Merge pull request #484 from nouiz/doc

NEWS.txt

 .. _NEWS:
 
+Update in the Trunk since the last release:
+
+* Added in the tutorial documentation on how to extend Theano.
+  This explains how to make a Theano Op from a Python function.
+  http://deeplearning.net/software/theano/tutorial/extending_theano.html
+
+
 =============
 Release Notes
 =============

doc/cifarSC2011/index.txt

@@ -65,7 +65,7 @@ from gurus on hand if you get stuck.
     introduction
     theano
     advanced_theano
-    extending_theano
+    /tutorial/extending_theano
     pyCUDA
     gpundarray
 

doc/index.txt

@@ -17,7 +17,7 @@ Theano has been powering large-scale computationally intensive scientific invest
 since 2007.  But it is also approachable enough to be used in the classroom
 (IFT6266 at the University of Montreal).
 
-.. image:: images/talk2010.gif
+.. image:: images/talk2010.png
     :scale: 75%
     :align: left
 

doc/cifarSC2011/extending_theano.txt → doc/tutorial/extending_theano.txt

@@ -211,5 +211,11 @@ Exercises 8
 
   - Our current element-wise fusion generates computation with only 1 output.
 
-
-
+GPU Op
+------
+
+Op that execute on the GPU should inherit from the
+``theano.sandbox.cuda.GpuOp`` and not ``theano.Op``. This allows Theano
+to make the distinction between both. Currently, we use this to test
+if the NVIDIA driver works correctly with our sum reduction code on the
+gpu.

doc/tutorial/index.txt

@@ -38,5 +38,6 @@ you out.
     gpu_data_convert
     shape_info
     remarks
+    extending_theano
     debug_faq
     faq