+ &jR.^RIt^RIt^RIHt^RIt^RIHtRt^RI H t RRlt R R ltR R ltR RltRRltRRltRRltRRltRRltR#RRllt!RR4t!RR4tR$RR lltR%R!R"llt R# ] dRtRtLwi;i)&NAny)runtimeTF)_get_device_indexc8V^8dQhR\P/#returnctypesCDLL)formats"i/Users/jameslopez/projects/CWCArchive/cwc-podcast/.venv/lib/python3.14/site-packages/torch/cuda/_utils.py __annotate__rs&++c^RIp\P!\VP R4^,44pVPVn VPVnVP VnVP$VnVP(VnV# \ \ 3dj\PR8Xd<\P!R\PP^, R24pL\P!R4pLi;i)rNamdhip64win32 amdhip64_.dllzlibamdhip64.so)rocm_sdkr r strfind_libraries ImportError IndexErrorsysplatformtorchversionhiphipGetErrorStringcuGetErrorStringhipModuleLoadDatacuModuleLoadDatahipModuleGetFunctioncuModuleGetFunctionhipModuleLaunchKernelcuLaunchKernelhipFuncSetAttributecuFuncSetAttribute)rlibs r_get_hip_runtime_libraryr,s 0kk#h55jA!DEF00C00C!66C22C 44C J  $0 <<7 "++ %--*;*;A*>)?tDEC++./C 0s9BAD 3D  D c8V^8dQhR\P/#rr )rs"rrr-s++6;;+rc\PR8Xd\P!R4#\P!R4#)rz nvcuda.dllz libcuda.so.1)rrr r rr_get_cuda_libraryr0-s, ||w{{<(({{>**rc8V^8dQhR\P/#rr )rs"rrr5s##&++#rcj\PP'd \4#\ 4#N)rrr r,r0r/rr_get_gpu_runtime_libraryr45s$ }}')) ""rc(V^8dQhR\RR/#r resultr Nint)rs"rrr=s 7 7 7 7rcV^8XdR#\P!4p\4pVPV\P!V44VP eVP P 4MRp\RV 24h)rNUnknown CUDA error CUDA error: )r c_char_pr4r"byrefvaluedecode RuntimeError)r7err_strlibcuda error_messages& r _check_cudarE=sn {ooG&(G VV\\'%:;")--"; AU m_5 66rc0V^8dQhR\R\/#)r r7r r)rs"rrrIsrcn\'g \R4hVvrV\PP8wdQ\P !V4wr4\ V\4'dVP4p\RV RV R24h\V4^8XdR#\V4^8Xd V^,#V#)aCheck a cuda.bindings (cuda-python) call result for errors. All cuda.bindings runtime calls return ``(error, *outputs)``. This helper unpacks the tuple, raises on non-success, and returns the outputs (``None`` for zero outputs, scalar for one, tuple otherwise). zcuda.bindings is not availabler<z ()N) _HAS_CUDA_BINDINGSrA_cuda_bindings_runtime cudaError_t cudaSuccesscudaGetErrorString isinstancebytesr@len)r7errout_rBs& r_check_cuda_bindingsrTIs  ;<<IC ! - - 9 9 : # 5 5   gu % %nn&G\#b ;<< 3x1} 3x1}1v Jrc8V^8dQhR\P/#rr )rs"rrrfsV[[rc4^RIp\P!\VP R4^,44pVPVnVPVnVP"VnVP&VnVP*VnVP.VnVP2VnVP6VnVP:VnVP>Vn V# \ \ 3d\PR8XdqRPR\PP^,R\PP^,.4p\P!RT R24pELB\P!R4pEL[i;i)rNhiprtcr0rz libhiprtc.so)!rr r rrrrrrjoinrrr hiprtcGetErrorStringnvrtcGetErrorStringhiprtcCreateProgramnvrtcCreateProgramhiprtcDestroyProgramnvrtcDestroyProgramhiprtcCompileProgramnvrtcCompileProgramhiprtcGetCodeSizenvrtcGetCUBINSize hiprtcGetCode nvrtcGetCUBINhiprtcGetProgramLogSizenvrtcGetProgramLogSizehiprtcGetProgramLognvrtcGetProgramLoghiprtcAddNameExpressionnvrtcAddNameExpressionhiprtcGetLoweredNamenvrtcGetLoweredName)rr+ version_strs r_get_hiprtc_libraryrpfsA .kk#h55h?BCD"66C 44C!66C!66C11C))C!$!.__annotate__s ? ?C ?D ?rc<VS8wdt\P!4pSPV\P!V44VPeVPP 4MRp\ RV 24hR#)Nr;r<)r r=r\r>r?r@rA)r7rBrD NVRTC_SUCCESSlibnvrtcs& r check_nvrtc#_nvrtc_compile..check_nvrtcso ] "oo'G  ( (g1F G==, $$&)  m_=> > #rutf-8z--offload-arch=z--gpu-architecture=sm_) include_pathsruz-Iz12.8zPCH requires CUDA 12.8+, got z--pchz.cuzKernel compilation failed: rX)& torch.cudar~encoderuget_device_propertiescurrent_devicerr gcnArchNamemajorminorappendrrrAssertionErrorrrrPr r=c_void_pr^r>rlrbc_size_trhcreate_string_bufferr?rjrAr@rdrfrnr`raw)rrrrrrrr source_bytespropsoptionsrcuda_include_paths cuda_path directoryoptionnvrtc_compatible_flagsr num_options options_arrayprog c_kernel_namereslog_sizelog binary_sizebinaryc_mangled_name mangled_namerrs&&&&&& @@r_nvrtc_compilers0$%HM ? ?!''0L! 001J1J1LM ==   $)$5$5#6 $)KK= !> G }});(<=DDFG/0B/CDKKMN8&v.' I;'..01(*I NNR {+224 5+ u}}!! "V + #@ASAS@T!UV V  LG$"F NN6==1 2#;< NN5KL5KTDKK(5KLMg,K__{2W=M ?? D## LL  m3  & & (      &&w/M//mDE  & &t[- HC m??$''fll8.DE))(..9##D#.9#)):J:J:L9MNOO//#K**4k1JKL  ( ():): ;F&&tV45__&N$$T=&,,~:VW'%++224     d!34 ::| ##oMs'R3cDa]tRtRtoV3RlRltV3RlRltRtVtR#) _CudaModuleiIc8<V^8dQhRS[PRR/#)r moduler Nr r)r __classdict__s"rr_CudaModule.__annotate__Js33v343rc Wn/VnR#r3)_module_kernels)selfrs&&r__init___CudaModule.__init__Js 02 rc$<V^8dQhRS[RR/#)r namer _CudaKernel)r)rrs"rrrNsVVV Vrc WP9dVPV,#^RIHpV!4p\P!4p\ VP \P!V4VPVPR444\W@P4pWPPV&V# \dp\RT R24ThRp?ii;i)r)r4rzNo kernel named 'z' in this moduleN) rtorch.cuda._utilsr4r rrEr&r>rrrrAAttributeError)rrr4rCfunckernelrQs&& r __getattr___CudaModule.__getattr__Ns == ==& & ?*,  V ++LL& dkk'6J  !||4F"(MM$ M V #4TF:J!KLRU U VsA-B55 CCC)rrN)__name__ __module__ __qualname____firstlineno__rr__static_attributes____classdictcell__rs@rrrIs33VVrrc^a]tRtRtoRtV3RlRltR V3RlRlltV3RlR ltR tVt R#) rigzL Represents a compiled CUDA kernel that can be called with PyTorch tensors. cR<V^8dQhRS[PRS[PRR/#)r rrr Nr)rrs"rr_CudaKernel.__annotate__ls*''V__'foo'$'rc,WnW n^VnR#)rN)rr_max_shared_mem_bytes)rrrs&&&rr_CudaKernel.__init__ls  %&"rNc <V^8dQhRS[S[S[S[3,RS[S[S[S[3,RS[R,RS[RS[R,RR/#)r gridblockargsN shared_memstreamr )rr9rr)rrs"rrrqsm_ _ CcM"_ S#s]#_ Tk _  _ d _  _ rc ^RIpVPPP4pV'g.p.p.p VEF|p \ WP 4'dV P 'g4V P'dV P4'g \R4h\P!V P44p VPV 4V P\P!V 44K\ V \4'd?\P !V 4p V P\P!V 44EK\ V \"4'd?\P$!V 4p V P\P!V 44EKh\'R\)V 4 24h \P\+V 4,!4p\-V 4F,wr\P.!V \P4W&K. Vf^RIpVPP34pVR 8dWVP4^8XgW@P48d6VP4^8XdRMRVP4 R2p\7RV RV R 24h\9VP;VP<V^,V^,V^,V^,V^,V^,VVP>VR4 4R#) a Call the compiled CUDA kernel Args: grid (tuple): Grid dimensions (grid_x, grid_y, grid_z) block (tuple): Block dimensions (block_x, block_y, block_z) args (list): List of arguments to pass to the kernel. PyTorch tensor arguments will be automatically converted to pointers. shared_mem (int): Shared memory size in bytes stream (torch.cuda.Stream): CUDA stream to use. If None, uses current stream. Nz?All tensor arguments must be CUDA tensors or pinned CPU tensorszUnsupported argument type: znot configuredzonly z bytes configuredzKernel requires z' bytes of shared memory (>= 48KB), but ze. Call kernel.set_shared_memory_config(shared_mem) after compilation and before launching the kernel.) rru_utilsr4rNTensoris_cudais_cpu is_pinned ValueErrorr rdata_ptrrr>r9c_intfloatc_double TypeErrortyperP enumeratecastrcurrent_streamrrArEr(r_as_parameter_)rrrrrrrrCprocessed_argsc_argsargptrrr c_args_arrayiconfigured_msgs&&&&&& r__call___CudaKernel.__call__qsX& **##<<>D13C#||,,{{{CJJJ3==??$Yooclln5%%c* fll3/0C%% S) fll512C''!??3/ fll845"=d3i[ IJJ+0#f+58 'FA$kk#v?LO( > ZZ..0F  "  & &! +z!!/ 01GG 783  & & ;   &6"r)rrr)rrrNrN) rrrr__doc__rrr rrrs@rrrgs+'' _ _ B(6(6rrc V^8dQhR\\,R\\,R,R\\\R3,,/#)r ptx kernel_namesNr r)rrOrrdict)rs"rrrs@-- u-$(I$4-4]*++-rc t^RIp\4p\V\4'dVP R4p\ P !4pVPP4pV;_uu_4\VP\ P!V4V44RRR4V'g \V4#/pVFcp\ P !4p\VP\ P!V4WGP R444\W4Wg&Ke V# +'giL;i)a Loads a CUDA module from PTX code and returns a module object that can access kernels. Args: ptx (bytes or str): The PTX code to load kernel_names (list, optional): List of kernel names to extract from the module. If None, will return a module object with __getattr__. Returns: object: If kernel_names is None, returns a module object with __getattr__ to access kernels. If kernel_names is provided, returns a dict mapping kernel names to _CudaKernel objects. Nr)rr4rNrrr rrurrEr$r>rr&r) rrrrCrrkernelsrrs && r_cuda_load_modulers '(G#sjj!__ F ZZ & & (F G,,V\\&-A3GH  6""G   ' ' T"FKK,@  $D1  N! s /0D'' D7 cHV^8dQhR\R\R\R\/#)r deviceoptional allow_cpur )rrr9)rs"rrr-s2@@ @@48@@rc$\V\4'dV#\V\4'd\P!V4p\V\P4'dHV'd!VP R9d\ RV 24hMVP R8wd\ RV 24h\PP4'g7\V\PP4'd VP#\WV4#)aGet the device index from :attr:`device`, which can be a torch.device object, a Python integer, or ``None``. If :attr:`device` is a torch.device object, returns the device index if it is a CUDA device. Note that for a CUDA device without a specified index, i.e., ``torch.device('cuda')``, this will return the current default CUDA device if :attr:`optional` is ``True``. If :attr:`allow_cpu` is ``True``, CPU devices will be accepted and ``-1`` will be returned in this case. If :attr:`device` is a Python integer, it is returned as is. If :attr:`device` is ``None``, this will return the current default CUDA device if :attr:`optional` is ``True``. ruz(Expected a cuda or cpu device, but got: z!Expected a cuda device, but got: )rucpu) rNr9rrrrrjit is_scriptingruidx_torch_get_device_index)rrrs&&&rrr-s &# &#f%&%,,'' {{/1 #KF8!TUU2 [[F "@IJ J 99 ! ! # # fejj// 0 0::  "6Y ??r)NNNFr3)FF)r rtypingrr cuda.bindingsrrJrIr torch._utilsrr"r,r0r4rErTrpr{r~rrrrrr/rrr&s   F.+# 7::6&$0O$dVV