Batch normalization optimizations for old gpu backend.

theano/sandbox/cuda/opt.py

@@ -3050,3 +3050,28 @@ conv_groupopt.register('local_abstractconv3d_gradinputs_gemm',
                        local_abstractconv3d_gradinputs_gemm, 30,
                        'conv_gemm',
                        'gpu', 'fast_compile', 'fast_run')
+
+
+# Register cuDNN batch normalization implementation
+abstract_batch_norm_groupopt = theano.gof.optdb.LocalGroupDB()
+abstract_batch_norm_groupopt.__name__ = "gpu_batchnorm_opts"
+register_opt('fast_compile')(abstract_batch_norm_groupopt)
+
+# cuDNN optimizations are only registered if cuDNN is available.
+# (we import these opts here instead of at the top of this file
+# to avoid a circular dependency problem with dnn)
+from .dnn import (local_abstract_batch_norm_train_cudnn,
+                  local_abstract_batch_norm_train_grad_cudnn,
+                  local_abstract_batch_norm_inference_cudnn)     # noqa: 402
+abstract_batch_norm_groupopt.register('local_abstract_batch_norm_train_dnn',
+                                      local_abstract_batch_norm_train_cudnn, 20,
+                                      'batchnorm_dnn',
+                                      'gpu', 'fast_compile', 'fast_run', 'cudnn')
+abstract_batch_norm_groupopt.register('local_abstract_batch_norm_train_grad_dnn',
+                                      local_abstract_batch_norm_train_grad_cudnn, 20,
+                                      'batchnorm_dnn',
+                                      'gpu', 'fast_compile', 'fast_run', 'cudnn')
+abstract_batch_norm_groupopt.register('local_abstract_batch_norm_inference_dnn',
+                                      local_abstract_batch_norm_inference_cudnn, 20,
+                                      'batchnorm_dnn',
+                                      'gpu', 'fast_compile', 'fast_run', 'cudnn')