docker环境下paddledev/paddle:gpu-demo-v0.9.0a0 ，执行quickstart，Cuda Error: CUDA driver version is insufficient for CUDA runtime version
Created by: beckett1124
概述：
docker环境下paddledev/paddle:gpu-demo-v0.9.0a0 ，执行quickstart，Cuda Error: CUDA driver version is insufficient for CUDA runtime version。
I1129 11:18:51.543231 335 Util.cpp:155] commandline: /usr/local/bin/../opt/paddle/bin/paddle_trainer --config=trainer_config.lr.py --save_dir=./output --trainer_count=4 --log_period=100 --num_passes=15 --use_gpu=True
F1129 11:18:51.543578 335 hl_cuda_device.cc:526] Check failed: cudaSuccess == cudaStat (0 vs. 35) Cuda Error: CUDA driver version is insufficient for CUDA runtime version
* Check failure stack trace: *
@ 0x7f6d59f61daa (unknown)
@ 0x7f6d59f61ce4 (unknown)
@ 0x7f6d59f616e6 (unknown)
@ 0x7f6d59f64687 (unknown)
@ 0x895bf0 hl_specify_devices_start()
@ 0x895dfd hl_start()
@ 0x813362 paddle::initMain()
@ 0x52ac5b main
@ 0x7f6d5916df45 (unknown)
@ 0x540c05 (unknown)
@ (nil) (unknown)
/usr/local/bin/paddle: line 109: 335 Aborted ${DEBUGGER} $MYDIR/../opt/paddle/bin/paddle_trainer ${@:2}
I1129 11:18:52.197794 356 Util.cpp:155] commandline: /usr/local/bin/../opt/paddle/bin/paddle_trainer --config=trainer_config.lr.py --use_gpu=false --job=test --init_model_path=output/pass-00003 --config_args=is_predict=1 --predict_output_dir=.
I1129 11:18:52.197998 356 Util.cpp:130] Calling runInitFunctions
I1129 11:18:52.198359 356 Util.cpp:143] Call runInitFunctions done.
[WARNING 2016-11-29 11:18:52,343 networks.py:1438] outputs routine try to calculate network's inputs and outputs order. It might not work well.Please see follow log carefully.
[INFO 2016-11-29 11:18:52,343 networks.py:1466] The input order is [word]
[INFO 2016-11-29 11:18:52,343 networks.py:1472] The output order is [maxid_layer_0, fc_layer_0]
I1129 11:18:52.350263 356 Trainer.cpp:149] trainer: in testing mode
I1129 11:18:52.350291 356 Trainer.cpp:156] trainer mode: Testing
I1129 11:18:52.451740 356 PyDataProvider2.cpp:257] loading dataprovider dataprovider_bow::process_predict
I1129 11:18:52.453099 356 GradientMachine.cpp:123] Loading parameters from output/pass-00003
I1129 11:18:52.453135 356 Parameter.cpp:344] missing parameters [output/pass-00003/_fc_layer_0.w0] while loading model.
F1129 11:18:52.453147 356 Parameter.cpp:350] _fc_layer_0.w0 missing, not allowed.
* Check failure stack trace: *
@ 0x7efc8abdedaa (unknown)
@ 0x7efc8abdece4 (unknown)
@ 0x7efc8abde6e6 (unknown)
@ 0x7efc8abe1687 (unknown)
@ 0x825ce2 paddle::Parameter::load()
@ 0x5633a6 paddle::GradientMachine::loadParameters()
@ 0x6b689c paddle::ParameterUtil::loadParametersWithPath()
@ 0x6c0a6d paddle::Tester::test()
@ 0x52b718 main
@ 0x7efc89deaf45 (unknown)
@ 0x540c05 (unknown)
@ (nil) (unknown)
/usr/local/bin/paddle: line 109: 356 Aborted ${DEBUGGER} $MYDIR/../opt/paddle/bin/paddle_trainer ${@:2}
