[TIR] Introduce BufferLoad/Store (apache#5205)

Co-authored-by: Siyuan Feng <[email protected]> This PR introduces BufferLoad/Store to TIR. The new nodes will replace Provide and Call with Tensor arguments in the subsequent refactors.
neo-ai · Apr 17, 2020 · 02c4118 · 02c4118
1 parent 0464084
commit 02c4118
Show file tree

Hide file tree

Showing 15 changed files with 590 additions and 326 deletions.
diff --git a/include/tvm/tir/buffer.h b/include/tvm/tir/buffer.h
@@ -25,9 +25,8 @@
 #define TVM_TIR_BUFFER_H_
 
 #include <tvm/node/container.h>
-#include <tvm/tir/expr.h>
-#include <tvm/tir/op.h>
-
+#include <tvm/ir/expr.h>
+#include <tvm/tir/var.h>
 #include <string>
 
 
@@ -36,6 +35,9 @@ namespace tir {
 // Internal node container Buffer
 class BufferNode;
 
+// forward declare Stmt
+class Stmt;
+
 /*! \brief buffer type */
 enum BufferType : int {
   kDefault = 1,
@@ -75,9 +77,9 @@ class Buffer : public ObjectRef {
    * \param offset The offset of ptr.
    */
   TVM_DLL PrimExpr access_ptr(int access_mask,
-                          DataType ptr_type = DataType::Handle(),
-                          int content_lanes = 1,
-                          PrimExpr offset = make_const(DataType::Int(32), 0)) const;
+                              DataType ptr_type = DataType::Handle(),
+                              int content_lanes = 1,
+                              PrimExpr offset = IntImm(DataType::Int(32), 0)) const;
   /*!
    * \brief Create an Expr that does a vector load at begin index.
    * \param begin The beginning index