\;jLddlZddlZddlZddlmZddlmZmZmZeres ddlm Z dZ ndZ dZ gdZ da Gdd Z dd Zd Zd Zd ZdZdZdZy)N)core) CUDAPlaceis_compiled_with_cudais_compiled_with_rocm) CUDAGraphcy)NTr bG:\00. PROJECTS\API\Inventory\templateJSON\kerjaOCR\Lib\site-packages\paddle/device/cuda/graphs.pyis_cuda_graph_supportedr sr cy)NFr r r r r r #sr )global thread_localrelaxedc4eZdZddZdZdZdZdZd dZy) rNctJdd|_|4ttjj dd}t |}||_|tvsJtj||_ y)NzFCUDA Graph is only supported on PaddlePaddle compiled with NVIDIA GPU.FLAGS_selected_gpusr) CoreCUDAGraph_graphintosenvirongetr_place ALL_MODESindex_mode)selfplacemode device_ids r __init__zCUDAGraph.__init__,sq  % T S T % =BJJNN+@!DEIi(E y   __T* r cXtj|j|jyN)r begin_capturerrrs r capture_beginzCUDAGraph.capture_begin9s##DKKz!wrap_cuda_graph..es(D";F";r ) to_static)Layer)rpaddlein_dynamic_modepaddle.base.frameworkr?r3 cuda_graph_idrgen_new_memory_pool_id ValueError paddle.jitrI paddle.nnrJr2forward_cuda_graph_capture_mode_cuda_graph_pool_id) rGr memory_poolr?graph_idmemory_pool_idrIrJ new_function mock_funcs ` r wrap_cuda_graphr[QsQ 9    ! ! #;}% ) #N E !*AACN^   3J^, ,s 2X = ;= =%X&L(E" ((  )-I&i() %   (5(L(L(N %  k5 )##77  ) -8,K,KI ) r c~|j|j|j|j|j |j |j |j|j|j|j|j|j|jy)z copy var desc from src to dst :param dst: framework.VarDesc(cpp), dst var desc, cpp VarDesc instance :param src: framework.VarDesc(cpp), src var desc, cpp VarDesc instance :return: no return N) set_shapeshape set_dtypedtype set_lod_level lod_levelset_typetypeset_persistable persistableset_is_parameter is_parameterset_stop_gradient stop_gradient)dstsrcs r copy_var_descrm~sMM#))MM#))cmm$LL())*#++,r cg}t|jD]-\}}||kr |jD]}|j|/t t |S)a1 find all inputs of ops after an idx, used to determine the logical output of a cuda graph section :param block: framework.Block, the original block :param begin_idx: int, from which idx (not include) to find the later ins :return: a list of inputs names for all ops behind begin_idx ) enumerateopsinput_arg_namesappendlistset)block begin_idxinsidxopin_names r all_inputs_of_later_opr{sW CUYY'R )  ))G JJw *( C>r ctjj}|j}|j}g}g}tj j }t||d} |D]} | jD]} |j| } |jj| jjd} t| | |j| dk(r&|j| dk(r|j| | j| dk(s|j| dkDs|j!| | j"D]y}|j|} |jj| jjd} t| | |j|dk(si|j|{|jj%}|j'| j|j)|| j+||j-|||gfS)a 1. Construct a new program for corresponding section 2. Find all the logical inputs and outputs of a program section :param section: list, one cuda graph section, list of ops :param origin_program: framework.Program, origin program :param section_idx: list, the section ops' idx corresponding to the cuda graph section, a list of idx :return: a new program for the cuda graph section the logical ins and outs of the cuda graph section asciir)rKstaticProgram global_blockrop_proto_and_checker_makerkOpRoleAttrNamer{rqvardescr0encodermcountrrremoveoutput_arg_names append_op copy_from _set_attrattr_sync_with_cpp)sectionorigin_program section_idxprogramru origin_blockrwoutsop_role_attr_name later_insryrzr new_var_descout_name new_op_descs r #construct_program_and_find_ins_outsrsmm##%G  "E!..0L C D77GGI&|[_EI))G""7+C ::>>#((//'*BCL , ,zz'"a'CIIg,>!,C 7#)Q.4::g3F3J G$*++H""8,C ::>>#((//'*BCL , ,zz(#q(  H%,jj**, bgg&/9J1KL58  S$K r cT|j}g}g}d}g}g}g}g}d} tjj} t tjj j } t tjj j} | | z} t|jD]\}}|jdk(s|jdk(r|jJd|jdr|jdrd}|jt|jtsJd|jj!d }t#|d k(sJd t |d }|| k(rUt#|d kDrt#|t#|k(sJd|D]}t |j| | k(xst |j| | k(}|jdk(xs|jdk(}|s|sdg}g}t#|t#|k(sJdt#|d kDr"|j%||j%|g}g}nt't#|D]*}|j%|||j%||,g}g}|j%||j%|#g}g}|} t#|t#|k(sJdt#|d kDr"|j%||j%||g}|g}|j%||j%|t#|t#|k(sJdt#|d kDr"|j%||j%||||fS)a get all sections that should run under cuda graph and the corresponding idx :param program: framework.Program, the original program :return: A list of cuda graph sections and the corresponding ops' idx in the block. The program is under is test or not. Fr}conditional_blockwhilez9Cuda graph not support conditional block op and while op.is_testTzcuda_graph_attr should be a strrCzccuda graph attr should have three fields: cuda graph mode, cuda graph memory pool id, cuda graph idrz@len of internal section should be equal with len of internal idxz-num of section's op is not equal with the idxz0num of section's op is not equal with num of idx)rrrrrOpRoleLossBackwardrorprd_cuda_graph_attrhas_attrrr2r3splitlenrrrange)rrucuda_graph_sections sections_idxrinternal_section internal_idxcurrent_section current_idxcurrent_cuda_graph_idr loss_op_rolebackward_op_roleloss_grad_op_rolerxrycuda_graph_attrslocal_cuda_graph_id internal_op loss_relatedsub_block_relatedis r get_cuda_graph_sectionsrs  "ELGLOK77GGIt66==BBCL4::AAJJK$'77UYY'R 77) )RWW-?##+ KJ K+ ;;y !bggi&8G   *##S 10 1 "2288= '(A- L -#&&6q&9": "&;;'(1,/0C$5ZYZ(8  0 01B CD+,(!(--.?@01%GG'::Pbgg>P*(+<02,+-L#&#73 +<$ONO$ #?3a7 3 : :? K , 3 3K @.0O*,K!C(8F#3'7#89'../?/BC#**<?;:$& ! &&r*""3'$& ! '&?+s0FEF'!+'..? '' 4#%$"e   # #B '    $k(p  3$ :9:  ?a""?3K(  g 55r c\|d}|d}|d} |j} t|D]} | j| dd} d} |D]>}|j|jj d}|d} t |d} n| | Jd| j dt|ztjjjd d }| j d t|ztjjjd d }tjj||}| j!| d d |i|||d|jdt#|jj$||| | d|j|jd y)a Use section_program and ins_and_outs to initialize a run_program_op, and replace the section_idx marks ops in the origin program. :param ins_and_outs: list, the logical ins and outs of the section program :param section_program: framework.Program, the partial program need to run under cuda graph :param section_idx: list, the idx need to be removed from origin program :param origin_program: framework.Program, the origin program :param cuda_graph_section: list, the ops in current sections, used to get the mode, memory pool id and is_test :param order: int, the order of current section, used to create unique cuda graph var :param is_test: bool, the program is running under is_test or not :return: no return rr@F)syncNrCz>mode and memory pool id should be specified in cuda graph attr cuda_graph_T)r0rdrfrjprogram_out_scope_ run_programX)OutOutScoper) rstart_op_index end_op_indexr program_idcuda_graph_capture_modecuda_graph_pool_iduse_interpretorcoreforward_global_blockbackward_global_block)rdinputsoutputsattrs)rreversed _remove_oprrr create_varr3rVarDescVarTypeRAW STEP_SCOPESrKutils _hash_with_id _insert_oprrp) ins_and_outssection_programrrcuda_graph_sectionorderrrwr insert_idxrrxr rXryrcuda_graph_var out_scope_varrs r replace_cuda_graph_sectionrPs, q/C ?DQJ!..0L $%0% DN    *''--c2E8D q]N  ! ^7HGH7",, SZ ' \\ ! ! % % -N!++ !CJ . \\ ! ! - - ,M++O\JJ Sz%' ,88: < < > B BC$'+"0#($3$@$@$B%4%A%A%C  r c  t|jdkDrtjdt |\}}}t|t|k(sJdg}g}t t|D]:}t |||||\}}|j||j|<t|t|k(sJdtt t|D] }t|||||||||||"|S)z replace the ops marked with cuda_graph_attr to run_program_op to use cuda graph :param program: framework.Program, the program to be transformed :return: the cuda graph section program, user should hold these programs! r@zSub block(s) has been detected in the program. Cuda graph not support op with sub block, and it will only handle the global block.z@num of cuda graph sections is not equal with num of idx sectionszJthe num of cuda graph sections should be equal with the num of new program)rr) rblockswarningswarnrrrrrrr) rrrrrsection_programsrrins_outss r cuda_graph_transformrsM 7>>Q  b 2I2.w " #s( JIJ L 3*+ ,$G  "G\!_% ! H%0 -  C% TST  eC 345 6" O Q  O   " 7 r )rrA)rrrK paddle.baserpaddle.base.corerrrrrr rrNr[rmr{rrrrr r r rs   #8#:;M 2  #7#7L*Z-""1 hy6xVr8r