when use GTX1070 and beta23, report error: CUDA_ERROR_ILLEGAL_ADDRESS

[P-Chao]

I can sucesseful run CPU version, and when use gpu cudnn5.1 win10 , another matconvnet program runs well, but when I run FCN training with 'opts.train.gpus = [1]', the error code:

I use matlab2017a, matconvnet1.0beta23, cudnn5.1, windows10, thank you very much.

More about the debug info,
when the first time enter vl_nnconv(), the param_size 1x1x4096x21 gpuarray, runs well.
when seconed time enter vl_nnconv(), the param_size 1x1x4090x4096 gpuarray, runs failed, report error:

`
cnn_train_dag: resetting GPU

ans =

CUDADevice with properties:

                  Name: 'GeForce GTX 1070'
                 Index: 1
     ComputeCapability: '6.1'
        SupportsDouble: 1
         DriverVersion: 8
        ToolkitVersion: 8
    MaxThreadsPerBlock: 1024
      MaxShmemPerBlock: 49152
    MaxThreadBlockSize: [1024 1024 64]
           MaxGridSize: [2.1475e+09 65535 65535]
             SIMDWidth: 32
           TotalMemory: 8.5899e+09
       AvailableMemory: 7.0877e+09
   MultiprocessorCount: 15
          ClockRateKHz: 1860000
           ComputeMode: 'Default'
  GPUOverlapsTransfers: 1
KernelExecutionTimeout: 1
      CanMapHostMemory: 1
       DeviceSupported: 1
        DeviceSelected: 1

train: epoch 01: 1/565:Error using vl_nnconv
An error occurred during PTX compilation of .
The information log was:

The error log was:

The CUDA error code was: CUDA_ERROR_ILLEGAL_ADDRESS.

Error in dagnn.Conv/backward (line 21)
[derInputs{1}, derParams{1}, derParams{2}] = vl_nnconv(...

Error in dagnn.Layer/backwardAdvanced (line 120)
[derInputs, derParams] = obj.backward ...

Error in dagnn.DagNN/eval (line 117)
obj.layers(l).block.backwardAdvanced(obj.layers(l)) ;

Error in cnn_train_dag>processEpoch (line 217)
net.eval(inputs, params.derOutputs, 'holdOn', s < params.numSubBatches) ;

Error in cnn_train_dag (line 87)
[net, state] = processEpoch(net, state, params, 'train') ;

Error in fcnTrain (line 98)
info = cnn_train_dag(net, imdb, getBatchWrapper(bopts), ...
`