0.52% : 0.000001s : 8: predicate.reset_defer_inline 0.85% : 0.000002s : 17: predicate.reshape_eliminate 0.83% : 0.000002s : 16: predicate.row_tensor_add_zeros_like 0.51% : 0.000001s : 8: predicate.row_tensor_eliminate 0.89% : 0.000002s : 16: predicate.same_eliminate 0.59% : 0.000001s : 16: predicate.set_cell_output_no_recompute 0.96% : 0.000002s : 16: predicate.shard_identity_eliminate 0.93% : 0.000002s : 16: predicate.special_op_eliminate 0.74% : 0.000002s : 16: predicate.specialize_transform 1.38% : 0.000003s : 16: predicate.split_environ_get_set_with_tuple_value 0.92% : 0.000002s : 16: predicate.stack_unstack_eliminate 0.49% : 0.000001s : 8: predicate.switch_call_monad_eliminater 0.86% : 0.000002s : 18: predicate.switch_defer_inline 1.62% : 0.000004s : 34: predicate.switch_layer_defer_inline 3.81% : 0.000009s : 67: predicate.switch_simplify 0.84% : 0.000002s : 17: predicate.tile_eliminate 0.78% : 0.000002s : 17: predicate.transpose_eliminate 1.67% : 0.000004s : 33: predicate.tuple_list_convert_item_index_to_positive 1.51% : 0.000004s : 33: predicate.tuple_list_get_item_const_eliminator 1.36% : 0.000003s : 33: predicate.tuple_list_get_item_depend_reorder 2.98% : 0.000007s : 49: predicate.tuple_list_get_item_eliminator 1.49% : 0.000004s : 33: predicate.tuple_list_get_set_item_eliminator 2.51% : 0.000006s : 49: predicate.tuple_list_set_item_eliminator 1.76% : 0.000004s : 33: predicate.tuple_to_list_eliminator_ 2.11% : 0.000005s : 50: predicate.updatestate_pure_node_eliminater 3.04% : 0.000008s : 66: predicate.updatestate_useless_node_eliminater 0.54% : 0.000001s : 8: predicate.value_based_eliminate 0.82% : 0.000002s : 16: predicate.virtual_dataset_eliminate 0.87% : 0.000002s : 16: predicate.virtual_output_eliminate 0.33% : 0.000001s : 8: predicate.virtual_view_grad_eliminate 0.45% : 0.000001s : 8: predicate.zero_like_fill_zero ------[func_graph_cloner_run.] 0.001270 6 52.22% : 0.000663s : 3: func_graph_cloner_run.FuncGraphClonerGraph 47.78% : 0.000607s : 3: func_graph_cloner_run.FuncGraphSpecializer ------[meta_graph.] 0.000000 0 ------[manager.] 0.000000 0 ------[pynative] 0.000000 0 ------[others.] 0.455907 192 0.00% : 0.000004s : 1: ForceFp32Comm 1.33% : 0.006081s : 1: add_attr 1.33% : 0.006061s : 1: add_attr_with_inline 0.00% : 0.000004s : 1: add_comm_op_reuse_tag 0.02% : 0.000085s : 1: add_recomputation 0.00% : 0.000005s : 1: assign_add_opt 0.01% : 0.000062s : 1: auto_monad 0.01% : 0.000034s : 1: auto_monad_reorder 0.00% : 0.000004s : 1: begin_end_overlap_inline 0.00% : 0.000007s : 1: bias_add_comm_swap 0.21% : 0.000956s : 1: bootstrap 0.01% : 0.000044s : 1: cconv 0.00% : 0.000012s : 1: comm_op_add_attrs 0.01% : 0.000025s : 1: control_data_broadcast_order 0.00% : 0.000013s : 1: convert_after_rewriter 0.01% : 0.000055s : 1: cse_after_recomputation 0.00% : 0.000005s : 1: dataset_repeat_opt 0.00% : 0.000006s : 1: detach_backward 0.00% : 0.000013s : 1: environ_conv 0.00% : 0.000023s : 1: event_method 0.00% : 0.000005s : 1: full_micro_interleaved_order_control 0.00% : 0.000015s : 1: get_jit_bprop_graph 0.00% : 0.000009s : 1: graph_reusing 0.00% : 0.000005s : 1: grouped_pairwise_exchange_alltoall 0.00% : 0.000004s : 1: handle_group_info 0.00% : 0.000006s : 1: inline 0.00% : 0.000006s : 1: insert-virtual-dataset 0.00% : 0.000004s : 1: interleave_parallel_branches 0.00% : 0.000004s : 1: interleave_split_concat_branches 0.00% : 0.000006s : 1: label_fine_grained_interleaved_index 0.00% : 0.000014s : 1: label_micro_interleaved_index 45.39% : 0.206923s : 1: loop_unroll 0.00% : 0.000004s : 1: merge_cast_opt 0.00% : 0.000006s : 1: micro_interleaved_order_control 0.21% : 0.000962s : 1: mutable_eliminate 0.00% : 0.000009s : 1: offloading_packed_experts 0.01% : 0.000055s : 1: opt.transform.loop_unroll_optimizer 0.01% : 0.000034s : 1: opt.transform.mutable_eliminate 0.43% : 0.001958s : 78: opt.transform.opt_a 0.02% : 0.000072s : 1: opt.transform.opt_after_cconv 0.01% : 0.000058s : 1: opt.transform.opt_after_jit_grad 0.07% : 0.000298s : 28: opt.transform.opt_b 0.03% : 0.000118s : 2: opt.transform.opt_trans_graph 0.02% : 0.000096s : 4: opt.transform.symbol_engine_opt 1.03% : 0.004713s : 1: opt_a 0.06% : 0.000289s : 1: opt_after_cconv 0.20% : 0.000900s : 1: opt_after_jit_grad 0.13% : 0.000593s : 1: opt_b 47.12% : 0.214843s : 1: optimize 0.01% : 0.000040s : 1: optimize_parallel_all_gather_comm 0.00% : 0.000010s : 1: order_py_execute_after_rewriter 0.01% : 0.000033s : 1: overlap_grad_flash_sp 0.00% : 0.000004s : 1: overlap_grad_matmul_and_grad_allreduce 0.00% : 0.000017s : 1: overlap_grad_ring_attention 0.00% : 0.000009s : 1: overlap_opt_shard_grad_in_pipeline 0.00% : 0.000004s : 1: overlap_opt_shard_in_pipeline 0.00% : 0.000017s : 1: overlap_param_gather 0.00% : 0.000005s : 1: overlap_recompute_allgather_and_fa_grad 0.00% : 0.000010s : 1: overlap_recompute_and_grad_model_parallel 0.00% : 0.000006s : 1: overlap_recompute_comm 0.00% : 0.000008s : 1: parallel-infer-symbol 0.00% : 0.000004s : 1: parallel-infer-symbol-second 0.00% : 0.000006s : 1: partial_unused_args_eliminate 0.00% : 0.000019s : 1: pipeline_parallel_scheduler 0.00% : 0.000005s : 1: pipeline_split 0.01% : 0.000051s : 1: pre_auto_parallel 0.01% : 0.000040s : 1: py_interpret_to_execute 0.01% : 0.000044s : 1: py_interpret_to_execute_after_opt_a 0.00% : 0.000007s : 1: remove_cast_before_assign_add 0.02% : 0.000090s : 1: remove_dup_value 0.12% : 0.000530s : 1: renormalize.infer 0.16% : 0.000737s : 1: renormalize.specialize 0.00% : 0.000006s : 1: reorder_send_recv_between_fp_bp 0.00% : 0.000010s : 1: rewriter_after_jit_bprop_graph 0.01% : 0.000063s : 1: rewriter_after_opt_a 0.02% : 0.000077s : 1: rewriter_before_opt_a 0.00% : 0.000005s : 1: slice_cell_reuse_recomputed_activation 0.00% : 0.000005s : 1: slice_recompute_activation 0.00% : 0.000005s : 1: split_layernorm_comm 0.00% : 0.000006s : 1: split_matmul_comm_elemetwise 0.00% : 0.000019s : 1: swap_dp_allreduce_reducescatter 0.06% : 0.000273s : 1: symbol_engine_optimizer 0.03% : 0.000156s : 1: tuple_transform 1.77% : 0.008087s : 1: type_inference . [hook] pytest_runtest_teardown:test_paged_attention_mla_combine_cache_norm[1-float16-16] tests/st/infer/ops/test_internal_ops/test_paged_attention.py::test_paged_attention_mla_combine_cache_norm[1-float16-16],max_mem:270.0M [WARNING] ME(168852:281473710628656,MainProcess):2026-01-29-17:38:26.458.720 [mindspore/context.py:1334] For 'context.set_context', the parameter 'device_target' will be deprecated and removed in a future version. Please use the api mindspore.set_device() instead. TotalTime = 0.181914, [21] [bootstrap]: 0.00114941 [type_inference]: 0.0090529 [event_method]: 1.877e-05 [auto_monad]: 6.402e-05 [graph_reusing]: 5.69e-06 [inline]: 2.69999e-06 [add_attr]: 0.1615, [1] [add_attr_with_inline]: 0.161481, [1] [Cycle 1]: 8.509e-05, [2] [tag_attr]: 2.54e-05 [meta_addattr_fg_expand]: 4.03999e-06 [parallel-infer-symbol]: 3.64002e-06 [pre_auto_parallel]: 4.792e-05 [insert-virtual-dataset]: 2.79001e-06 [parallel-infer-symbol-second]: 6.69999e-07 [dataset_repeat_opt]: 1.79e-06 [pipeline_split]: 1.71e-06 [optimize]: 0.00908608, [53] [py_interpret_to_execute]: 4.424e-05 [rewriter_before_opt_a]: 7.994e-05 [opt_a]: 0.00530256, [2] [Cycle 1]: 0.00394907, [45] [expand_dump_flag]: 3.33e-06 [switch_simplify]: 4.497e-05 [loop_unroll]: 4.512e-05 [a_1]: 0.00085697 [with_stream_mark]: 3.393e-05 [recompute_prepare]: 2.286e-05 [updatestate_depend_eliminate]: 9.14e-06 [updatestate_assign_eliminate]: 8.32998e-06 [updatestate_loads_eliminate]: 5.24e-06 [parameter_eliminate]: 2.06998e-06 [a_2]: 0.00019227 [accelerated_algorithm]: 1.394e-05 [shard]: 2.73998e-06 [meta_shard_fg_expand]: 3.58999e-06 [shard_inline]: 1.092e-05 [merge_send_recv]: 1.221e-05 [auto_parallel]: 1.008e-05 [parallel]: 3.63e-05 [flash_sp]: 1.284e-05 [merge_comm]: 5.36998e-06 [allreduce_fusion]: 4.94e-06 [matmul_add_comm_reduction]: 1.28e-05 [allreduce_slice_to_reducescatter]: 9.20001e-07 [virtual_shard_identity]: 2.108e-05 [virtual_dataset]: 1.21e-05 [get_grad_eliminate_]: 1.156e-05 [virtual_output]: 1.263e-05 [merge_forward]: 6.02001e-06 [cell_reuse_recompute_pass]: 1.93002e-06 [offload_activation]: 1.722e-05 [cell_reuse_handle_not_recompute_node_pass]: 3.083e-05 [merge_recompute_call_nodes]: 1.75001e-06 [before_grad]: 1.998e-05 [set_forward_comm_id_for_comm_node_pass]: 8.82e-06 [meta_fg_expand]: 4.94003e-06 [flash_sp_send_recv_attached]: 2.99001e-06 [receive_attached]: 2.78e-06 [after_resolve]: 2.675e-05 [a_after_grad]: 0.00047409 [renormalize]: 0.00130373 [add_forward_monad_depend]: 9.88002e-06 [auto_monad_grad]: 2.37001e-06 [auto_monad_eliminator]: 2.204e-05 [cse]: 7.909e-05 [a_3]: 8.553e-05 [Cycle 2]: 0.0013394, [45] [expand_dump_flag]: 2.91e-06 [switch_simplify]: 1.29e-05 [loop_unroll]: 1.069e-05 [a_1]: 0.00027064 [with_stream_mark]: 2.419e-05 [recompute_prepare]: 1.198e-05 [updatestate_depend_eliminate]: 5.59e-06 [updatestate_assign_eliminate]: 4.83001e-06 [updatestate_loads_eliminate]: 5.04998e-06 [parameter_eliminate]: 2.29999e-06 [a_2]: 0.00012966 [accelerated_algorithm]: 2.704e-05 [shard]: 2.49999e-06 [meta_shard_fg_expand]: 2.74001e-06 [shard_inline]: 1.033e-05 [merge_send_recv]: 1.124e-05 [auto_parallel]: 1.144e-05 [parallel]: 9.71e-06 [flash_sp]: 3.74002e-06 [merge_comm]: 7.83001e-06 [allreduce_fusion]: 5.12e-06 [matmul_add_comm_reduction]: 1.048e-05 [allreduce_slice_to_reducescatter]: 9.89996e-07 [virtual_shard_identity]: 1.494e-05 [virtual_dataset]: 1.175e-05 [get_grad_eliminate_]: 1.413e-05 [virtual_output]: 1.17e-05 [merge_forward]: 6.19001e-06 [cell_reuse_recompute_pass]: 2.68998e-06 [offload_activation]: 1.297e-05 [cell_reuse_handle_not_recompute_node_pass]: 2.831e-05 [merge_recompute_call_nodes]: 1.36998e-06 [before_grad]: 2.337e-05 [set_forward_comm_id_for_comm_node_pass]: 5.46e-06 [meta_fg_expand]: 3.82002e-06 [flash_sp_send_recv_attached]: 2.16998e-06 [receive_attached]: 2.48998e-06 [after_resolve]: 3.236e-05 [a_after_grad]: 3.184e-05 [renormalize]: 6.00121e-08 [add_forward_monad_depend]: 2.63e-06 [auto_monad_grad]: 1.96e-06 [auto_monad_eliminator]: 1.401e-05 [cse]: 6.679e-05 [a_3]: 9.73e-05 [py_interpret_to_execute_after_opt_a]: 5.238e-05 [slice_cell_reuse_recomputed_activation]: 2.17001e-06 [rewriter_after_opt_a]: 5.992e-05 [convert_after_rewriter]: 9.82999e-06 [order_py_execute_after_rewriter]: 6.79001e-06 [mutable_eliminate]: 0.00101421 [opt_b]: 0.00052109, [1] [Cycle 1]: 0.00051127, [7] [b_1]: 0.00031866 [b_2]: 1.391e-05 [updatestate_depend_eliminate]: 1.223e-05 [updatestate_assign_eliminate]: 4.98001e-06 [updatestate_loads_eliminate]: 7.96001e-06 [renormalize]: 7.29982e-07 [cse]: 0.00010337 [optimize_parallel_all_gather_comm]: 2.862e-05 [overlap_param_gather]: 2.70002e-06 [cconv]: 4.275e-05 [loop_unroll]: 0.00068492 [opt_after_cconv]: 0.0002391, [1] [Cycle 1]: 0.00023166, [7] [c_1]: 9.155e-05 [parameter_eliminate]: 5.78997e-06 [updatestate_depend_eliminate]: 9.05999e-06 [updatestate_assign_eliminate]: 4.66002e-06 [updatestate_loads_eliminate]: 4.57e-06 [cse]: 6.644e-05 [renormalize]: 9.20001e-07 [remove_dup_value]: 7.536e-05 [tuple_transform]: 0.0001451, [1] [Cycle 1]: 0.00013913, [4] [d_1]: 9.877e-05 [none_parameter_eliminate]: 2.56998e-06 [renormalize]: 2.40019e-07 [switch_simplify]: 1.207e-05 [partial_unused_args_eliminate]: 1.89e-06 [add_recomputation]: 8.577e-05 [cse_after_recomputation]: 4.123e-05, [1] [Cycle 1]: 3.618e-05, [1] [cse]: 2.893e-05 [environ_conv]: 9.96e-06 [swap_dp_allreduce_reducescatter]: 6.73e-06 [bias_add_comm_swap]: 3.23e-06 [label_micro_interleaved_index]: 7.2e-06 [label_fine_grained_interleaved_index]: 2.54999e-06 [merge_cast_opt]: 1.56002e-06 [slice_recompute_activation]: 2.20002e-06 [micro_interleaved_order_control]: 2.53e-06 [assign_add_opt]: 1.28002e-06 [ForceFp32Comm]: 9.50007e-07 [remove_cast_before_assign_add]: 1.31002e-06 [full_micro_interleaved_order_control]: 2.35002e-06 [reorder_send_recv_between_fp_bp]: 3.14001e-06 [comm_op_add_attrs]: 1.09e-06 [add_comm_op_reuse_tag]: 1.00999e-06 [interleave_split_concat_branches]: 1.29e-06 [interleave_parallel_branches]: 1.05001e-06 [overlap_opt_shard_in_pipeline]: 1.88997e-06 [overlap_opt_shard_grad_in_pipeline]: 1.72999e-06 [control_data_broadcast_order]: 1.927e-05 [grouped_pairwise_exchange_alltoall]: 1.52999e-06 [offloading_packed_experts]: 6.43998e-06 [overlap_recompute_and_grad_model_parallel]: 6.36e-06 [overlap_grad_matmul_and_grad_allreduce]: 1.29e-06 [overlap_recompute_allgather_and_fa_grad]: 1.32999e-06 [overlap_recompute_comm]: 2.55002e-06 [overlap_grad_ring_attention]: 6.02999e-06 [overlap_grad_flash_sp]: 2.549e-05 [begin_end_overlap_inline]: 4.89992e-07 [split_matmul_comm_elemetwise]: 2.05002e-06 [split_layernorm_comm]: 2.11998e-06 [handle_group_info]: 1.17e-06 [symbol_engine_optimizer]: 0.000216, [1] [Cycle 1]: 0.00021025, [6] [build]: 9.021e-05 [elim_shapecalc]: 2.079e-05 [elim_not_effective]: 2.139e-05 [opt_reshape]: 1.3e-05 [fold_const_symbol]: 2.278e-05 [renormalize]: 2.19996e-07 [detach_backward]: 2.71e-06 [pipeline_parallel_scheduler]: 1.44e-06 [auto_monad_reorder]: 3.122e-05 [get_jit_bprop_graph]: 2.21e-06 [rewriter_after_jit_bprop_graph]: 7.21001e-06 [opt_after_jit_grad]: 0.00066179 [validate]: 6.132e-05 Sums bootstrap : 0.001149s : 6.08% type_inference : 0.009053s : 47.88% event_method : 0.000019s : 0.10% auto_monad : 0.000064s : 0.34% graph_reusing : 0.000006s : 0.03% inline : 0.000003s : 0.01% add_attr.add_attr_with_inline.tag_attr : 0.000025s : 0.13% add_attr.add_attr_with_inline.meta_addattr_fg_expand : 0.000004s : 0.02% parallel-infer-symbol : 0.000004s : 0.02% pre_auto_parallel : 0.000048s : 0.25% insert-virtual-dataset : 0.000003s : 0.01% parallel-infer-symbol-second : 0.000001s : 0.00% dataset_repeat_opt : 0.000002s : 0.01% pipeline_split : 0.000002s : 0.01% optimize.py_interpret_to_execute : 0.000044s : 0.23% optimize.rewriter_before_opt_a : 0.000080s : 0.42% optimize.opt_a.expand_dump_flag : 0.000006s : 0.03% optimize.opt_a.switch_simplify : 0.000058s : 0.31% optimize.opt_a.loop_unroll : 0.000056s : 0.30% optimize.opt_a.a_1 : 0.001128s : 5.96% optimize.opt_a.with_stream_mark : 0.000058s : 0.31% optimize.opt_a.recompute_prepare : 0.000035s : 0.18% optimize.opt_a.updatestate_depend_eliminate : 0.000015s : 0.08% optimize.opt_a.updatestate_assign_eliminate : 0.000013s : 0.07% optimize.opt_a.updatestate_loads_eliminate : 0.000010s : 0.05% optimize.opt_a.parameter_eliminate : 0.000004s : 0.02% optimize.opt_a.a_2 : 0.000322s : 1.70% optimize.opt_a.accelerated_algorithm : 0.000041s : 0.22% optimize.opt_a.shard : 0.000005s : 0.03% optimize.opt_a.meta_shard_fg_expand : 0.000006s : 0.03% optimize.opt_a.shard_inline : 0.000021s : 0.11% optimize.opt_a.merge_send_recv : 0.000023s : 0.12% optimize.opt_a.auto_parallel : 0.000022s : 0.11% optimize.opt_a.parallel : 0.000046s : 0.24% optimize.opt_a.flash_sp : 0.000017s : 0.09% optimize.opt_a.merge_comm : 0.000013s : 0.07% optimize.opt_a.allreduce_fusion : 0.000010s : 0.05% optimize.opt_a.matmul_add_comm_reduction : 0.000023s : 0.12% optimize.opt_a.allreduce_slice_to_reducescatter : 0.000002s : 0.01% optimize.opt_a.virtual_shard_identity : 0.000036s : 0.19% optimize.opt_a.virtual_dataset : 0.000024s : 0.13% optimize.opt_a.get_grad_eliminate_ : 0.000026s : 0.14% optimize.opt_a.virtual_output : 0.000024s : 0.13% optimize.opt_a.merge_forward : 0.000012s : 0.06% optimize.opt_a.cell_reuse_recompute_pass : 0.000005s : 0.02% optimize.opt_a.offload_activation : 0.000030s : 0.16% optimize.opt_a.cell_reuse_handle_not_recompute_node_pass : 0.000059s : 0.31% optimize.opt_a.merge_recompute_call_nodes : 0.000003s : 0.02% optimize.opt_a.before_grad : 0.000043s : 0.23% optimize.opt_a.set_forward_comm_id_for_comm_node_pass : 0.000014s : 0.08% optimize.opt_a.meta_fg_expand : 0.000009s : 0.05% optimize.opt_a.flash_sp_send_recv_attached : 0.000005s : 0.03% optimize.opt_a.receive_attached : 0.000005s : 0.03% optimize.opt_a.after_resolve : 0.000059s : 0.31% optimize.opt_a.a_after_grad : 0.000506s : 2.68% optimize.opt_a.renormalize : 0.001304s : 6.90% optimize.opt_a.add_forward_monad_depend : 0.000013s : 0.07% optimize.opt_a.auto_monad_grad : 0.000004s : 0.02% optimize.opt_a.auto_monad_eliminator : 0.000036s : 0.19% optimize.opt_a.cse : 0.000146s : 0.77% optimize.opt_a.a_3 : 0.000183s : 0.97% optimize.py_interpret_to_execute_after_opt_a : 0.000052s : 0.28% optimize.slice_cell_reuse_recomputed_activation : 0.000002s : 0.01% optimize.rewriter_after_opt_a : 0.000060s : 0.32% optimize.convert_after_rewriter : 0.000010s : 0.05% optimize.order_py_execute_after_rewriter : 0.000007s : 0.04% optimize.mutable_eliminate : 0.001014s : 5.36% optimize.opt_b.b_1 : 0.000319s : 1.69% optimize.opt_b.b_2 : 0.000014s : 0.07% optimize.opt_b.updatestate_depend_eliminate : 0.000012s : 0.06% optimize.opt_b.updatestate_assign_eliminate : 0.000005s : 0.03% optimize.opt_b.updatestate_loads_eliminate : 0.000008s : 0.04% optimize.opt_b.renormalize : 0.000001s : 0.00% optimize.opt_b.cse : 0.000103s : 0.55% optimize.optimize_parallel_all_gather_comm : 0.000029s : 0.15% optimize.overlap_param_gather : 0.000003s : 0.01% optimize.cconv : 0.000043s : 0.23% optimize.loop_unroll : 0.000685s : 3.62% optimize.opt_after_cconv.c_1 : 0.000092s : 0.48% optimize.opt_after_cconv.parameter_eliminate : 0.000006s : 0.03% optimize.opt_after_cconv.updatestate_depend_eliminate : 0.000009s : 0.05% optimize.opt_after_cconv.updatestate_assign_eliminate : 0.000005s : 0.02% optimize.opt_after_cconv.updatestate_loads_eliminate : 0.000005s : 0.02% optimize.opt_after_cconv.cse : 0.000066s : 0.35% optimize.opt_after_cconv.renormalize : 0.000001s : 0.00% optimize.remove_dup_value : 0.000075s : 0.40% optimize.tuple_transform.d_1 : 0.000099s : 0.52% optimize.tuple_transform.none_parameter_eliminate : 0.000003s : 0.01% optimize.tuple_transform.renormalize : 0.000000s : 0.00% optimize.tuple_transform.switch_simplify : 0.000012s : 0.06% optimize.partial_unused_args_eliminate : 0.000002s : 0.01% optimize.add_recomputation : 0.000086s : 0.45% optimize.cse_after_recomputation.cse : 0.000029s : 0.15% optimize.environ_conv : 0.000010s : 0.05% optimize.swap_dp_allreduce_reducescatter : 0.000007s : 0.04% optimize.bias_add_comm_swap : 0.000003s : 0.02% optimize.label_micro_interleaved_index : 0.000007s : 0.04% optimize.label_fine_grained_interleaved_index : 0.000003s : 0.01% optimize.merge_cast_opt : 0.000002s : 0.01% optimize.slice_recompute_activation : 0.000002s : 0.01% optimize.micro_interleaved_order_control : 0.000003s : 0.01% optimize.assign_add_opt : 0.000001s : 0.01% optimize.ForceFp32Comm : 0.000001s : 0.01% optimize.remove_cast_before_assign_add : 0.000001s : 0.01% optimize.full_micro_interleaved_order_control : 0.000002s : 0.01% optimize.reorder_send_recv_between_fp_bp : 0.000003s : 0.02% optimize.comm_op_add_attrs : 0.000001s : 0.01% optimize.add_comm_op_reuse_tag : 0.000001s : 0.01% optimize.interleave_split_concat_branches : 0.000001s : 0.01% optimize.interleave_parallel_branches : 0.000001s : 0.01% optimize.overlap_opt_shard_in_pipeline : 0.000002s : 0.01% optimize.overlap_opt_shard_grad_in_pipeline : 0.000002s : 0.01% optimize.control_data_broadcast_order : 0.000019s : 0.10% optimize.grouped_pairwise_exchange_alltoall : 0.000002s : 0.01% optimize.offloading_packed_experts : 0.000006s : 0.03% optimize.overlap_recompute_and_grad_model_parallel : 0.000006s : 0.03% optimize.overlap_grad_matmul_and_grad_allreduce : 0.000001s : 0.01% optimize.overlap_recompute_allgather_and_fa_grad : 0.000001s : 0.01% optimize.overlap_recompute_comm : 0.000003s : 0.01% optimize.overlap_grad_ring_attention : 0.000006s : 0.03% optimize.overlap_grad_flash_sp : 0.000025s : 0.13% optimize.begin_end_overlap_inline : 0.000000s : 0.00% optimize.split_matmul_comm_elemetwise : 0.000002s : 0.01% optimize.split_layernorm_comm : 0.000002s : 0.01% optimize.handle_group_info : 0.000001s : 0.01% optimize.symbol_engine_optimizer.build : 0.000090s : 0.48% optimize.symbol_engine_optimizer.elim_shapecalc : 0.000021s : 0.11% optimize.symbol_engine_optimizer.elim_not_effective : 0.000021s : 0.11% optimize.symbol_engine_optimizer.opt_reshape : 0.000013s : 0.07% optimize.symbol_engine_optimizer.fold_const_symbol : 0.000023s : 0.12% optimize.symbol_engine_optimizer.renormalize : 0.000000s : 0.00% detach_backward : 0.000003s : 0.01% pipeline_parallel_scheduler : 0.000001s : 0.01% auto_monad_reorder : 0.000031s : 0.17% get_jit_bprop_graph : 0.000002s : 0.01% rewriter_after_jit_bprop_graph : 0.000007s : 0.04% opt_after_jit_grad : 0.000662s : 3.50% validate : 0.000061s : 0.32% Time group info: ------[substitution.] 0.000225 33 1.65% : 0.000004s : 2: substitution.elim_not_effective 3.95% : 0.000009s : 2: substitution.fold_const_symbol 3.39% : 0.000008s : 8: substitution.graph_param_transform 82.02% : 0.000184s : 1: substitution.inline 2.26% : 0.000005s : 4: substitution.j_node_and_user_rematch 2.34% : 0.000005s : 4: substitution.remove_not_recompute_node 4.39% : 0.000010s : 12: substitution.replace_old_param ------[type_inference.] 0.008897 2 84.50% : 0.007519s : 1: type_inference.infer 15.50% : 0.001379s : 1: type_inference.specialize ------[replace.] 0.000028 1 100.00% : 0.000028s : 1: replace.inline ------[match.] 0.000183 1 100.00% : 0.000183s : 1: match.inline ------[predicate.] 0.000689 1879 0.31% : 0.000002s : 17: predicate.accumulaten_eliminater 0.40% : 0.000003s : 8: predicate.ad_related_special_op_eliminate 0.32% : 0.000002s : 16: predicate.addn_check_dump 0.31% : 0.000002s : 17: predicate.addn_zero_filter 0.26% : 0.000002s : 17: predicate.adjust_all_reduce_mul_add 0.78% : 0.000005s : 33: predicate.arithmetic_simplify 0.28% : 0.000002s : 17: predicate.cast_eliminate 0.32% : 0.000002s : 16: predicate.check_bprop_eliminate 0.25% : 0.000002s : 16: predicate.compare_switch_simplify 0.09% : 0.000001s : 8: predicate.const_output_eliminate 0.25% : 0.000002s : 16: predicate.depend_value_elim 0.29% : 0.000002s : 17: predicate.dict_get_item_const_eliminator 0.33% : 0.000002s : 17: predicate.dict_get_item_eliminator 0.28% : 0.000002s : 17: predicate.dict_set_item_eliminator 0.48% : 0.000003s : 16: predicate.dumpgradient_eliminate 0.13% : 0.000001s : 8: predicate.elim_not_effective 0.20% : 0.000001s : 8: predicate.elim_shapecalc_of_broadcastargs 0.46% : 0.000003s : 25: predicate.environ_add_const_eliminate 0.40% : 0.000003s : 25: predicate.environ_get_add_eliminate 0.41% : 0.000003s : 25: predicate.environ_get_depend_swap 0.68% : 0.000005s : 41: predicate.environ_get_eliminate 0.40% : 0.000003s : 25: predicate.environ_get_set_eliminate 0.29% : 0.000002s : 18: predicate.exchange_switch_depend_value 0.51% : 0.000004s : 18: predicate.float_depend_g_call 0.25% : 0.000002s : 16: predicate.float_environ_get_switch 0.42% : 0.000003s : 24: predicate.float_tuple_getitem_switch 0.09% : 0.000001s : 8: predicate.fold_const_symbol 0.30% : 0.000002s : 16: predicate.get_grad_eliminate 0.13% : 0.000001s : 8: predicate.graph_param_transform 0.24% : 0.000002s : 16: predicate.incorporate_call 0.21% : 0.000001s : 16: predicate.incorporate_call_switch 1.85% : 0.000013s : 83: predicate.inline 64.01% : 0.000441s : 16: predicate.inline_without_move 0.17% : 0.000001s : 16: predicate.j_node_and_user_rematch 0.38% : 0.000003s : 16: predicate.less_batch_normalization 0.67% : 0.000005s : 33: predicate.list_to_tuple_eliminator_ 0.81% : 0.000006s : 50: predicate.load_eliminater 0.49% : 0.000003s : 8: predicate.loop_unroll_after_grad 0.45% : 0.000003s : 25: predicate.loop_unroll_before_grad 0.68% : 0.000005s : 33: predicate.make_slice_get_slice_eliminator 0.28% : 0.000002s : 16: predicate.merge_addn 0.26% : 0.000002s : 16: predicate.micro_step_allgather_replace 0.29% : 0.000002s : 16: predicate.mini_step_allgather_replace 0.26% : 0.000002s : 17: predicate.minmaximum_grad 0.74% : 0.000005s : 8: predicate.mutable_eliminate 0.17% : 0.000001s : 8: predicate.opt_reshape 0.18% : 0.000001s : 8: predicate.parallel_virtual_node 0.38% : 0.000003s : 18: predicate.partial_defer_inline 0.41% : 0.000003s : 25: predicate.partial_eliminate 0.30% : 0.000002s : 17: predicate.print_const_string_wrapper 0.27% : 0.000002s : 16: predicate.reduce_all_const_elim 0.49% : 0.000003s : 17: predicate.reduce_eliminate 0.85% : 0.000006s : 50: predicate.redundant_stop_gradient_eliminater 0.24% : 0.000002s : 16: predicate.remove_not_recompute_node 0.47% : 0.000003s : 33: predicate.replace_applicator 0.22% : 0.000001s : 16: predicate.replace_old_param 0.17% : 0.000001s : 8: predicate.reset_defer_inline 0.32% : 0.000002s : 17: predicate.reshape_eliminate 0.30% : 0.000002s : 16: predicate.row_tensor_add_zeros_like 0.17% : 0.000001s : 8: predicate.row_tensor_eliminate 0.36% : 0.000002s : 16: predicate.same_eliminate 0.29% : 0.000002s : 16: predicate.set_cell_output_no_recompute 0.42% : 0.000003s : 16: predicate.shard_identity_eliminate 0.31% : 0.000002s : 16: predicate.special_op_eliminate 0.27% : 0.000002s : 16: predicate.specialize_transform 0.39% : 0.000003s : 16: predicate.split_environ_get_set_with_tuple_value 0.56% : 0.000004s : 16: predicate.stack_unstack_eliminate 0.15% : 0.000001s : 8: predicate.switch_call_monad_eliminater 0.30% : 0.000002s : 18: predicate.switch_defer_inline 0.59% : 0.000004s : 34: predicate.switch_layer_defer_inline 1.33% : 0.000009s : 67: predicate.switch_simplify 0.38% : 0.000003s : 17: predicate.tile_eliminate 0.31% : 0.000002s : 17: predicate.transpose_eliminate 0.66% : 0.000005s : 33: predicate.tuple_list_convert_item_index_to_positive 0.54% : 0.000004s : 33: predicate.tuple_list_get_item_const_eliminator 0.56% : 0.000004s : 33: predicate.tuple_list_get_item_depend_reorder 1.17% : 0.000008s : 49: predicate.tuple_list_get_item_eliminator 0.59% : 0.000004s : 33: predicate.tuple_list_get_set_item_eliminator 0.95% : 0.000007s : 49: predicate.tuple_list_set_item_eliminator 0.58% : 0.000004s : 33: predicate.tuple_to_list_eliminator_ 0.76% : 0.000005s : 50: predicate.updatestate_pure_node_eliminater 1.05% : 0.000007s : 66: predicate.updatestate_useless_node_eliminater 0.17% : 0.000001s : 8: predicate.value_based_eliminate 0.29% : 0.000002s : 16: predicate.virtual_dataset_eliminate 0.31% : 0.000002s : 16: predicate.virtual_output_eliminate 0.12% : 0.000001s : 8: predicate.virtual_view_grad_eliminate 0.20% : 0.000001s : 8: predicate.zero_like_fill_zero ------[func_graph_cloner_run.] 0.001442 6 46.73% : 0.000674s : 3: func_graph_cloner_run.FuncGraphClonerGraph 53.27% : 0.000768s : 3: func_graph_cloner_run.FuncGraphSpecializer ------[meta_graph.] 0.000000 0 ------[manager.] 0.000000 0 ------[pynative] 0.000000 0 ------[others.] 0.356784 192 0.00% : 0.000004s : 1: ForceFp32Comm 45.27% : 0.161507s : 1: add_attr 45.26% : 0.161486s : 1: add_attr_with_inline 0.00% : 0.000004s : 1: add_comm_op_reuse_tag 0.03% : 0.000090s : 1: add_recomputation 0.00% : 0.000005s : 1: assign_add_opt 0.02% : 0.000070s : 1: auto_monad 0.01% : 0.000036s : 1: auto_monad_reorder 0.00% : 0.000003s : 1: begin_end_overlap_inline 0.00% : 0.000006s : 1: bias_add_comm_swap 0.34% : 0.001196s : 1: bootstrap 0.01% : 0.000052s : 1: cconv 0.00% : 0.000004s : 1: comm_op_add_attrs 0.01% : 0.000023s : 1: control_data_broadcast_order 0.00% : 0.000013s : 1: convert_after_rewriter 0.01% : 0.000044s : 1: cse_after_recomputation 0.00% : 0.000005s : 1: dataset_repeat_opt 0.00% : 0.000006s : 1: detach_backward 0.00% : 0.000013s : 1: environ_conv 0.01% : 0.000026s : 1: event_method 0.00% : 0.000005s : 1: full_micro_interleaved_order_control 0.00% : 0.000006s : 1: get_jit_bprop_graph 0.00% : 0.000010s : 1: graph_reusing 0.00% : 0.000004s : 1: grouped_pairwise_exchange_alltoall 0.00% : 0.000004s : 1: handle_group_info 0.00% : 0.000006s : 1: inline 0.00% : 0.000006s : 1: insert-virtual-dataset 0.00% : 0.000004s : 1: interleave_parallel_branches 0.00% : 0.000004s : 1: interleave_split_concat_branches 0.00% : 0.000007s : 1: label_fine_grained_interleaved_index 0.00% : 0.000010s : 1: label_micro_interleaved_index 0.20% : 0.000697s : 1: loop_unroll 0.00% : 0.000004s : 1: merge_cast_opt 0.00% : 0.000009s : 1: micro_interleaved_order_control 0.29% : 0.001032s : 1: mutable_eliminate 0.00% : 0.000010s : 1: offloading_packed_experts 0.01% : 0.000027s : 1: opt.transform.loop_unroll_optimizer 0.01% : 0.000037s : 1: opt.transform.mutable_eliminate 0.70% : 0.002499s : 78: opt.transform.opt_a 0.03% : 0.000090s : 1: opt.transform.opt_after_cconv 0.01% : 0.000048s : 1: opt.transform.opt_after_jit_grad 0.08% : 0.000288s : 28: opt.transform.opt_b 0.03% : 0.000108s : 2: opt.transform.opt_trans_graph 0.02% : 0.000073s : 4: opt.transform.symbol_engine_opt 1.49% : 0.005307s : 1: opt_a 0.07% : 0.000243s : 1: opt_after_cconv 0.19% : 0.000676s : 1: opt_after_jit_grad 0.15% : 0.000530s : 1: opt_b 2.55% : 0.009094s : 1: optimize 0.01% : 0.000037s : 1: optimize_parallel_all_gather_comm 0.00% : 0.000010s : 1: order_py_execute_after_rewriter 0.01% : 0.000029s : 1: overlap_grad_flash_sp 0.00% : 0.000004s : 1: overlap_grad_matmul_and_grad_allreduce 0.00% : 0.000013s : 1: overlap_grad_ring_attention 0.00% : 0.000010s : 1: overlap_opt_shard_grad_in_pipeline 0.00% : 0.000009s : 1: overlap_opt_shard_in_pipeline 0.00% : 0.000007s : 1: overlap_param_gather 0.00% : 0.000004s : 1: overlap_recompute_allgather_and_fa_grad 0.00% : 0.000010s : 1: overlap_recompute_and_grad_model_parallel 0.00% : 0.000005s : 1: overlap_recompute_comm 0.00% : 0.000007s : 1: parallel-infer-symbol 0.00% : 0.000004s : 1: parallel-infer-symbol-second 0.00% : 0.000005s : 1: partial_unused_args_eliminate 0.00% : 0.000010s : 1: pipeline_parallel_scheduler 0.00% : 0.000005s : 1: pipeline_split 0.01% : 0.000053s : 1: pre_auto_parallel 0.01% : 0.000049s : 1: py_interpret_to_execute 0.02% : 0.000057s : 1: py_interpret_to_execute_after_opt_a 0.00% : 0.000012s : 1: remove_cast_before_assign_add 0.02% : 0.000080s : 1: remove_dup_value 0.16% : 0.000556s : 1: renormalize.infer 0.21% : 0.000733s : 1: renormalize.specialize 0.00% : 0.000006s : 1: reorder_send_recv_between_fp_bp 0.00% : 0.000011s : 1: rewriter_after_jit_bprop_graph 0.02% : 0.000064s : 1: rewriter_after_opt_a 0.02% : 0.000084s : 1: rewriter_before_opt_a 0.00% : 0.000006s : 1: slice_cell_reuse_recomputed_activation 0.00% : 0.000005s : 1: slice_recompute_activation 0.00% : 0.000006s : 1: split_layernorm_comm 0.00% : 0.000005s : 1: split_matmul_comm_elemetwise 0.00% : 0.000010s : 1: swap_dp_allreduce_reducescatter 0.06% : 0.000219s : 1: symbol_engine_optimizer 0.04% : 0.000148s : 1: tuple_transform 2.54% : 0.009080s : 1: type_inference . [hook] pytest_runtest_teardown:test_paged_attention_mla_combine_cache_norm[1-float16-128] tests/st/infer/ops/test_internal_ops/test_paged_attention.py::test_paged_attention_mla_combine_cache_norm[1-float16-128],max_mem:270.0M [WARNING] ME(168852:281473710628656,MainProcess):2026-01-29-17:38:33.657.037 [mindspore/context.py:1334] For 'context.set_context', the parameter 'device_target' will be deprecated and removed in a future version. Please use the api mindspore.set_device() instead. TotalTime = 0.209617, [21] [bootstrap]: 0.00105276 [type_inference]: 0.0111909 [event_method]: 1.695e-05 [auto_monad]: 8.939e-05 [graph_reusing]: 5.37999e-06 [inline]: 2.91999e-06 [add_attr]: 0.00592542, [1] [add_attr_with_inline]: 0.00590867, [1] [Cycle 1]: 7.354e-05, [2] [tag_attr]: 2.144e-05 [meta_addattr_fg_expand]: 4.07e-06 [parallel-infer-symbol]: 3.21001e-06 [pre_auto_parallel]: 4.459e-05 [insert-virtual-dataset]: 2.56e-06 [parallel-infer-symbol-second]: 6.80011e-07 [dataset_repeat_opt]: 1.96e-06 [pipeline_split]: 1.81e-06 [optimize]: 0.1898, [53] [py_interpret_to_execute]: 3.409e-05 [rewriter_before_opt_a]: 6.886e-05 [opt_a]: 0.00467033, [2] [Cycle 1]: 0.00326806, [45] [expand_dump_flag]: 2.83e-06 [switch_simplify]: 4.45e-05 [loop_unroll]: 2.772e-05 [a_1]: 0.00091731 [with_stream_mark]: 2.707e-05 [recompute_prepare]: 1.824e-05 [updatestate_depend_eliminate]: 6.17001e-06 [updatestate_assign_eliminate]: 8.62e-06 [updatestate_loads_eliminate]: 5.17e-06 [parameter_eliminate]: 1.97001e-06 [a_2]: 0.00016042 [accelerated_algorithm]: 1.299e-05 [shard]: 2.69001e-06 [meta_shard_fg_expand]: 2.91e-06 [shard_inline]: 1.054e-05 [merge_send_recv]: 1.074e-05 [auto_parallel]: 8.3e-06 [parallel]: 3.011e-05 [flash_sp]: 1.035e-05 [merge_comm]: 5.37001e-06 [allreduce_fusion]: 5.15001e-06 [matmul_add_comm_reduction]: 1.218e-05 [allreduce_slice_to_reducescatter]: 7.40023e-07 [virtual_shard_identity]: 1.42e-05 [virtual_dataset]: 1.27e-05 [get_grad_eliminate_]: 1.112e-05 [virtual_output]: 1.125e-05 [merge_forward]: 6.78998e-06 [cell_reuse_recompute_pass]: 1.79998e-06 [offload_activation]: 1.35e-05 [cell_reuse_handle_not_recompute_node_pass]: 1.899e-05 [merge_recompute_call_nodes]: 1.38002e-06 [before_grad]: 1.708e-05 [set_forward_comm_id_for_comm_node_pass]: 5.36002e-06 [meta_fg_expand]: 4.35e-06 [flash_sp_send_recv_attached]: 2.75997e-06 [receive_attached]: 2.58e-06 [after_resolve]: 2.986e-05 [a_after_grad]: 2.08e-05 [renormalize]: 0.00115654 [add_forward_monad_depend]: 6.45002e-06 [auto_monad_grad]: 2.46998e-06 [auto_monad_eliminator]: 2.594e-05 [cse]: 8.385e-05 [a_3]: 8.442e-05 [Cycle 2]: 0.00138619, [45] [expand_dump_flag]: 2.17999e-06 [switch_simplify]: 1.242e-05 [loop_unroll]: 1.043e-05 [a_1]: 0.00024705 [with_stream_mark]: 2.322e-05 [recompute_prepare]: 1.065e-05 [updatestate_depend_eliminate]: 5.89e-06 [updatestate_assign_eliminate]: 4.45999e-06 [updatestate_loads_eliminate]: 5.01997e-06 [parameter_eliminate]: 1.37e-06 [a_2]: 0.00013121 [accelerated_algorithm]: 1.838e-05 [shard]: 1.90001e-06 [meta_shard_fg_expand]: 2.07999e-06 [shard_inline]: 1.626e-05 [merge_send_recv]: 1.003e-05 [auto_parallel]: 1.635e-05 [parallel]: 9.47001e-06 [flash_sp]: 4.05e-06 [merge_comm]: 5.52999e-06 [allreduce_fusion]: 7.15e-06 [matmul_add_comm_reduction]: 1.118e-05 [allreduce_slice_to_reducescatter]: 5.69999e-07 [virtual_shard_identity]: 1.36e-05 [virtual_dataset]: 1.433e-05 [get_grad_eliminate_]: 1.27e-05 [virtual_output]: 1.07e-05 [merge_forward]: 5.34e-06 [cell_reuse_recompute_pass]: 1.93002e-06 [offload_activation]: 1.355e-05 [cell_reuse_handle_not_recompute_node_pass]: 2.012e-05 [merge_recompute_call_nodes]: 1.22e-06 [before_grad]: 1.659e-05 [set_forward_comm_id_for_comm_node_pass]: 5.42001e-06 [meta_fg_expand]: 3.53e-06 [flash_sp_send_recv_attached]: 9.00007e-07 [receive_attached]: 1.58002e-06 [after_resolve]: 2.316e-05 [a_after_grad]: 1.756e-05 [renormalize]: 9.00181e-08 [add_forward_monad_depend]: 2.62001e-06 [auto_monad_grad]: 1.81998e-06 [auto_monad_eliminator]: 1.181e-05 [cse]: 6.559e-05 [a_3]: 7.614e-05 [py_interpret_to_execute_after_opt_a]: 3.781e-05 [slice_cell_reuse_recomputed_activation]: 1.87999e-06 [rewriter_after_opt_a]: 6.318e-05 [convert_after_rewriter]: 9.39e-06 [order_py_execute_after_rewriter]: 7.58001e-06 [mutable_eliminate]: 0.00075531 [opt_b]: 0.00058609, [1] [Cycle 1]: 0.00057387, [7] [b_1]: 0.0003526 [b_2]: 1.441e-05 [updatestate_depend_eliminate]: 1.035e-05 [updatestate_assign_eliminate]: 4.87e-06 [updatestate_loads_eliminate]: 4.38999e-06 [renormalize]: 6.39993e-07 [cse]: 7.928e-05 [optimize_parallel_all_gather_comm]: 3.207e-05 [overlap_param_gather]: 2.03002e-06 [cconv]: 3.814e-05 [loop_unroll]: 0.00077192 [opt_after_cconv]: 0.00024771, [1] [Cycle 1]: 0.00023999, [7] [c_1]: 6.781e-05 [parameter_eliminate]: 5.05001e-06 [updatestate_depend_eliminate]: 8.18001e-06 [updatestate_assign_eliminate]: 4.36002e-06 [updatestate_loads_eliminate]: 4.50001e-06 [cse]: 6.846e-05 [renormalize]: 4.30009e-07 [remove_dup_value]: 8.641e-05 [tuple_transform]: 0.0001866, [1] [Cycle 1]: 0.00017491, [4] [d_1]: 0.00010094 [none_parameter_eliminate]: 2.04999e-06 [renormalize]: 2.3999e-07 [switch_simplify]: 1.455e-05 [partial_unused_args_eliminate]: 1.92001e-06 [add_recomputation]: 8.258e-05 [cse_after_recomputation]: 5.919e-05, [1] [Cycle 1]: 5.137e-05, [1] [cse]: 4.032e-05 [environ_conv]: 8.84998e-06 [swap_dp_allreduce_reducescatter]: 1.03e-05 [bias_add_comm_swap]: 3.58e-06 [label_micro_interleaved_index]: 5.97999e-06 [label_fine_grained_interleaved_index]: 2.63998e-06 [merge_cast_opt]: 1.47001e-06 [slice_recompute_activation]: 2.13998e-06 [micro_interleaved_order_control]: 2.63e-06 [assign_add_opt]: 1.52001e-06 [ForceFp32Comm]: 9.89996e-07 [remove_cast_before_assign_add]: 1.06002e-06 [full_micro_interleaved_order_control]: 2.28998e-06 [reorder_send_recv_between_fp_bp]: 2.84999e-06 [comm_op_add_attrs]: 1.05001e-06 [add_comm_op_reuse_tag]: 1.00001e-06 [interleave_split_concat_branches]: 1.18001e-06 [interleave_parallel_branches]: 1.07e-06 [overlap_opt_shard_in_pipeline]: 1.20999e-06 [overlap_opt_shard_grad_in_pipeline]: 1.66002e-06 [control_data_broadcast_order]: 1.773e-05 [grouped_pairwise_exchange_alltoall]: 1.45999e-06 [offloading_packed_experts]: 5.76e-06 [overlap_recompute_and_grad_model_parallel]: 6.24999e-06 [overlap_grad_matmul_and_grad_allreduce]: 1.17999e-06 [overlap_recompute_allgather_and_fa_grad]: 1.25999e-06 [overlap_recompute_comm]: 2.61e-06 [overlap_grad_ring_attention]: 6.02001e-06 [overlap_grad_flash_sp]: 3.125e-05 [begin_end_overlap_inline]: 5.69999e-07 [split_matmul_comm_elemetwise]: 2.36e-06 [split_layernorm_comm]: 2.14999e-06 [handle_group_info]: 1.53002e-06 [symbol_engine_optimizer]: 0.18161, [1] [Cycle 1]: 0.181596, [6] [build]: 0.181291 [elim_shapecalc]: 6.055e-05 [elim_not_effective]: 3.994e-05 [opt_reshape]: 1.394e-05 [fold_const_symbol]: 3.396e-05 [renormalize]: 5.60016e-07 [detach_backward]: 4.76002e-06 [pipeline_parallel_scheduler]: 2.15002e-06 [auto_monad_reorder]: 5e-05 [get_jit_bprop_graph]: 2.37999e-06 [rewriter_after_jit_bprop_graph]: 1.301e-05 [opt_after_jit_grad]: 0.00111356 [validate]: 7.275e-05 Sums bootstrap : 0.001053s : 0.52% type_inference : 0.011191s : 5.55% event_method : 0.000017s : 0.01% auto_monad : 0.000089s : 0.04% graph_reusing : 0.000005s : 0.00% inline : 0.000003s : 0.00% add_attr.add_attr_with_inline.tag_attr : 0.000021s : 0.01% add_attr.add_attr_with_inline.meta_addattr_fg_expand : 0.000004s : 0.00% parallel-infer-symbol : 0.000003s : 0.00% pre_auto_parallel : 0.000045s : 0.02% insert-virtual-dataset : 0.000003s : 0.00% parallel-infer-symbol-second : 0.000001s : 0.00% dataset_repeat_opt : 0.000002s : 0.00% pipeline_split : 0.000002s : 0.00% optimize.py_interpret_to_execute : 0.000034s : 0.02% optimize.rewriter_before_opt_a : 0.000069s : 0.03% optimize.opt_a.expand_dump_flag : 0.000005s : 0.00% optimize.opt_a.switch_simplify : 0.000057s : 0.03% optimize.opt_a.loop_unroll : 0.000038s : 0.02% optimize.opt_a.a_1 : 0.001164s : 0.58% optimize.opt_a.with_stream_mark : 0.000050s : 0.02% optimize.opt_a.recompute_prepare : 0.000029s : 0.01% optimize.opt_a.updatestate_depend_eliminate : 0.000012s : 0.01% optimize.opt_a.updatestate_assign_eliminate : 0.000013s : 0.01% optimize.opt_a.updatestate_loads_eliminate : 0.000010s : 0.01% optimize.opt_a.parameter_eliminate : 0.000003s : 0.00% optimize.opt_a.a_2 : 0.000292s : 0.14% optimize.opt_a.accelerated_algorithm : 0.000031s : 0.02% optimize.opt_a.shard : 0.000005s : 0.00% optimize.opt_a.meta_shard_fg_expand : 0.000005s : 0.00% optimize.opt_a.shard_inline : 0.000027s : 0.01% optimize.opt_a.merge_send_recv : 0.000021s : 0.01% optimize.opt_a.auto_parallel : 0.000025s : 0.01% optimize.opt_a.parallel : 0.000040s : 0.02% optimize.opt_a.flash_sp : 0.000014s : 0.01% optimize.opt_a.merge_comm : 0.000011s : 0.01% optimize.opt_a.allreduce_fusion : 0.000012s : 0.01% optimize.opt_a.matmul_add_comm_reduction : 0.000023s : 0.01% optimize.opt_a.allreduce_slice_to_reducescatter : 0.000001s : 0.00% optimize.opt_a.virtual_shard_identity : 0.000028s : 0.01% optimize.opt_a.virtual_dataset : 0.000027s : 0.01% optimize.opt_a.get_grad_eliminate_ : 0.000024s : 0.01% optimize.opt_a.virtual_output : 0.000022s : 0.01% optimize.opt_a.merge_forward : 0.000012s : 0.01% optimize.opt_a.cell_reuse_recompute_pass : 0.000004s : 0.00% optimize.opt_a.offload_activation : 0.000027s : 0.01% optimize.opt_a.cell_reuse_handle_not_recompute_node_pass : 0.000039s : 0.02% optimize.opt_a.merge_recompute_call_nodes : 0.000003s : 0.00% optimize.opt_a.before_grad : 0.000034s : 0.02% optimize.opt_a.set_forward_comm_id_for_comm_node_pass : 0.000011s : 0.01% optimize.opt_a.meta_fg_expand : 0.000008s : 0.00% optimize.opt_a.flash_sp_send_recv_attached : 0.000004s : 0.00% optimize.opt_a.receive_attached : 0.000004s : 0.00% optimize.opt_a.after_resolve : 0.000053s : 0.03% optimize.opt_a.a_after_grad : 0.000038s : 0.02% optimize.opt_a.renormalize : 0.001157s : 0.57% optimize.opt_a.add_forward_monad_depend : 0.000009s : 0.00% optimize.opt_a.auto_monad_grad : 0.000004s : 0.00% optimize.opt_a.auto_monad_eliminator : 0.000038s : 0.02% optimize.opt_a.cse : 0.000149s : 0.07% optimize.opt_a.a_3 : 0.000161s : 0.08% optimize.py_interpret_to_execute_after_opt_a : 0.000038s : 0.02% optimize.slice_cell_reuse_recomputed_activation : 0.000002s : 0.00% optimize.rewriter_after_opt_a : 0.000063s : 0.03% optimize.convert_after_rewriter : 0.000009s : 0.00% optimize.order_py_execute_after_rewriter : 0.000008s : 0.00% optimize.mutable_eliminate : 0.000755s : 0.37% optimize.opt_b.b_1 : 0.000353s : 0.17% optimize.opt_b.b_2 : 0.000014s : 0.01% optimize.opt_b.updatestate_depend_eliminate : 0.000010s : 0.01% optimize.opt_b.updatestate_assign_eliminate : 0.000005s : 0.00% optimize.opt_b.updatestate_loads_eliminate : 0.000004s : 0.00% optimize.opt_b.renormalize : 0.000001s : 0.00% optimize.opt_b.cse : 0.000079s : 0.04% optimize.optimize_parallel_all_gather_comm : 0.000032s : 0.02% optimize.overlap_param_gather : 0.000002s : 0.00% optimize.cconv : 0.000038s : 0.02% optimize.loop_unroll : 0.000772s : 0.38% optimize.opt_after_cconv.c_1 : 0.000068s : 0.03% optimize.opt_after_cconv.parameter_eliminate : 0.000005s : 0.00% optimize.opt_after_cconv.updatestate_depend_eliminate : 0.000008s : 0.00% optimize.opt_after_cconv.updatestate_assign_eliminate : 0.000004s : 0.00% optimize.opt_after_cconv.updatestate_loads_eliminate : 0.000005s : 0.00% optimize.opt_after_cconv.cse : 0.000068s : 0.03% optimize.opt_after_cconv.renormalize : 0.000000s : 0.00% optimize.remove_dup_value : 0.000086s : 0.04% optimize.tuple_transform.d_1 : 0.000101s : 0.05% optimize.tuple_transform.none_parameter_eliminate : 0.000002s : 0.00% optimize.tuple_transform.renormalize : 0.000000s : 0.00% optimize.tuple_transform.switch_simplify : 0.000015s : 0.01% optimize.partial_unused_args_eliminate : 0.000002s : 0.00% optimize.add_recomputation : 0.000083s : 0.04% optimize.cse_after_recomputation.cse : 0.000040s : 0.02% optimize.environ_conv : 0.000009s : 0.00% optimize.swap_dp_allreduce_reducescatter : 0.000010s : 0.01% optimize.bias_add_comm_swap : 0.000004s : 0.00% optimize.label_micro_interleaved_index : 0.000006s : 0.00% optimize.label_fine_grained_interleaved_index : 0.000003s : 0.00% optimize.merge_cast_opt : 0.000001s : 0.00% optimize.slice_recompute_activation : 0.000002s : 0.00% optimize.micro_interleaved_order_control : 0.000003s : 0.00% optimize.assign_add_opt : 0.000002s : 0.00% optimize.ForceFp32Comm : 0.000001s : 0.00% optimize.remove_cast_before_assign_add : 0.000001s : 0.00% optimize.full_micro_interleaved_order_control : 0.000002s : 0.00% optimize.reorder_send_recv_between_fp_bp : 0.000003s : 0.00% optimize.comm_op_add_attrs : 0.000001s : 0.00% optimize.add_comm_op_reuse_tag : 0.000001s : 0.00% optimize.interleave_split_concat_branches : 0.000001s : 0.00% optimize.interleave_parallel_branches : 0.000001s : 0.00% optimize.overlap_opt_shard_in_pipeline : 0.000001s : 0.00% optimize.overlap_opt_shard_grad_in_pipeline : 0.000002s : 0.00% optimize.control_data_broadcast_order : 0.000018s : 0.01% optimize.grouped_pairwise_exchange_alltoall : 0.000001s : 0.00% optimize.offloading_packed_experts : 0.000006s : 0.00% optimize.overlap_recompute_and_grad_model_parallel : 0.000006s : 0.00% optimize.overlap_grad_matmul_and_grad_allreduce : 0.000001s : 0.00% optimize.overlap_recompute_allgather_and_fa_grad : 0.000001s : 0.00% optimize.overlap_recompute_comm : 0.000003s : 0.00% optimize.overlap_grad_ring_attention : 0.000006s : 0.00% optimize.overlap_grad_flash_sp : 0.000031s : 0.02% optimize.begin_end_overlap_inline : 0.000001s : 0.00% optimize.split_matmul_comm_elemetwise : 0.000002s : 0.00% optimize.split_layernorm_comm : 0.000002s : 0.00% optimize.handle_group_info : 0.000002s : 0.00% optimize.symbol_engine_optimizer.build : 0.181291s : 89.84% optimize.symbol_engine_optimizer.elim_shapecalc : 0.000061s : 0.03% optimize.symbol_engine_optimizer.elim_not_effective : 0.000040s : 0.02% optimize.symbol_engine_optimizer.opt_reshape : 0.000014s : 0.01% optimize.symbol_engine_optimizer.fold_const_symbol : 0.000034s : 0.02% optimize.symbol_engine_optimizer.renormalize : 0.000001s : 0.00% detach_backward : 0.000005s : 0.00% pipeline_parallel_scheduler : 0.000002s : 0.00% auto_monad_reorder : 0.000050s : 0.02% get_jit_bprop_graph : 0.000002s : 0.00% rewriter_after_jit_bprop_graph : 0.000013s : 0.01% opt_after_jit_grad : 0.001114s : 0.55% validate : 0.000073s : 0.04% Time group info: ------[substitution.] 0.000196 33 2.69% : 0.000005s : 2: substitution.elim_not_effective 7.87% : 0.000015s : 2: substitution.fold_const_symbol 4.07% : 0.000008s : 8: substitution.graph_param_transform 75.87% : 0.000148s : 1: substitution.inline 2.35% : 0.000005s : 4: substitution.j_node_and_user_rematch 2.89% : 0.000006s : 4: substitution.remove_not_recompute_node 4.26% : 0.000008s : 12: substitution.replace_old_param ------[type_inference.] 0.011074 2 88.39% : 0.009787s : 1: type_inference.infer 11.61% : 0.001286s : 1: type_inference.specialize ------[replace.] 0.000023 1 100.00% : 0.000023s : 1: replace.inline ------[match.] 0.000147 1 100.00% : 0.000147s : 1: match.inline ------[predicate.] 0.000242 1879 0.77% : 0.000002s : 17: predicate.accumulaten_eliminater 1.16% : 0.000003s : 8: predicate.ad_related_special_op_eliminate 0.66% : 0.000002s : 16: predicate.addn_check_dump 0.80% : 0.000002s : 17: predicate.addn_zero_filter 0.69% : 0.000002s : 17: predicate.adjust_all_reduce_mul_add 2.09% : 0.000005s : 33: predicate.arithmetic_simplify 1.14% : 0.000003s : 17: predicate.cast_eliminate 0.84% : 0.000002s : 16: predicate.check_bprop_eliminate 0.73% : 0.000002s : 16: predicate.compare_switch_simplify 0.26% : 0.000001s : 8: predicate.const_output_eliminate 0.71% : 0.000002s : 16: predicate.depend_value_elim 0.87% : 0.000002s : 17: predicate.dict_get_item_const_eliminator 0.95% : 0.000002s : 17: predicate.dict_get_item_eliminator 0.85% : 0.000002s : 17: predicate.dict_set_item_eliminator 1.54% : 0.000004s : 16: predicate.dumpgradient_eliminate 0.51% : 0.000001s : 8: predicate.elim_not_effective 2.58% : 0.000006s : 8: predicate.elim_shapecalc_of_broadcastargs 1.11% : 0.000003s : 25: predicate.environ_add_const_eliminate 1.07% : 0.000003s : 25: predicate.environ_get_add_eliminate 1.06% : 0.000003s : 25: predicate.environ_get_depend_swap 1.81% : 0.000004s : 41: predicate.environ_get_eliminate 1.05% : 0.000003s : 25: predicate.environ_get_set_eliminate 0.87% : 0.000002s : 18: predicate.exchange_switch_depend_value 1.48% : 0.000004s : 18: predicate.float_depend_g_call 0.71% : 0.000002s : 16: predicate.float_environ_get_switch 1.09% : 0.000003s : 24: predicate.float_tuple_getitem_switch 0.29% : 0.000001s : 8: predicate.fold_const_symbol 0.82% : 0.000002s : 16: predicate.get_grad_eliminate 0.38% : 0.000001s : 8: predicate.graph_param_transform 0.65% : 0.000002s : 16: predicate.incorporate_call 0.60% : 0.000001s : 16: predicate.incorporate_call_switch 5.29% : 0.000013s : 83: predicate.inline 0.87% : 0.000002s : 16: predicate.inline_without_move 0.47% : 0.000001s : 16: predicate.j_node_and_user_rematch 1.04% : 0.000003s : 16: predicate.less_batch_normalization 1.75% : 0.000004s : 33: predicate.list_to_tuple_eliminator_ 2.11% : 0.000005s : 50: predicate.load_eliminater 1.24% : 0.000003s : 8: predicate.loop_unroll_after_grad 1.15% : 0.000003s : 25: predicate.loop_unroll_before_grad 1.78% : 0.000004s : 33: predicate.make_slice_get_slice_eliminator 0.76% : 0.000002s : 16: predicate.merge_addn 0.78% : 0.000002s : 16: predicate.micro_step_allgather_replace 0.73% : 0.000002s : 16: predicate.mini_step_allgather_replace 0.73% : 0.000002s : 17: predicate.minmaximum_grad 1.43% : 0.000003s : 8: predicate.mutable_eliminate 0.48% : 0.000001s : 8: predicate.opt_reshape 0.42% : 0.000001s : 8: predicate.parallel_virtual_node 1.12% : 0.000003s : 18: predicate.partial_defer_inline 1.20% : 0.000003s : 25: predicate.partial_eliminate 0.80% : 0.000002s : 17: predicate.print_const_string_wrapper 0.77% : 0.000002s : 16: predicate.reduce_all_const_elim 1.06% : 0.000003s : 17: predicate.reduce_eliminate 2.27% : 0.000005s : 50: predicate.redundant_stop_gradient_eliminater 0.67% : 0.000002s : 16: predicate.remove_not_recompute_node 1.34% : 0.000003s : 33: predicate.replace_applicator 0.65% : 0.000002s : 16: predicate.replace_old_param 0.53% : 0.000001s : 8: predicate.reset_defer_inline 0.83% : 0.000002s : 17: predicate.reshape_eliminate 0.84% : 0.000002s : 16: predicate.row_tensor_add_zeros_like 0.42% : 0.000001s : 8: predicate.row_tensor_eliminate 0.89% : 0.000002s : 16: predicate.same_eliminate 0.57% : 0.000001s : 16: predicate.set_cell_output_no_recompute 1.13% : 0.000003s : 16: predicate.shard_identity_eliminate 0.88% : 0.000002s : 16: predicate.special_op_eliminate 0.76% : 0.000002s : 16: predicate.specialize_transform 1.22% : 0.000003s : 16: predicate.split_environ_get_set_with_tuple_value 0.96% : 0.000002s : 16: predicate.stack_unstack_eliminate 0.54% : 0.000001s : 8: predicate.switch_call_monad_eliminater 0.89% : 0.000002s : 18: predicate.switch_defer_inline 1.56% : 0.000004s : 34: predicate.switch_layer_defer_inline 3.74% : 0.000009s : 67: predicate.switch_simplify 0.78% : 0.000002s : 17: predicate.tile_eliminate 0.76% : 0.000002s : 17: predicate.transpose_eliminate 1.63% : 0.000004s : 33: predicate.tuple_list_convert_item_index_to_positive 1.52% : 0.000004s : 33: predicate.tuple_list_get_item_const_eliminator 1.65% : 0.000004s : 33: predicate.tuple_list_get_item_depend_reorder 2.85% : 0.000007s : 49: predicate.tuple_list_get_item_eliminator 1.55% : 0.000004s : 33: predicate.tuple_list_get_set_item_eliminator 2.54% : 0.000006s : 49: predicate.tuple_list_set_item_eliminator 1.65% : 0.000004s : 33: predicate.tuple_to_list_eliminator_ 2.16% : 0.000005s : 50: predicate.updatestate_pure_node_eliminater 2.97% : 0.000007s : 66: predicate.updatestate_useless_node_eliminater 0.46% : 0.000001s : 8: predicate.value_based_eliminate 0.88% : 0.000002s : 16: predicate.virtual_dataset_eliminate 0.89% : 0.000002s : 16: predicate.virtual_output_eliminate 0.34% : 0.000001s : 8: predicate.virtual_view_grad_eliminate 0.55% : 0.000001s : 8: predicate.zero_like_fill_zero ------[func_graph_cloner_run.] 0.001552 6 54.58% : 0.000847s : 3: func_graph_cloner_run.FuncGraphClonerGraph 45.42% : 0.000705s : 3: func_graph_cloner_run.FuncGraphSpecializer ------[meta_graph.] 0.000000 0 ------[manager.] 0.000000 0 ------[pynative] 0.000000 0 ------[others.] 0.408972 192 0.00% : 0.000004s : 1: ForceFp32Comm 1.45% : 0.005932s : 1: add_attr 1.45% : 0.005914s : 1: add_attr_with_inline 0.00% : 0.000004s : 1: add_comm_op_reuse_tag 0.02% : 0.000088s : 1: add_recomputation 0.00% : 0.000004s : 1: assign_add_opt 0.02% : 0.000097s : 1: auto_monad 0.01% : 0.000054s : 1: auto_monad_reorder 0.00% : 0.000003s : 1: begin_end_overlap_inline 0.00% : 0.000007s : 1: bias_add_comm_swap 0.27% : 0.001095s : 1: bootstrap 0.01% : 0.000043s : 1: cconv 0.00% : 0.000012s : 1: comm_op_add_attrs 0.01% : 0.000022s : 1: control_data_broadcast_order 0.00% : 0.000013s : 1: convert_after_rewriter 0.02% : 0.000062s : 1: cse_after_recomputation 0.00% : 0.000005s : 1: dataset_repeat_opt 0.00% : 0.000008s : 1: detach_backward 0.00% : 0.000012s : 1: environ_conv 0.01% : 0.000024s : 1: event_method 0.00% : 0.000005s : 1: full_micro_interleaved_order_control 0.01% : 0.000026s : 1: get_jit_bprop_graph 0.00% : 0.000009s : 1: graph_reusing 0.00% : 0.000004s : 1: grouped_pairwise_exchange_alltoall 0.00% : 0.000004s : 1: handle_group_info 0.00% : 0.000009s : 1: inline 0.00% : 0.000006s : 1: insert-virtual-dataset 0.00% : 0.000007s : 1: interleave_parallel_branches 0.00% : 0.000004s : 1: interleave_split_concat_branches 0.00% : 0.000009s : 1: label_fine_grained_interleaved_index 0.00% : 0.000009s : 1: label_micro_interleaved_index 0.19% : 0.000786s : 1: loop_unroll 0.00% : 0.000004s : 1: merge_cast_opt 0.00% : 0.000006s : 1: micro_interleaved_order_control 0.19% : 0.000765s : 1: mutable_eliminate 0.00% : 0.000009s : 1: offloading_packed_experts 0.01% : 0.000026s : 1: opt.transform.loop_unroll_optimizer 0.01% : 0.000027s : 1: opt.transform.mutable_eliminate 0.48% : 0.001960s : 78: opt.transform.opt_a 0.02% : 0.000066s : 1: opt.transform.opt_after_cconv 0.01% : 0.000053s : 1: opt.transform.opt_after_jit_grad 0.08% : 0.000308s : 28: opt.transform.opt_b 0.03% : 0.000113s : 2: opt.transform.opt_trans_graph 0.03% : 0.000136s : 4: opt.transform.symbol_engine_opt 1.14% : 0.004674s : 1: opt_a 0.06% : 0.000256s : 1: opt_after_cconv 0.28% : 0.001127s : 1: opt_after_jit_grad 0.15% : 0.000594s : 1: opt_b 46.41% : 0.189808s : 1: optimize 0.01% : 0.000036s : 1: optimize_parallel_all_gather_comm 0.00% : 0.000011s : 1: order_py_execute_after_rewriter 0.01% : 0.000035s : 1: overlap_grad_flash_sp 0.00% : 0.000004s : 1: overlap_grad_matmul_and_grad_allreduce 0.00% : 0.000010s : 1: overlap_grad_ring_attention 0.00% : 0.000005s : 1: overlap_opt_shard_grad_in_pipeline 0.00% : 0.000004s : 1: overlap_opt_shard_in_pipeline 0.00% : 0.000008s : 1: overlap_param_gather 0.00% : 0.000008s : 1: overlap_recompute_allgather_and_fa_grad 0.00% : 0.000010s : 1: overlap_recompute_and_grad_model_parallel 0.00% : 0.000005s : 1: overlap_recompute_comm 0.00% : 0.000007s : 1: parallel-infer-symbol 0.00% : 0.000004s : 1: parallel-infer-symbol-second 0.00% : 0.000006s : 1: partial_unused_args_eliminate 0.00% : 0.000006s : 1: pipeline_parallel_scheduler 0.00% : 0.000005s : 1: pipeline_split 0.01% : 0.000049s : 1: pre_auto_parallel 0.01% : 0.000039s : 1: py_interpret_to_execute 0.01% : 0.000042s : 1: py_interpret_to_execute_after_opt_a 0.00% : 0.000008s : 1: remove_cast_before_assign_add 0.02% : 0.000091s : 1: remove_dup_value 0.13% : 0.000514s : 1: renormalize.infer 0.15% : 0.000631s : 1: renormalize.specialize 0.00% : 0.000006s : 1: reorder_send_recv_between_fp_bp 0.00% : 0.000017s : 1: rewriter_after_jit_bprop_graph 0.02% : 0.000068s : 1: rewriter_after_opt_a 0.02% : 0.000073s : 1: rewriter_before_opt_a 0.00% : 0.000005s : 1: slice_cell_reuse_recomputed_activation 0.00% : 0.000005s : 1: slice_recompute_activation 0.00% : 0.000009s : 1: split_layernorm_comm 0.00% : 0.000006s : 1: split_matmul_comm_elemetwise 0.00% : 0.000014s : 1: swap_dp_allreduce_reducescatter 44.41% : 0.181615s : 1: symbol_engine_optimizer 0.05% : 0.000190s : 1: tuple_transform 2.74% : 0.011216s : 1: type_inference . [hook] pytest_runtest_teardown:test_paged_attention_mla_combine_cache_norm[1-float32-16] tests/st/infer/ops/test_internal_ops/test_paged_attention.py::test_paged_attention_mla_combine_cache_norm[1-float32-16],max_mem:270.0M [WARNING] ME(168852:281473710628656,MainProcess):2026-01-29-17:38:42.233.978 [mindspore/context.py:1334] For 'context.set_context', the parameter 'device_target' will be deprecated and removed in a future version. Please use the api mindspore.set_device() instead. TotalTime = 0.338727, [21] [bootstrap]: 0.00086647 [type_inference]: 0.179305 [event_method]: 1.84e-05 [auto_monad]: 5.975e-05 [graph_reusing]: 5.35999e-06 [inline]: 2.51e-06 [add_attr]: 0.00803916, [1] [add_attr_with_inline]: 0.00801457, [1] [Cycle 1]: 9.56e-05, [2] [tag_attr]: 2.74e-05 [meta_addattr_fg_expand]: 6.23e-06 [parallel-infer-symbol]: 4.07e-06 [pre_auto_parallel]: 4.397e-05 [insert-virtual-dataset]: 2.43e-06 [parallel-infer-symbol-second]: 6.29982e-07 [dataset_repeat_opt]: 2.11003e-06 [pipeline_split]: 1.60001e-06 [optimize]: 0.0140421, [53] [py_interpret_to_execute]: 8.002e-05 [rewriter_before_opt_a]: 0.00013145 [opt_a]: 0.00551164, [2] [Cycle 1]: 0.00415867, [45] [expand_dump_flag]: 3.43999e-06 [switch_simplify]: 5.687e-05 [loop_unroll]: 3.485e-05 [a_1]: 0.00099177 [with_stream_mark]: 3.419e-05 [recompute_prepare]: 2.106e-05 [updatestate_depend_eliminate]: 7.25998e-06 [updatestate_assign_eliminate]: 8.94e-06 [updatestate_loads_eliminate]: 7.82002e-06 [parameter_eliminate]: 2.22999e-06 [a_2]: 0.00018928 [accelerated_algorithm]: 2.753e-05 [shard]: 3.08e-06 [meta_shard_fg_expand]: 4.32e-06 [shard_inline]: 1.836e-05 [merge_send_recv]: 1.461e-05 [auto_parallel]: 1.101e-05 [parallel]: 3.85e-05 [flash_sp]: 1.274e-05 [merge_comm]: 5.81998e-06 [allreduce_fusion]: 5.58002e-06 [matmul_add_comm_reduction]: 1.801e-05 [allreduce_slice_to_reducescatter]: 1.15001e-06 [virtual_shard_identity]: 1.871e-05 [virtual_dataset]: 2.326e-05 [get_grad_eliminate_]: 2.011e-05 [virtual_output]: 1.587e-05 [merge_forward]: 7.38e-06 [cell_reuse_recompute_pass]: 2.06e-06 [offload_activation]: 3.24e-05 [cell_reuse_handle_not_recompute_node_pass]: 2.502e-05 [merge_recompute_call_nodes]: 1.44998e-06 [before_grad]: 2.496e-05 [set_forward_comm_id_for_comm_node_pass]: 6.83998e-06 [meta_fg_expand]: 5.13002e-06 [flash_sp_send_recv_attached]: 3.95998e-06 [receive_attached]: 2.89999e-06 [after_resolve]: 4.781e-05 [a_after_grad]: 4.666e-05 [renormalize]: 0.00159967 [add_forward_monad_depend]: 8.52998e-06 [auto_monad_grad]: 2.61999e-06 [auto_monad_eliminator]: 2.376e-05 [cse]: 9.302e-05 [a_3]: 8.749e-05 [Cycle 2]: 0.00133259, [45] [expand_dump_flag]: 2.71e-06 [switch_simplify]: 1.656e-05 [loop_unroll]: 1.586e-05 [a_1]: 0.00024753 [with_stream_mark]: 4.469e-05 [recompute_prepare]: 1.599e-05 [updatestate_depend_eliminate]: 6.19001e-06 [updatestate_assign_eliminate]: 4.63999e-06 [updatestate_loads_eliminate]: 4.60001e-06 [parameter_eliminate]: 2.11998e-06 [a_2]: 0.00012935 [accelerated_algorithm]: 1.109e-05 [shard]: 2.44999e-06 [meta_shard_fg_expand]: 3.51001e-06 [shard_inline]: 1.353e-05 [merge_send_recv]: 8.33999e-06 [auto_parallel]: 1.647e-05 [parallel]: 8.80999e-06 [flash_sp]: 4.08999e-06 [merge_comm]: 4.97e-06 [allreduce_fusion]: 1.116e-05 [matmul_add_comm_reduction]: 1.082e-05 [allreduce_slice_to_reducescatter]: 6.19999e-07 [virtual_shard_identity]: 1.273e-05 [virtual_dataset]: 1.37e-05 [get_grad_eliminate_]: 1.01e-05 [virtual_output]: 1.702e-05 [merge_forward]: 5.04003e-06 [cell_reuse_recompute_pass]: 2.51e-06 [offload_activation]: 1.188e-05 [cell_reuse_handle_not_recompute_node_pass]: 2.088e-05 [merge_recompute_call_nodes]: 9.80013e-07 [before_grad]: 2.758e-05 [set_forward_comm_id_for_comm_node_pass]: 5.42001e-06 [meta_fg_expand]: 3.95e-06 [flash_sp_send_recv_attached]: 1.49998e-06 [receive_attached]: 2.07001e-06 [after_resolve]: 3.875e-05 [a_after_grad]: 3.138e-05 [renormalize]: 6.99947e-08 [add_forward_monad_depend]: 2.48e-06 [auto_monad_grad]: 1.94e-06 [auto_monad_eliminator]: 1.658e-05 [cse]: 6.442e-05 [a_3]: 0.00011016 [py_interpret_to_execute_after_opt_a]: 3.884e-05 [slice_cell_reuse_recomputed_activation]: 2.32999e-06 [rewriter_after_opt_a]: 8.178e-05 [convert_after_rewriter]: 1.32e-05 [order_py_execute_after_rewriter]: 7.01001e-06 [mutable_eliminate]: 0.00090344 [opt_b]: 0.0005099, [1] [Cycle 1]: 0.00049967, [7] [b_1]: 0.00029841 [b_2]: 1.78e-05 [updatestate_depend_eliminate]: 1.036e-05 [updatestate_assign_eliminate]: 4.82e-06 [updatestate_loads_eliminate]: 4.60001e-06 [renormalize]: 6.89994e-07 [cse]: 9.158e-05 [optimize_parallel_all_gather_comm]: 2.445e-05 [overlap_param_gather]: 4.70999e-06 [cconv]: 3.698e-05 [loop_unroll]: 0.00069189 [opt_after_cconv]: 0.00489946, [1] [Cycle 1]: 0.00488637, [7] [c_1]: 5.993e-05 [parameter_eliminate]: 5.25001e-06 [updatestate_depend_eliminate]: 9.15001e-06 [updatestate_assign_eliminate]: 5.17e-06 [updatestate_loads_eliminate]: 4.53999e-06 [cse]: 0.00470742 [renormalize]: 1.25999e-06 [remove_dup_value]: 8.553e-05 [tuple_transform]: 0.00019196, [1] [Cycle 1]: 0.00018157, [4] [d_1]: 0.00012804 [none_parameter_eliminate]: 4.42e-06 [renormalize]: 1.69995e-07 [switch_simplify]: 1.355e-05 [partial_unused_args_eliminate]: 2.19001e-06 [add_recomputation]: 9.613e-05 [cse_after_recomputation]: 4.482e-05, [1] [Cycle 1]: 3.983e-05, [1] [cse]: 3.243e-05 [environ_conv]: 9.35001e-06 [swap_dp_allreduce_reducescatter]: 8.02e-06 [bias_add_comm_swap]: 4.1e-06 [label_micro_interleaved_index]: 7.79002e-06 [label_fine_grained_interleaved_index]: 3.00002e-06 [merge_cast_opt]: 1.99e-06 [slice_recompute_activation]: 2.22999e-06 [micro_interleaved_order_control]: 2.76e-06 [assign_add_opt]: 1.52001e-06 [ForceFp32Comm]: 1.59998e-06 [remove_cast_before_assign_add]: 1.02e-06 [full_micro_interleaved_order_control]: 2.44001e-06 [reorder_send_recv_between_fp_bp]: 2.77002e-06 [comm_op_add_attrs]: 1.12999e-06 [add_comm_op_reuse_tag]: 1.15999e-06 [interleave_split_concat_branches]: 1.25999e-06 [interleave_parallel_branches]: 1.04e-06 [overlap_opt_shard_in_pipeline]: 1.70001e-06 [overlap_opt_shard_grad_in_pipeline]: 2.61e-06 [control_data_broadcast_order]: 2.19e-05 [grouped_pairwise_exchange_alltoall]: 1.81998e-06 [offloading_packed_experts]: 6.10002e-06 [overlap_recompute_and_grad_model_parallel]: 6.79999e-06 [overlap_grad_matmul_and_grad_allreduce]: 1.68002e-06 [overlap_recompute_allgather_and_fa_grad]: 1.40001e-06 [overlap_recompute_comm]: 2.65002e-06 [overlap_grad_ring_attention]: 9.37999e-06 [overlap_grad_flash_sp]: 2.651e-05 [begin_end_overlap_inline]: 7.7e-07 [split_matmul_comm_elemetwise]: 2.09e-06 [split_layernorm_comm]: 2.34999e-06 [handle_group_info]: 1.04e-06 [symbol_engine_optimizer]: 0.0002395, [1] [Cycle 1]: 0.0002301, [6] [build]: 9.139e-05 [elim_shapecalc]: 3.043e-05 [elim_not_effective]: 2.183e-05 [opt_reshape]: 1.402e-05 [fold_const_symbol]: 2.584e-05 [renormalize]: 2.80008e-07 [detach_backward]: 2.71e-06 [pipeline_parallel_scheduler]: 1.42e-06 [auto_monad_reorder]: 2.848e-05 [get_jit_bprop_graph]: 2.01e-06 [rewriter_after_jit_bprop_graph]: 5.84999e-06 [opt_after_jit_grad]: 0.00098246 [validate]: 0.135092 Sums bootstrap : 0.000866s : 0.26% type_inference : 0.179305s : 54.50% event_method : 0.000018s : 0.01% auto_monad : 0.000060s : 0.02% graph_reusing : 0.000005s : 0.00% inline : 0.000003s : 0.00% add_attr.add_attr_with_inline.tag_attr : 0.000027s : 0.01% add_attr.add_attr_with_inline.meta_addattr_fg_expand : 0.000006s : 0.00% parallel-infer-symbol : 0.000004s : 0.00% pre_auto_parallel : 0.000044s : 0.01% insert-virtual-dataset : 0.000002s : 0.00% parallel-infer-symbol-second : 0.000001s : 0.00% dataset_repeat_opt : 0.000002s : 0.00% pipeline_split : 0.000002s : 0.00% optimize.py_interpret_to_execute : 0.000080s : 0.02% optimize.rewriter_before_opt_a : 0.000131s : 0.04% optimize.opt_a.expand_dump_flag : 0.000006s : 0.00% optimize.opt_a.switch_simplify : 0.000073s : 0.02% optimize.opt_a.loop_unroll : 0.000051s : 0.02% optimize.opt_a.a_1 : 0.001239s : 0.38% optimize.opt_a.with_stream_mark : 0.000079s : 0.02% optimize.opt_a.recompute_prepare : 0.000037s : 0.01% optimize.opt_a.updatestate_depend_eliminate : 0.000013s : 0.00% optimize.opt_a.updatestate_assign_eliminate : 0.000014s : 0.00% optimize.opt_a.updatestate_loads_eliminate : 0.000012s : 0.00% optimize.opt_a.parameter_eliminate : 0.000004s : 0.00% optimize.opt_a.a_2 : 0.000319s : 0.10% optimize.opt_a.accelerated_algorithm : 0.000039s : 0.01% optimize.opt_a.shard : 0.000006s : 0.00% optimize.opt_a.meta_shard_fg_expand : 0.000008s : 0.00% optimize.opt_a.shard_inline : 0.000032s : 0.01% optimize.opt_a.merge_send_recv : 0.000023s : 0.01% optimize.opt_a.auto_parallel : 0.000027s : 0.01% optimize.opt_a.parallel : 0.000047s : 0.01% optimize.opt_a.flash_sp : 0.000017s : 0.01% optimize.opt_a.merge_comm : 0.000011s : 0.00% optimize.opt_a.allreduce_fusion : 0.000017s : 0.01% optimize.opt_a.matmul_add_comm_reduction : 0.000029s : 0.01% optimize.opt_a.allreduce_slice_to_reducescatter : 0.000002s : 0.00% optimize.opt_a.virtual_shard_identity : 0.000031s : 0.01% optimize.opt_a.virtual_dataset : 0.000037s : 0.01% optimize.opt_a.get_grad_eliminate_ : 0.000030s : 0.01% optimize.opt_a.virtual_output : 0.000033s : 0.01% optimize.opt_a.merge_forward : 0.000012s : 0.00% optimize.opt_a.cell_reuse_recompute_pass : 0.000005s : 0.00% optimize.opt_a.offload_activation : 0.000044s : 0.01% optimize.opt_a.cell_reuse_handle_not_recompute_node_pass : 0.000046s : 0.01% optimize.opt_a.merge_recompute_call_nodes : 0.000002s : 0.00% optimize.opt_a.before_grad : 0.000053s : 0.02% optimize.opt_a.set_forward_comm_id_for_comm_node_pass : 0.000012s : 0.00% optimize.opt_a.meta_fg_expand : 0.000009s : 0.00% optimize.opt_a.flash_sp_send_recv_attached : 0.000005s : 0.00% optimize.opt_a.receive_attached : 0.000005s : 0.00% optimize.opt_a.after_resolve : 0.000087s : 0.03% optimize.opt_a.a_after_grad : 0.000078s : 0.02% optimize.opt_a.renormalize : 0.001600s : 0.49% optimize.opt_a.add_forward_monad_depend : 0.000011s : 0.00% optimize.opt_a.auto_monad_grad : 0.000005s : 0.00% optimize.opt_a.auto_monad_eliminator : 0.000040s : 0.01% optimize.opt_a.cse : 0.000157s : 0.05% optimize.opt_a.a_3 : 0.000198s : 0.06% optimize.py_interpret_to_execute_after_opt_a : 0.000039s : 0.01% optimize.slice_cell_reuse_recomputed_activation : 0.000002s : 0.00% optimize.rewriter_after_opt_a : 0.000082s : 0.02% optimize.convert_after_rewriter : 0.000013s : 0.00% optimize.order_py_execute_after_rewriter : 0.000007s : 0.00% optimize.mutable_eliminate : 0.000903s : 0.27% optimize.opt_b.b_1 : 0.000298s : 0.09% optimize.opt_b.b_2 : 0.000018s : 0.01% optimize.opt_b.updatestate_depend_eliminate : 0.000010s : 0.00% optimize.opt_b.updatestate_assign_eliminate : 0.000005s : 0.00% optimize.opt_b.updatestate_loads_eliminate : 0.000005s : 0.00% optimize.opt_b.renormalize : 0.000001s : 0.00% optimize.opt_b.cse : 0.000092s : 0.03% optimize.optimize_parallel_all_gather_comm : 0.000024s : 0.01% optimize.overlap_param_gather : 0.000005s : 0.00% optimize.cconv : 0.000037s : 0.01% optimize.loop_unroll : 0.000692s : 0.21% optimize.opt_after_cconv.c_1 : 0.000060s : 0.02% optimize.opt_after_cconv.parameter_eliminate : 0.000005s : 0.00% optimize.opt_after_cconv.updatestate_depend_eliminate : 0.000009s : 0.00% optimize.opt_after_cconv.updatestate_assign_eliminate : 0.000005s : 0.00% optimize.opt_after_cconv.updatestate_loads_eliminate : 0.000005s : 0.00% optimize.opt_after_cconv.cse : 0.004707s : 1.43% optimize.opt_after_cconv.renormalize : 0.000001s : 0.00% optimize.remove_dup_value : 0.000086s : 0.03% optimize.tuple_transform.d_1 : 0.000128s : 0.04% optimize.tuple_transform.none_parameter_eliminate : 0.000004s : 0.00% optimize.tuple_transform.renormalize : 0.000000s : 0.00% optimize.tuple_transform.switch_simplify : 0.000014s : 0.00% optimize.partial_unused_args_eliminate : 0.000002s : 0.00% optimize.add_recomputation : 0.000096s : 0.03% optimize.cse_after_recomputation.cse : 0.000032s : 0.01% optimize.environ_conv : 0.000009s : 0.00% optimize.swap_dp_allreduce_reducescatter : 0.000008s : 0.00% optimize.bias_add_comm_swap : 0.000004s : 0.00% optimize.label_micro_interleaved_index : 0.000008s : 0.00% optimize.label_fine_grained_interleaved_index : 0.000003s : 0.00% optimize.merge_cast_opt : 0.000002s : 0.00% optimize.slice_recompute_activation : 0.000002s : 0.00% optimize.micro_interleaved_order_control : 0.000003s : 0.00% optimize.assign_add_opt : 0.000002s : 0.00% optimize.ForceFp32Comm : 0.000002s : 0.00% optimize.remove_cast_before_assign_add : 0.000001s : 0.00% optimize.full_micro_interleaved_order_control : 0.000002s : 0.00% optimize.reorder_send_recv_between_fp_bp : 0.000003s : 0.00% optimize.comm_op_add_attrs : 0.000001s : 0.00% optimize.add_comm_op_reuse_tag : 0.000001s : 0.00% optimize.interleave_split_concat_branches : 0.000001s : 0.00% optimize.interleave_parallel_branches : 0.000001s : 0.00% optimize.overlap_opt_shard_in_pipeline : 0.000002s : 0.00% optimize.overlap_opt_shard_grad_in_pipeline : 0.000003s : 0.00% optimize.control_data_broadcast_order : 0.000022s : 0.01% optimize.grouped_pairwise_exchange_alltoall : 0.000002s : 0.00% optimize.offloading_packed_experts : 0.000006s : 0.00% optimize.overlap_recompute_and_grad_model_parallel : 0.000007s : 0.00% optimize.overlap_grad_matmul_and_grad_allreduce : 0.000002s : 0.00% optimize.overlap_recompute_allgather_and_fa_grad : 0.000001s : 0.00% optimize.overlap_recompute_comm : 0.000003s : 0.00% optimize.overlap_grad_ring_attention : 0.000009s : 0.00% optimize.overlap_grad_flash_sp : 0.000027s : 0.01% optimize.begin_end_overlap_inline : 0.000001s : 0.00% optimize.split_matmul_comm_elemetwise : 0.000002s : 0.00% optimize.split_layernorm_comm : 0.000002s : 0.00% optimize.handle_group_info : 0.000001s : 0.00% optimize.symbol_engine_optimizer.build : 0.000091s : 0.03% optimize.symbol_engine_optimizer.elim_shapecalc : 0.000030s : 0.01% optimize.symbol_engine_optimizer.elim_not_effective : 0.000022s : 0.01% optimize.symbol_engine_optimizer.opt_reshape : 0.000014s : 0.00% optimize.symbol_engine_optimizer.fold_const_symbol : 0.000026s : 0.01% optimize.symbol_engine_optimizer.renormalize : 0.000000s : 0.00% detach_backward : 0.000003s : 0.00% pipeline_parallel_scheduler : 0.000001s : 0.00% auto_monad_reorder : 0.000028s : 0.01% get_jit_bprop_graph : 0.000002s : 0.00% rewriter_after_jit_bprop_graph : 0.000006s : 0.00% opt_after_jit_grad : 0.000982s : 0.30% validate : 0.135092s : 41.06% Time group info: ------[substitution.] 0.000203 33 1.75% : 0.000004s : 2: substitution.elim_not_effective 4.08% : 0.000008s : 2: substitution.fold_const_symbol 4.76% : 0.000010s : 8: substitution.graph_param_transform 79.89% : 0.000163s : 1: substitution.inline 2.08% : 0.000004s : 4: substitution.j_node_and_user_rematch 2.62% : 0.000005s : 4: substitution.remove_not_recompute_node 4.81% : 0.000010s : 12: substitution.replace_old_param ------[type_inference.] 0.179183 2 99.34% : 0.178000s : 1: type_inference.infer 0.66% : 0.001183s : 1: type_inference.specialize ------[replace.] 0.000030 1 100.00% : 0.000030s : 1: replace.inline ------[match.] 0.000161 1 100.00% : 0.000161s : 1: match.inline ------[predicate.] 0.000251 1879 0.83% : 0.000002s : 17: predicate.accumulaten_eliminater 1.43% : 0.000004s : 8: predicate.ad_related_special_op_eliminate 0.81% : 0.000002s : 16: predicate.addn_check_dump 0.90% : 0.000002s : 17: predicate.addn_zero_filter 0.75% : 0.000002s : 17: predicate.adjust_all_reduce_mul_add 2.40% : 0.000006s : 33: predicate.arithmetic_simplify 0.89% : 0.000002s : 17: predicate.cast_eliminate 0.85% : 0.000002s : 16: predicate.check_bprop_eliminate 0.73% : 0.000002s : 16: predicate.compare_switch_simplify 0.26% : 0.000001s : 8: predicate.const_output_eliminate 0.79% : 0.000002s : 16: predicate.depend_value_elim 0.94% : 0.000002s : 17: predicate.dict_get_item_const_eliminator 1.10% : 0.000003s : 17: predicate.dict_get_item_eliminator 0.79% : 0.000002s : 17: predicate.dict_set_item_eliminator 1.28% : 0.000003s : 16: predicate.dumpgradient_eliminate 0.31% : 0.000001s : 8: predicate.elim_not_effective 0.59% : 0.000001s : 8: predicate.elim_shapecalc_of_broadcastargs 1.07% : 0.000003s : 25: predicate.environ_add_const_eliminate 1.07% : 0.000003s : 25: predicate.environ_get_add_eliminate 1.05% : 0.000003s : 25: predicate.environ_get_depend_swap 1.78% : 0.000004s : 41: predicate.environ_get_eliminate 1.03% : 0.000003s : 25: predicate.environ_get_set_eliminate 0.80% : 0.000002s : 18: predicate.exchange_switch_depend_value 1.48% : 0.000004s : 18: predicate.float_depend_g_call 0.70% : 0.000002s : 16: predicate.float_environ_get_switch 1.13% : 0.000003s : 24: predicate.float_tuple_getitem_switch 0.26% : 0.000001s : 8: predicate.fold_const_symbol 1.00% : 0.000003s : 16: predicate.get_grad_eliminate 0.38% : 0.000001s : 8: predicate.graph_param_transform 0.76% : 0.000002s : 16: predicate.incorporate_call 0.63% : 0.000002s : 16: predicate.incorporate_call_switch 5.36% : 0.000013s : 83: predicate.inline 1.07% : 0.000003s : 16: predicate.inline_without_move 0.49% : 0.000001s : 16: predicate.j_node_and_user_rematch 1.11% : 0.000003s : 16: predicate.less_batch_normalization 1.96% : 0.000005s : 33: predicate.list_to_tuple_eliminator_ 2.26% : 0.000006s : 50: predicate.load_eliminater 1.03% : 0.000003s : 8: predicate.loop_unroll_after_grad 1.38% : 0.000003s : 25: predicate.loop_unroll_before_grad 1.72% : 0.000004s : 33: predicate.make_slice_get_slice_eliminator 0.74% : 0.000002s : 16: predicate.merge_addn 0.66% : 0.000002s : 16: predicate.micro_step_allgather_replace 0.71% : 0.000002s : 16: predicate.mini_step_allgather_replace 0.73% : 0.000002s : 17: predicate.minmaximum_grad 1.11% : 0.000003s : 8: predicate.mutable_eliminate 0.45% : 0.000001s : 8: predicate.opt_reshape 0.65% : 0.000002s : 8: predicate.parallel_virtual_node 1.22% : 0.000003s : 18: predicate.partial_defer_inline 1.12% : 0.000003s : 25: predicate.partial_eliminate 0.76% : 0.000002s : 17: predicate.print_const_string_wrapper 0.73% : 0.000002s : 16: predicate.reduce_all_const_elim 1.34% : 0.000003s : 17: predicate.reduce_eliminate 2.16% : 0.000005s : 50: predicate.redundant_stop_gradient_eliminater 0.83% : 0.000002s : 16: predicate.remove_not_recompute_node 1.29% : 0.000003s : 33: predicate.replace_applicator 0.63% : 0.000002s : 16: predicate.replace_old_param 0.39% : 0.000001s : 8: predicate.reset_defer_inline 0.97% : 0.000002s : 17: predicate.reshape_eliminate 0.84% : 0.000002s : 16: predicate.row_tensor_add_zeros_like 0.47% : 0.000001s : 8: predicate.row_tensor_eliminate 0.87% : 0.000002s : 16: predicate.same_eliminate 0.62% : 0.000002s : 16: predicate.set_cell_output_no_recompute 1.03% : 0.000003s : 16: predicate.shard_identity_eliminate 0.78% : 0.000002s : 16: predicate.special_op_eliminate 0.78% : 0.000002s : 16: predicate.specialize_transform 1.26% : 0.000003s : 16: predicate.split_environ_get_set_with_tuple_value 0.89% : 0.000002s : 16: predicate.stack_unstack_eliminate 0.37% : 0.000001s : 8: predicate.switch_call_monad_eliminater 0.94% : 0.000002s : 18: predicate.switch_defer_inline 1.58% : 0.000004s : 34: predicate.switch_layer_defer_inline 3.89% : 0.000010s : 67: predicate.switch_simplify 1.01% : 0.000003s : 17: predicate.tile_eliminate 0.94% : 0.000002s : 17: predicate.transpose_eliminate 1.66% : 0.000004s : 33: predicate.tuple_list_convert_item_index_to_positive 1.61% : 0.000004s : 33: predicate.tuple_list_get_item_const_eliminator 1.47% : 0.000004s : 33: predicate.tuple_list_get_item_depend_reorder 3.28% : 0.000008s : 49: predicate.tuple_list_get_item_eliminator 1.69% : 0.000004s : 33: predicate.tuple_list_get_set_item_eliminator 2.44% : 0.000006s : 49: predicate.tuple_list_set_item_eliminator 1.65% : 0.000004s : 33: predicate.tuple_to_list_eliminator_ 2.10% : 0.000005s : 50: predicate.updatestate_pure_node_eliminater 2.91% : 0.000007s : 66: predicate.updatestate_useless_node_eliminater 0.44% : 0.000001s : 8: predicate.value_based_eliminate 1.05% : 0.000003s : 16: predicate.virtual_dataset_eliminate 0.84% : 0.000002s : 16: predicate.virtual_output_eliminate 0.36% : 0.000001s : 8: predicate.virtual_view_grad_eliminate 0.57% : 0.000001s : 8: predicate.zero_like_fill_zero ------[func_graph_cloner_run.] 0.001546 6 45.14% : 0.000698s : 3: func_graph_cloner_run.FuncGraphClonerGraph 54.86% : 0.000848s : 3: func_graph_cloner_run.FuncGraphSpecializer ------[meta_graph.] 0.000000 0 ------[manager.] 0.000000 0 ------[pynative] 0.000000 0 ------[others.] 0.230071 192 0.00% : 0.000004s : 1: ForceFp32Comm 3.50% : 0.008047s : 1: add_attr 3.49% : 0.008024s : 1: add_attr_with_inline 0.00% : 0.000004s : 1: add_comm_op_reuse_tag 0.04% : 0.000100s : 1: add_recomputation 0.00% : 0.000004s : 1: assign_add_opt 0.03% : 0.000065s : 1: auto_monad 0.01% : 0.000033s : 1: auto_monad_reorder 0.00% : 0.000003s : 1: begin_end_overlap_inline 0.00% : 0.000007s : 1: bias_add_comm_swap 0.39% : 0.000905s : 1: bootstrap 0.02% : 0.000041s : 1: cconv 0.00% : 0.000005s : 1: comm_op_add_attrs 0.01% : 0.000025s : 1: control_data_broadcast_order 0.01% : 0.000017s : 1: convert_after_rewriter 0.02% : 0.000048s : 1: cse_after_recomputation 0.00% : 0.000005s : 1: dataset_repeat_opt 0.00% : 0.000006s : 1: detach_backward 0.01% : 0.000012s : 1: environ_conv 0.01% : 0.000025s : 1: event_method 0.00% : 0.000005s : 1: full_micro_interleaved_order_control 0.00% : 0.000006s : 1: get_jit_bprop_graph 0.00% : 0.000009s : 1: graph_reusing 0.00% : 0.000005s : 1: grouped_pairwise_exchange_alltoall 0.00% : 0.000004s : 1: handle_group_info 0.00% : 0.000006s : 1: inline 0.00% : 0.000006s : 1: insert-virtual-dataset 0.00% : 0.000004s : 1: interleave_parallel_branches 0.00% : 0.000004s : 1: interleave_split_concat_branches 0.00% : 0.000006s : 1: label_fine_grained_interleaved_index 0.00% : 0.000011s : 1: label_micro_interleaved_index 0.31% : 0.000703s : 1: loop_unroll 0.00% : 0.000005s : 1: merge_cast_opt 0.00% : 0.000006s : 1: micro_interleaved_order_control 0.40% : 0.000916s : 1: mutable_eliminate 0.00% : 0.000009s : 1: offloading_packed_experts 0.01% : 0.000026s : 1: opt.transform.loop_unroll_optimizer 0.01% : 0.000026s : 1: opt.transform.mutable_eliminate 0.98% : 0.002266s : 78: opt.transform.opt_a 0.03% : 0.000058s : 1: opt.transform.opt_after_cconv 0.03% : 0.000059s : 1: opt.transform.opt_after_jit_grad 0.12% : 0.000272s : 28: opt.transform.opt_b 0.06% : 0.000136s : 2: opt.transform.opt_trans_graph 0.04% : 0.000087s : 4: opt.transform.symbol_engine_opt 2.40% : 0.005519s : 1: opt_a 2.13% : 0.004907s : 1: opt_after_cconv 0.43% : 0.000995s : 1: opt_after_jit_grad 0.22% : 0.000517s : 1: opt_b 6.11% : 0.014049s : 1: optimize 0.01% : 0.000029s : 1: optimize_parallel_all_gather_comm 0.00% : 0.000010s : 1: order_py_execute_after_rewriter 0.01% : 0.000030s : 1: overlap_grad_flash_sp 0.00% : 0.000005s : 1: overlap_grad_matmul_and_grad_allreduce 0.01% : 0.000013s : 1: overlap_grad_ring_attention 0.00% : 0.000006s : 1: overlap_opt_shard_grad_in_pipeline 0.00% : 0.000005s : 1: overlap_opt_shard_in_pipeline 0.00% : 0.000008s : 1: overlap_param_gather 0.00% : 0.000010s : 1: overlap_recompute_allgather_and_fa_grad 0.01% : 0.000013s : 1: overlap_recompute_and_grad_model_parallel 0.00% : 0.000005s : 1: overlap_recompute_comm 0.00% : 0.000008s : 1: parallel-infer-symbol 0.00% : 0.000004s : 1: parallel-infer-symbol-second 0.00% : 0.000009s : 1: partial_unused_args_eliminate 0.00% : 0.000005s : 1: pipeline_parallel_scheduler 0.00% : 0.000005s : 1: pipeline_split 0.02% : 0.000049s : 1: pre_auto_parallel 0.04% : 0.000085s : 1: py_interpret_to_execute 0.02% : 0.000044s : 1: py_interpret_to_execute_after_opt_a 0.00% : 0.000004s : 1: remove_cast_before_assign_add 0.04% : 0.000091s : 1: remove_dup_value 0.27% : 0.000610s : 1: renormalize.infer 0.42% : 0.000971s : 1: renormalize.specialize 0.00% : 0.000006s : 1: reorder_send_recv_between_fp_bp 0.00% : 0.000009s : 1: rewriter_after_jit_bprop_graph 0.04% : 0.000092s : 1: rewriter_after_opt_a 0.06% : 0.000141s : 1: rewriter_before_opt_a 0.00% : 0.000006s : 1: slice_cell_reuse_recomputed_activation 0.00% : 0.000005s : 1: slice_recompute_activation 0.00% : 0.000006s : 1: split_layernorm_comm 0.00% : 0.000005s : 1: split_matmul_comm_elemetwise 0.00% : 0.000011s : 1: swap_dp_allreduce_reducescatter 0.11% : 0.000243s : 1: symbol_engine_optimizer 0.08% : 0.000195s : 1: tuple_transform 77.95% : 0.179330s : 1: type_inference . [hook] pytest_runtest_teardown:test_paged_attention_mla_combine_cache_norm[1-float32-128] tests/st/infer/ops/test_internal_ops/test_paged_attention.py::test_paged_attention_mla_combine_cache_norm[1-float32-128],max_mem:270.0M =============================== warnings summary =============================== ../../../../../../../../../../usr/local/Ascend/cann-8.5.0/python/site-packages/tbe/dsl/classifier/transdata/transdata_classifier.py:222 /usr/local/Ascend/cann-8.5.0/python/site-packages/tbe/dsl/classifier/transdata/transdata_classifier.py:222: DeprecationWarning: invalid escape sequence \B """ ../../../../../../../../../../usr/local/Ascend/cann-8.5.0/python/site-packages/tbe/dsl/unify_schedule/vector/transdata/common/graph/transdata_graph_info.py:143 /usr/local/Ascend/cann-8.5.0/python/site-packages/tbe/dsl/unify_schedule/vector/transdata/common/graph/transdata_graph_info.py:143: DeprecationWarning: invalid escape sequence \c """ ../../../../../../../../../../usr/local/Ascend/cann-8.5.0/python/site-packages/tbe/dsl/unify_schedule/vector/transdata/common/graph/transdata_graph_info.py:170 /usr/local/Ascend/cann-8.5.0/python/site-packages/tbe/dsl/unify_schedule/vector/transdata/common/graph/transdata_graph_info.py:170: DeprecationWarning: invalid escape sequence \c """ ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/numpy/core/getlimits.py:549 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/numpy/core/getlimits.py:549: UserWarning: The value of the smallest subnormal for type is zero. setattr(self, word, getattr(machar, word).flat[0]) ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/numpy/core/getlimits.py:89 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/numpy/core/getlimits.py:89: UserWarning: The value of the smallest subnormal for type is zero. return self._float_to_str(self.smallest_subnormal) ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/numpy/core/getlimits.py:549 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/numpy/core/getlimits.py:549: UserWarning: The value of the smallest subnormal for type is zero. setattr(self, word, getattr(machar, word).flat[0]) ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/numpy/core/getlimits.py:89 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/numpy/core/getlimits.py:89: UserWarning: The value of the smallest subnormal for type is zero. return self._float_to_str(self.smallest_subnormal) ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/batchnorm_fold2.py:57 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/batchnorm_fold2.py:57: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("batchnorm_fold2") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/batchnorm_fold2_grad.py:56 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/batchnorm_fold2_grad.py:56: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("batchnorm_fold2_grad") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/batchnorm_fold2_grad_reduce.py:48 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/batchnorm_fold2_grad_reduce.py:48: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("batchnorm_fold2_grad_reduce") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/correction_mul.py:51 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/correction_mul.py:51: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("correction_mul") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/correction_mul_grad.py:51 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/correction_mul_grad.py:51: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("correction_mul_grad") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/correction_mul_grad.py:143 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/correction_mul_grad.py:143: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("correction_mul_grad_reduce") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perlayer.py:50 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perlayer.py:50: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("fake_learned_scale_quant_perlayer") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perlayer_grad.py:92 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perlayer_grad.py:92: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("fake_learned_scale_quant_perlayer_grad_d") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perlayer_grad_reduce.py:49 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perlayer_grad_reduce.py:49: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("fake_learned_scale_quant_perlayer_grad_d_reduce") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perchannel.py:50 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perchannel.py:50: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("fake_learned_scale_quant_perchannel") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perchannel_grad.py:91 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perchannel_grad.py:91: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("fake_learned_scale_quant_perchannel_grad_d") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perchannel_grad_reduce.py:48 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_learned_scale_quant_perchannel_grad_reduce.py:48: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("fake_learned_scale_quant_perchannel_grad_d_reduce") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_quant_perchannel.py:52 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_quant_perchannel.py:52: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("fake_quant_perchannel") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_quant_perchannel_grad.py:81 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_quant_perchannel_grad.py:81: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("fake_quant_perchannel_grad") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_quant_perlayer.py:54 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_quant_perlayer.py:54: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("fake_quant_per_layer") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_quant_perlayer_grad.py:81 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/fake_quant_perlayer_grad.py:81: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("fake_quant_per_layer_grad") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/minmax_update_perchannel.py:50 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/minmax_update_perchannel.py:50: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("minmax_update_perchannel") ../../../../../../../../anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/minmax_update_perlayer.py:50 /home/jenkins/anaconda3/envs/ci39/lib/python3.9/site-packages/mindspore/ops/_op_impl/_custom_op/minmax_update_perlayer.py:50: DeprecationWarning: te_fusion.fusion_manager.fusion_manager.register is deprecated,please replace it with tbe.common.register.register_op_compute @fusion_manager.register("minmax_update_perlayer") -- Docs: https://docs.pytest.org/en/stable/warnings.html ================== 8 passed, 25 warnings in 121.62s (0:02:01) ==================