Make the exercise section appear in the tutorial index page.

doc/tutorial/adding.txt

@@ -195,9 +195,10 @@ with NumPy arrays may be found here: :ref:`tensor creation<libdoc_tensor_creatio
    program will use 32- or 64-bit integers (``i`` prefix vs. the ``l`` prefix)
    and floats (``f`` prefix vs. the ``d`` prefix).
 
--------------------------------------------
 
-**Exercise**
+
+Exercise
+========
 
 .. code-block:: python
 

doc/tutorial/loop.txt

@@ -84,10 +84,10 @@ The full documentation can be found in the library: :ref:`Scan <lib_scan>`.
   # 19.0
 
 
--------------------------------------------
 
 
-**Exercise**
+Exercise
+========
 
 Run both examples.
 

doc/tutorial/modes.txt

@@ -36,9 +36,9 @@ variables, type this from the command-line:
 
 For more detail, see :ref:`Configuration <libdoc_config>` in the library.
 
--------------------------------------------
 
-**Exercise**
+Exercise
+========
 
 
 Consider the logistic regression:

doc/tutorial/using_gpu.txt

@@ -320,9 +320,9 @@ To change the value of a ``shared`` variable, e.g. to provide new data to proces
 use ``shared_variable.set_value(new_value)``. For a lot more detail about this,
 see :ref:`aliasing`.
 
--------------------------------------------
 
-**Exercise**
+Exercise
+========
 
 Consider again the logistic regression:
 
@@ -551,15 +551,14 @@ you feel competent enough, you may try yourself on the corresponding exercises.
   assert numpy.allclose(dest, a*b)
   print dest
 
--------------------------------------------
 
-**Exercise**
+Exercise
+========
 
 Run the preceding example.
 
 Modify and execute to work for a matrix of shape (20, 10).
 
--------------------------------------------
 
 
 .. _pyCUDA_theano:
@@ -615,9 +614,9 @@ Use this code to test it:
 >>> assert numpy.allclose(f(xv), xv*2)
 >>> print numpy.asarray(f(xv))
 
--------------------------------------------
 
-**Exercise**
+Exercise
+========
 
 
 Run the preceding example.