Add in the doc cuDNN BN and new CPU parallelized operations

doc/library/sandbox/cuda/dnn.txt

@@ -173,3 +173,14 @@ Softmax Ops
 .. automodule:: theano.sandbox.cuda.dnn
    :noindex:
    :members: GpuDnnSoftmax, GpuDnnSoftmaxGrad
+
+
+
+.. _libdoc_cuda_dnn_bn:
+
+Batch Normalization
+===================
+
+.. automodule:: theano.sandbox.cuda.dnn
+   :noindex:
+   :members: dnn_batch_normalization_train, dnn_batch_normalization_test

doc/library/tensor/nnet/bn.txt

@@ -10,4 +10,7 @@
 .. moduleauthor:: LISA
 
 
+.. seealso:: :ref:`cuDNN batch normalization <libdoc_cuda_dnn_bn>`
+    must be added manually.
+
 .. autofunction:: theano.tensor.nnet.bn.batch_normalization

doc/tutorial/multi_cores.txt

@@ -4,6 +4,13 @@
 Multi cores support in Theano
 =============================
 
+Convolution and Pooling
+=======================
+
+Since Theano 0.9dev2, the convolution and pooling are parallelized on
+CPU.
+
+
 BLAS operation
 ==============
 

setup.py

@@ -54,7 +54,7 @@ PLATFORMS           = ["Windows", "Linux", "Solaris", "Mac OS-X", "Unix"]
 MAJOR               = 0
 MINOR               = 9
 MICRO               = 0
-SUFFIX              = "dev1"  # Should be blank except for rc's, betas, etc.
+SUFFIX              = "dev2"  # Should be blank except for rc's, betas, etc.
 ISRELEASED          = False
 
 VERSION             = '%d.%d.%d%s' % (MAJOR, MINOR, MICRO, SUFFIX)

theano/sandbox/cuda/dnn.py

@@ -2672,6 +2672,8 @@ def dnn_batch_normalization_train(inputs, gamma, beta, mode='per-activation',
 
     Notes
     -----
+    Request cuDNN 5 and Theano 0.9dev2 or more recent.
+
     For 4d tensors, returned values are equivalent to:
 
     >>> axes = 0 if mode == 'per-activation' else (0, 2, 3)
@@ -2736,6 +2738,8 @@ def dnn_batch_normalization_test(inputs, gamma, beta, mean, var,
 
     Notes
     -----
+    Request cuDNN 5 and Theano 0.9dev2 or more recent.
+
     For 4d tensors, the returned value is equivalent to:
 
     >>> axes = (0,) if mode == 'per-activation' else (0, 2, 3)