Closed joshuia closed 5 years ago
all error message:
Check failed: error == cudaSuccess CUDA error CUDA driver version is insufficient for CUDA runtime version at /lustre04/scratch/zhuzhaoc/conda/conda-bld/graphvite_1565206157156/work/include/core/solver.h:182 Check failure stack trace: @ 0x7f92dc7294dd google::LogMessage::Fail() @ 0x7f92dc731071 google::LogMessage::SendToLog() @ 0x7f92dc728ecd google::LogMessage::Flush() @ 0x7f92dc72a76a google::LogMessageFatal::~LogMessageFatal() @ 0x7f92db37e0b9 graphvite::CudaCheck() @ 0x7f92db496105 graphvite::SolverMixin<>::SolverMixin() @ 0x7f92db4d46b5 _ZZN8pybind1112cpp_function10initializeIZNS_6detail8initimpl11constructorIJSt6vectorIiSaIiEEimEE7executeINS_6class_IN9graphvite11GraphSolverILm128EfjEEJEEEJNS_10call_guardIJNS_18gil_scoped_releaseEEEENS_5arg_vESI_SI_ELi0EEEvRT_DpRKT0_EUlRNS2_16value_and_holderES7_imE_vJSQ_S7_imEJNS_4nameENS_9is_methodENS_7siblingENS2_24is_new_style_constructorESH_SI_SI_SI_EEEvOSJ_PFT0_DpT1_EDpRKT2_ENUlRNS2_13function_callEE1_4FUNES17 @ 0x7f92db403469 pybind11::cpp_function::dispatcher() @ 0x7f92e458fb73 PyObject_Call @ 0x7f92e459ea4d instancemethod_call @ 0x7f92e458fb73 PyObject_Call @ 0x7f92e45e8a84 slot_tp_init @ 0x7f92e45e533b type_call @ 0x7f92e458fb73 PyObject_Call @ 0x7f92e4624a6e PyEval_EvalFrameEx @ 0x7f92e462ba99 PyEval_EvalCodeEx @ 0x7f92e45b46da function_call @ 0x7f92e458fb73 PyObject_Call @ 0x7f92e45e1f96 slot_tp_new @ 0x7f92e45e52f5 type_call @ 0x7f92e458fb73 PyObject_Call @ 0x7f92e4626119 PyEval_EvalFrameEx @ 0x7f92e462ba99 PyEval_EvalCodeEx @ 0x7f92e45b47c7 function_call @ 0x7f92e458fb73 PyObject_Call @ 0x7f92e4624a6e PyEval_EvalFrameEx @ 0x7f92e462ba99 PyEval_EvalCodeEx @ 0x7f92e45b47c7 function_call @ 0x7f92e458fb73 PyObject_Call @ 0x7f92e4624a6e PyEval_EvalFrameEx @ 0x7f92e462ba99 PyEval_EvalCodeEx @ 0x7f92e45b47c7 function_call
Sorry I haven't seen this issue before. Could you try conda list graphvite
and check if the signature is cuda92?
conda list graphvite
# Name Version Build Channel
graphvite 0.1.0 py27cuda100 milagraph
So it means when I used conda install -c milagraph graphvite
, it built with cuda 10.0?Could you tell me how to build with cuda 9.2?
OK, I found that conda install -c milagraph graphvite=0.1.0=py27cuda92
is OK, Thanks!
when I use "graphvite baseline quick start", it shows:
Check failed: error == cudaSuccess CUDA error CUDA driver version is insufficient for CUDA runtime version at /lustre04/scratch/zhuzhaoc/conda/conda-bld/graphvite_1565206157156/work/include/core/solver.h:182
but "nvcc -V" shows:
nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2018 NVIDIA Corporation Built on Wed_Apr_11_23:16:29_CDT_2018 Cuda compilation tools, release 9.2, V9.2.88
is cuda 9.2 error ?