checkin domain

python/tvm/__init__.py

@@ -5,3 +5,4 @@ from .op import *
 from .expr import Var, const
 from .expr_util import *
 from .tensor import Tensor
+from .domain import RDom, Range

python/tvm/domain.py

@@ -0,0 +1,38 @@
+from __future__ import absolute_import as _abs
+from . import expr as _expr
+from . import expr_util as _expr_util
+
+
+class Range(object):
+    """Represent a range in one dimension.
+    """
+    def __init__(self, begin, end=None):
+        if end is None:
+            end = begin
+            begin = _expr.const(0)
+        self.begin = _expr._symbol(begin)
+        self.end = _expr._symbol(end)
+        self.extent = _expr_util.simplify(end - begin)
+
+    def __str__(self):
+        return "(%s, %s)" % (
+            _expr_util.format_str(self.begin),
+            _expr_util.format_str(self.end))
+
+    def __repr__(self):
+        return self.__str__()
+
+class RDom(object):
+    """reduction Domain
+    """
+    def __init__(self, domain):
+        if isinstance(domain, Range):
+            domain = [domain]
+        self.index = []
+        self.domain = domain
+        for i in range(len(domain)):
+            self.index.append(_expr.Var("rd_index_%d_" % i))
+
+
+"""Use list of ranges as domain"""
+Domain = list

python/tvm/expr.py

@@ -108,7 +108,27 @@ class UnaryOpExpr(Expr):
         self.src = _symbol(src)
 
     def children(self):
-        return (self.src)
+        return (self.src,)
+
+
+class ReduceExpr(Expr):
+    def __init__(self, op, src,  rdom):
+        self.op = op
+        self.src = src
+        self.rdom = rdom
+
+    def children(self):
+        return (self.src,)
+
+
+class TensorReadExpr(Expr):
+    """Tensor read expression, tensor[indices]"""
+    def __init__(self, tensor, indices):
+        self.tensor = tensor
+        self.indices = indices
+
+    def children(self):
+        return self.indices
 
 
 def const(value):

python/tvm/expr_util.py

@@ -2,7 +2,6 @@
 from __future__ import absolute_import as _abs
 from . import expr as _expr
 from . import op as _op
-from . import tensor as _tensor
 
 def expr_with_new_children(e, children):
     """Returns same expr as e but with new children
@@ -50,10 +49,27 @@ def transform(e, f):
     result : return value of f
         The final result of transformation.
     """
-    assert isinstance(e, _expr.Expr)
+    if not isinstance(e, _expr.Expr):
+        raise TypeError("Cannot handle type %s" % type(e))
     return f(e , [transform(c, f) for c in e.children()])
 
 
+def visit(e, f):
+    """Apply f to each element of e
+
+    Parameters
+    ----------
+    e : Expr
+       The input expression.
+
+    f : function with signiture (e)
+    """
+    assert isinstance(e, _expr.Expr)
+    for c in e.children():
+        visit(c, f)
+    f(e)
+
+
 def format_str(expr):
     """change expression to string.
 
@@ -76,12 +92,15 @@ def format_str(expr):
             return str(e.value)
         elif isinstance(e, _expr.Var):
             return e.name
-        elif isinstance(e, _tensor.TensorReadExpr):
+        elif isinstance(e, _expr.TensorReadExpr):
             return "%s(%s)" % (e.tensor.name, ','.join(result_children))
+        elif isinstance(e, _expr.ReduceExpr):
+            return e.op.format_reduce_str(result_children[0], e.rdom.domain)
         else:
             raise TypeError("Do not know how to handle type " + str(type(e)))
     return transform(expr, make_str)
 
+
 def simplify(expr):
     """simplify expression
 

python/tvm/op.py

@@ -22,15 +22,20 @@ def canonical_to_expr(c):
     else:
         return _expr.const(0)
 
+
 class BinaryOp(object):
     """Base class of binary operator"""
     def __call__(self, lhs, rhs):
         return _expr.BinaryOpExpr(self, lhs, rhs)
 
+
 class AddOp(BinaryOp):
     def format_str(self, lhs, rhs):
         return '(%s + %s)' % (lhs, rhs)
 
+    def format_reduce_str(self, src, rd):
+        return "reduce_sum(%s, rdom=%s)" % (src, str(rd))
+
     def canonical(self, lhs, rhs):
         lhs = lhs.copy()
         for k, v in rhs.items():
@@ -40,6 +45,7 @@ class AddOp(BinaryOp):
                 lhs[k] = v
         return lhs
 
+
 class SubOp(BinaryOp):
     def format_str(self, lhs, rhs):
         return '(%s - %s)' % (lhs, rhs)
@@ -53,6 +59,7 @@ class SubOp(BinaryOp):
                 lhs[k] = -v
         return lhs
 
+
 class MulOp(BinaryOp):
     def format_str(self, lhs, rhs):
         return '(%s * %s)' % (lhs, rhs)
@@ -72,6 +79,7 @@ class MulOp(BinaryOp):
             return rhs
         return {elhs * erhs: 1}
 
+
 class DivOp(BinaryOp):
     def format_str(self, lhs, rhs):
         return '(%s / %s)' % (lhs, rhs)
@@ -86,6 +94,7 @@ class DivOp(BinaryOp):
         elhs = canonical_to_expr(lhs)
         return {elhs / erhs: 1}
 
+
 class MaxOp(BinaryOp):
     def format_str(self, lhs, rhs):
         return 'max(%s, %s)' % (lhs, rhs)
@@ -97,6 +106,7 @@ class MaxOp(BinaryOp):
             return lhs if ediff.value >= 0 else rhs
         return {MaxOp()(lhs, rhs): 1}
 
+
 class MinOp(BinaryOp):
     def format_str(self, lhs, rhs):
         return 'min(%s, %s)' % (lhs, rhs)
@@ -120,3 +130,16 @@ _expr.__addop__ = add
 _expr.__subop__ = sub
 _expr.__mulop__ = mul
 _expr.__divop__ = div
+
+
+def reduce_sum(expr, rdom):
+    return _expr.ReduceExpr(add, expr, rdom)
+
+def reduce_prod(expr, rdom):
+    return _expr.ReduceExpr(mul, expr, rdom)
+
+def reduce_min(expr, rdom):
+    return _expr.ReduceExpr(min, expr, rdom)
+
+def reduce_max(expr, rdom):
+    return _expr.ReduceExpr(max, expr, rdom)

python/tvm/tensor.py

@@ -1,33 +1,70 @@
 from __future__ import absolute_import as _abs
 from . import expr as _expr
-
-class TensorReadExpr(_expr.Expr):
-    def __init__(self, tensor, indices):
-        self.tensor = tensor
-        self.indices = indices
-
-    def children(self):
-        return self.indices
+from . import expr_util as _expr_util
 
 
 class Tensor(object):
-    def __init__(self, ndim, fcompute=None, name=None):
+    def __init__(self, ndim, fcompute=None, name=None, shape=None):
         self.ndim = ndim
         if fcompute:
             arg_names = fcompute.func_code.co_varnames
             assert(len(arg_names) == ndim)
             self.dim_index = [_expr.Var(n) for n in arg_names]
             self.expr = fcompute(*self.dim_index)
+            if shape is None:
+                raise ValueError("argument shape need to be given for intermediate tensor")
+            self.shape = shape
         else:
             self.expr = None
             self.dim_index = None
             shape_name = '_shape'
             if name: shape_name = name + shape_name
-            self.shape = tuple(_expr.Var("%s_%d_" % (shape_name, i)) for i in range(ndim))
+            self.shape = shape if shape else tuple(
+                _expr.Var("%s_%d_" % (shape_name, i)) for i in range(ndim))
 
         self.name = name if name else "TensorObj"
+        self.inputs = None
 
     def __call__(self, *indices):
         if len(indices) != self.ndim:
             raise ValueError("Need to provide %d index in tensor slice" % self.ndim)
-        return TensorReadExpr(self, indices)
+        return _expr.TensorReadExpr(self, indices)
+
+    def input_tensors(self):
+        """List of input tensors to this tensor.
+
+        Returns
+        -------
+        inputs : list of input tensors
+        """
+        if self.inputs is not None:
+            return self.inputs
+        self.inputs = []
+        if self.expr:
+            def collect(e):
+                if isinstance(e, _expr.TensorReadExpr):
+                    self.inputs.append(e.tensor)
+            _expr_util.visit(self.expr, collect)
+        return self.inputs
+
+    def infer_input_domains(self, out_domain):
+        """Infer the input domains of each domain given output domains
+
+        Parameters
+        ----------
+        out_domain : list of Range
+            Domain of each dimension.
+
+        Returns
+        -------
+        in_domains: dict Tensor->Domain
+        """
+        assert self.expr
+        assert len(out_domain) == len(self.dim_index)
+        index_domains = {
+            self.dim_index[i] : out_domain[i] for i in range(len(out_domain))
+        }
+        def collect(e):
+            if isinstance(e, _expr.TensorReadExpr):
+                self.inputs.append(e.tensor)
+        _expr_util.visit(self.expr, collect)

tests/python/test_tensor.py

@@ -3,8 +3,27 @@ import tvm
 def test_tensor():
     A = tvm.Tensor(2, name='A')
     B = tvm.Tensor(2, name='B')
-    T = tvm.Tensor(3, lambda i, j, k: A(i, k) * B(j, k))
+    T = tvm.Tensor(3, lambda i, j, k: A(i, k) * B(j, k),
+                   shape=(A.shape[0], B.shape[0], A.shape[1]))
     print(tvm.format_str(T.expr))
 
+def test_tensor_inputs():
+    A = tvm.Tensor(2, name='A')
+    B = tvm.Tensor(2, name='B')
+    T = tvm.Tensor(3, lambda i, j, k: A(i, k) * B(j, k),
+                   shape=(A.shape[0], B.shape[0], A.shape[1]))
+    assert(T.input_tensors() == [A, B])
+
+def test_tensor_reduce():
+    A = tvm.Tensor(2, name='A')
+    B = tvm.Tensor(2, name='B')
+    T = tvm.Tensor(3, lambda i, j, k: A(i, k) * B(j, k),
+                   shape=(A.shape[0], B.shape[0], A.shape[1]))
+    rd = tvm.RDom(tvm.Range(A.shape[1]))
+    C = tvm.Tensor(2, lambda i, j: tvm.reduce_sum(T(i, j, rd.index[0]), rdom=rd),
+                   shape=(A.shape[0], B.shape[0]))
+    print(tvm.format_str(C.expr))
+
 if __name__ == "__main__":
-    test_tensor()
+    test_tensor_inputs()
+    test_tensor_reduce()