Merge pull request #963 from lamblin/fix_cxx_noblas

theano/sparse/opt.py

@@ -675,7 +675,8 @@ local_usmm = gof.opt.PatternSub(
     (theano.tensor.sub, 'z',
      (theano.tensor.mul,
       {'pattern': 'alpha',
-       'constraint': lambda expr: numpy.all(expr.type.broadcastable)},
+       'constraint': lambda expr: (numpy.all(expr.type.broadcastable) and
+                                   theano.config.blas.ldflags)},
     (sparse._dot, 'x', 'y'))),
     (usmm, (theano.tensor.neg, 'alpha'), 'x', 'y', 'z'))
 register_specialize(local_usmm, name="local_usmm")
@@ -1646,6 +1647,9 @@ sampling_dot_csr = SamplingDotCSR()
 # register a specialization to replace SamplingDot -> SamplingDotCsr
 @gof.local_optimizer([sparse.sampling_dot])
 def local_sampling_dot_csr(node):
+    if not theano.config.blas.ldflags:
+        # The C implementation of SamplingDotCsr relies on BLAS routines
+        return
     if node.op == sparse.sampling_dot:
         x, y, p = node.inputs
         if p.type.format == 'csr':
@@ -1656,6 +1660,7 @@ def local_sampling_dot_csr(node):
 
             return [sparse.CSR(z_data, z_ind, z_ptr, p_shape)]
     return False
+
 sparse.register_specialize(local_sampling_dot_csr,
                            'cxx_only',
                            name='local_sampling_dot_csr')

theano/sparse/tests/test_basic.py

@@ -1204,7 +1204,16 @@ class UsmmTests(unittest.TestCase):
 
             fast_compile = theano.config.mode == "FAST_COMPILE"
 
-            if (y.type.dtype == up and format1 == 'csc' and format2 == 'dense'
+            if not theano.config.blas.ldflags:
+                # Usmm should not be inserted, because it relies on BLAS
+                assert len(topo) == 4, topo
+                assert isinstance(topo[0].op, theano.sparse.Dot)
+                assert isinstance(topo[1].op, theano.tensor.DimShuffle)
+                assert (isinstance(topo[2].op, theano.tensor.Elemwise) and
+                        isinstance(topo[2].op.scalar_op, theano.scalar.Mul))
+                assert (isinstance(topo[3].op, theano.tensor.Elemwise) and
+                        isinstance(topo[3].op.scalar_op, theano.scalar.Sub))
+            elif (y.type.dtype == up and format1 == 'csc' and format2 == 'dense'
                 and not fast_compile and theano.config.cxx and
                 up in ('float32', 'float64')):
                 # The op UsmmCscDense should be inserted

theano/sparse/tests/test_opt.py

@@ -130,5 +130,11 @@ def test_local_sampling_dot_csr():
                             sparse.sampling_dot(*inputs),
                             mode=mode)
 
+        if theano.config.blas.ldflags:
             assert not any(isinstance(node.op, sparse.SamplingDot) for node
                        in f.maker.fgraph.toposort())
+        else:
+            # SamplingDotCSR's C implementation needs blas, so it should not
+            # be inserted
+            assert not any(isinstance(node.op, sparse.opt.SamplingDotCSR) for node
+                       in f.maker.fgraph.toposort())

theano/tensor/nnet/Conv3D.py

@@ -51,7 +51,7 @@ class Conv3D(theano.Op):
         return "Conv3D"
 
     def c_code_cache_version(self):
-        return (2,)
+        return (3,)
 
 
     def make_node(self, V, W, b, d):
@@ -338,7 +338,8 @@ class Conv3D(theano.Op):
         #if the data types are not mixed, we can insert special case optimizations based on BLAS
         VV, WV, bv, dv = node.inputs
         HV = node.outputs[0]
-        if VV.dtype == WV.dtype and HV.dtype == VV.dtype:
+        if (theano.config.blas.ldflags and
+            VV.dtype == WV.dtype and HV.dtype == VV.dtype):
             if VV.dtype == 'float64':
                 gemv = 'dgemv_'
             elif VV.dtype == 'float32':