[TOPI] parallel schedule improve for x86 & layout_transform support (#1130)

* add layout_transform. add schedule support for ndim>=5 for x86

* fix lint

topi/include/topi/nn.h

@@ -482,5 +482,27 @@ inline tvm::Tensor group_conv2d_ngchw(const tvm::Tensor& I,
   return tvm::compute(output_shape, l, name, tag);
 }
 
+using FLayoutIndicesTransform = std::function<Array<Expr>(const Array<Var>& indices)>;
+/*!
+ * \brief Transform the layout according to the mapping function \p to_src_indices.
+ * \param src the source input.
+ * \param dst_shape the output shape.
+ * \param to_src_indices the mapping function from input index to output index.
+ * \param name output tensor name.
+ * \param tag output tensor tag.
+ * \return A tensor with shape \p dst_shape.
+ */
+inline Tensor layout_transform(const Tensor& src,
+                               const Array<Expr>& dst_shape,
+                               const FLayoutIndicesTransform& to_src_indices,
+                               const std::string name = "layout_transform",
+                               const std::string tag = kInjective) {
+  auto src_shape = src->shape;
+  return compute(
+  dst_shape, [&](const Array<Var>& dst_indices) {
+    return src(to_src_indices(dst_indices));
+  }, name, tag);
+}
+
 }  // namespace topi
 #endif  // TOPI_NN_H_

topi/python/topi/generic/nn.py

+# pylint: disable=invalid-name,unused-argument
 """Generic nn operators"""
 from __future__ import absolute_import as _abs
 import tvm
@@ -54,6 +55,32 @@ def schedule_conv2d_nhwc(outs):
 
 
 @tvm.target.generic_func
+def schedule_conv2d_NCHWc(num_filter, kernel_size, strides, padding, outs):
+    """Schedule for conv2d_NCHW[x]c
+
+    Parameters
+    ----------
+    num_filter: int
+                The number of filter, i.e., the output channel.
+    kernel_size: tuple of int
+                 (kernel_height, kernel_width)
+    strides: tuple of int
+             (stride_of_height, stride_of_width)
+    padding: tuple of int
+             (pad_of_height, pad_of_width)
+    outs: Array of Tensor
+          The computation graph description of conv2d_NCHWc
+          in the format of an array of tensors.
+
+    Returns
+    -------
+    sch: Schedule
+        The computation schedule for the op.
+    """
+    return _default_schedule(outs, False)
+
+
+@tvm.target.generic_func
 def schedule_conv2d_transpose_nchw(outs):
     """Schedule for conv2d_transpose_nchw
 

topi/python/topi/x86/__init__.py

@@ -7,3 +7,4 @@ from .binarize_pack import schedule_binarize_pack
 from .binary_dense import schedule_binary_dense
 from .nn import *
 from .injective import *
+from .pooling import schedule_pool, schedule_global_pool

topi/python/topi/x86/injective.py

@@ -23,13 +23,11 @@ def schedule_injective(outs):
     x = outs[0]
     s = tvm.create_schedule([x.op for x in outs])
     tvm.schedule.AutoInlineInjective(s)
-    if len(s[x].op.axis) == 4:
-        n, c, _, _ = s[x].op.axis
-        fused = s[x].fuse(n, c) # for nhwc layout, fuse n and h
+    if len(s[x].op.axis) >= 5:
+        fused = s[x].fuse(s[x].op.axis[0], s[x].op.axis[1], s[x].op.axis[2])
         s[x].parallel(fused)
-    elif len(s[x].op.axis) == 5:
-        n, C, h, _, _ = s[x].op.axis
-        fused = s[x].fuse(n, C, h)
+    elif len(s[x].op.axis) >= 3:
+        fused = s[x].fuse(s[x].op.axis[0], s[x].op.axis[1])
         s[x].parallel(fused)
     else:
         s[x].parallel(s[x].op.axis[0])

topi/python/topi/x86/nn.py

@@ -5,23 +5,6 @@ import tvm
 from .. import generic
 from .. import tag
 
-def _default_schedule(outs, auto_inline):
-    """Default schedule for x86."""
-    x = outs[0]
-    s = tvm.create_schedule([x.op for x in outs])
-    if auto_inline:
-        tvm.schedule.AutoInlineInjective(s)
-        s[x].fuse(s[x].op.axis)
-        return s
-    if len(s[x].op.axis) == 4:
-        n, c, _, _ = s[x].op.axis
-        fused = s[x].fuse(n, c) # for nhwc layout, fuse n and h
-        s[x].parallel(fused)
-    else:
-        s[x].parallel(s[x].op.axis[0])
-    return s
-
-
 @generic.schedule_softmax.register(["cpu"])
 def schedule_softmax(outs):
     """Schedule for softmax
@@ -37,25 +20,19 @@ def schedule_softmax(outs):
     sch: Schedule
         The computation schedule for the op.
     """
-    return _default_schedule(outs, False)
-
-
-@generic.schedule_pool.register(["cpu"])
-def schedule_pool(outs):
-    """Schedule for pool
-
-    Parameters
-    ----------
-    outs: Array of Tensor
-          The computation graph description of pool
-          in the format of an array of tensors.
-
-    Returns
-    -------
-    sch: Schedule
-        The computation schedule for the op.
-    """
-    return _default_schedule(outs, False)
+    outs = [outs] if isinstance(outs, tvm.tensor.Tensor) else outs
+    x = outs[0]
+    s = tvm.create_schedule([x.op for x in outs])
+    tvm.schedule.AutoInlineInjective(s)
+    if len(s[x].op.axis) >= 5:
+        fused = s[x].fuse(s[x].op.axis[0], s[x].op.axis[1], s[x].op.axis[2])
+        s[x].parallel(fused)
+    elif len(s[x].op.axis) >= 3:
+        fused = s[x].fuse(s[x].op.axis[0], s[x].op.axis[1])
+        s[x].parallel(fused)
+    else:
+        s[x].parallel(s[x].op.axis[0])
+    return s
 
 
 @generic.schedule_dense.register(["cpu"])

topi/python/topi/x86/pooling.py

+# pylint: disable=invalid-name, unused-variable
+"""Schedule for pooling operators"""
+import tvm
+from .. import generic
+from .. import tag
+
+def _parallel_sch(sch):
+    if len(sch.op.axis) >= 5:
+        fused = sch.fuse(sch.op.axis[0], sch.op.axis[1], sch.op.axis[2])
+        sch.parallel(fused)
+    elif len(sch.op.axis) >= 3:
+        n, c, _, _ = sch.op.axis
+        fused = sch.fuse(n, c) # for nhwc layout, fuse n and h
+        sch.parallel(fused)
+    else:
+        sch.parallel(sch.op.axis[0])
+
+
+@generic.schedule_pool.register(["cpu"])
+def schedule_pool(outs):
+    """Schedule for pool
+
+    Parameters
+    ----------
+    outs: Array of Tensor
+          The computation graph description of pool
+          in the format of an array of tensors.
+
+    Returns
+    -------
+    sch: Schedule
+        The computation schedule for the op.
+    """
+    outs = [outs] if isinstance(outs, tvm.tensor.Tensor) else outs
+    s = tvm.create_schedule([x.op for x in outs])
+
+    def _schedule(PaddedInput, Pool):
+        if isinstance(PaddedInput.op, tvm.tensor.ComputeOp):
+            s[PaddedInput].compute_inline()
+        _parallel_sch(s[Pool])
+
+    def traverse(OP):
+        """Internal travserse function"""
+        # inline all one-to-one-mapping operators except the last stage (output)
+        if tag.is_broadcast(OP.tag):
+            if OP not in s.outputs:
+                s[OP].compute_inline()
+            for tensor in OP.input_tensors:
+                if tensor.op.input_tensors:
+                    traverse(tensor.op)
+        # schedule pool
+        elif OP.tag.startswith('pool'):
+            PaddedInput = OP.input_tensors[0]
+            Pool = OP.output(0)
+            _schedule(PaddedInput, Pool)
+        else:
+            raise RuntimeError("Unsupported operator: %s" % OP.tag)
+    traverse(outs[0].op)
+    return s
+
+
+@generic.schedule_global_pool.register(["cpu"])
+def schedule_global_pool(outs):
+    """Schedule for global pool
+
+    Parameters
+    ----------
+    outs: Array of Tensor
+          The computation graph description of pool
+          in the format of an array of tensors.
+
+    Returns
+    -------
+    sch: Schedule
+        The computation schedule for the op.
+    """
+    outs = [outs] if isinstance(outs, tvm.tensor.Tensor) else outs
+    s = tvm.create_schedule([x.op for x in outs])
+    def traverse(OP):
+        """Internal travserse function"""
+        # inline all one-to-one-mapping operators except the last stage (output)
+        if tag.is_broadcast(OP.tag):
+            if OP not in s.outputs:
+                s[OP].compute_inline()
+            for tensor in OP.input_tensors:
+                if tensor.op.input_tensors:
+                    traverse(tensor.op)
+        # schedule pool
+        elif OP.tag.startswith('global_pool'):
+            Pool = OP.output(0)
+            _parallel_sch(s[Pool])
+        else:
+            raise RuntimeError("Unsupported operator: %s" % OP.tag)
+    traverse(outs[0].op)
+    return s