add GPUCorrMM conv3d gradinputs alternative

theano/gpuarray/opt.py

@@ -1980,6 +1980,31 @@ def local_abstractconv3d_gradinputs_gemm(node):
     return [rval]
 
 
+@local_optimizer([AbstractConv3d_gradInputs])
+def local_abstractconv3d_gradinputs_gemm_alt(node):
+    if not isinstance(node.op, AbstractConv3d_gradInputs):
+        return None
+    kern, topgrad, shape = node.inputs
+    if not isinstance(kern.type, GpuArrayType) or \
+            not isinstance(topgrad.type, GpuArrayType):
+        return None
+    border_mode = node.op.border_mode
+    subsample = node.op.subsample
+    filter_dilation = node.op.filter_dilation
+
+    if border_mode == 'valid' and subsample == (1, 1, 1):
+        if not node.op.filter_flip:
+            kern = kern[:, :, ::-1, ::-1, ::-1]
+        rval = GpuCorr3dMM(border_mode='full',
+                           subsample=subsample,
+                           filter_dilation=filter_dilation)(
+            gpu_contiguous(topgrad),
+            gpu_contiguous(kern.dimshuffle(1, 0, 2, 3, 4)))
+        return [rval]
+    else:
+        return None
+
+
 class ConvMetaOptimizer(LocalMetaOptimizer):
 
     def __init__(self, optimizers=()):
@@ -2680,6 +2705,7 @@ conv_metaopt.register([local_abstractconv_cudnn_alternative])
 conv_metaopt.register([local_abstractconv3d2d])
 conv_metaopt.register([local_abstractconv3d_alt])
 conv_metaopt.register([local_abstractconv3d_gemm_gradweights_alt])
+conv_metaopt.register([local_abstractconv3d_gradinputs_gemm_alt])
 abstractconv_groupopt.register('conv_metaopt', conv_metaopt, 'conv_meta', position=0)
 
 # Register cuDNN batch normalization implementation