[Mlir-commits] [mlir] f981121 - [mlir][sparse] refined doc of sparse tensor ops

[Aart Bik]

Author: Aart Bik
Date: 2022-09-01T09:25:23-07:00
New Revision: f981121a66a40dfa8c60b6cb8f594f4fb97408f2

URL: https://github.com/llvm/llvm-project/commit/f981121a66a40dfa8c60b6cb8f594f4fb97408f2
DIFF: https://github.com/llvm/llvm-project/commit/f981121a66a40dfa8c60b6cb8f594f4fb97408f2.diff

LOG: [mlir][sparse] refined doc of sparse tensor ops

Reviewed By: bixia

Differential Revision: https://reviews.llvm.org/D133086

Added: 
    

Modified: 
    mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensorOps.td

Removed: 
    


################################################################################
diff  --git a/mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensorOps.td b/mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensorOps.td
index 25bc16fec96c7..3e1564f201cc0 100644
--- a/mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensorOps.td
+++ b/mlir/include/mlir/Dialect/SparseTensor/IR/SparseTensorOps.td
@@ -100,8 +100,8 @@ def SparseTensor_ToPointersOp : SparseTensor_Op<"pointers", [NoSideEffect]>,
     `bufferization.to_memref` operation in the sense that it provides a bridge
     between a tensor world view and a bufferized world view. Unlike the
     `bufferization.to_memref` operation, however, this sparse operation actually
-    lowers into a call into a support library to obtain access to the
-    pointers array.
+    lowers into code that extracts the pointers array from the sparse storage
+    scheme (either by calling a support library or through direct code).
 
     Example:
 
@@ -125,8 +125,8 @@ def SparseTensor_ToIndicesOp : SparseTensor_Op<"indices", [NoSideEffect]>,
     `bufferization.to_memref` operation in the sense that it provides a bridge
     between a tensor world view and a bufferized world view. Unlike the
     `bufferization.to_memref` operation, however, this sparse operation actually
-    lowers into a call into a support library to obtain access to the
-    indices array.
+    lowers into code that extracts the indices array from the sparse storage
+    scheme (either by calling a support library or through direct code).
 
     Example:
 
@@ -150,8 +150,8 @@ def SparseTensor_ToValuesOp : SparseTensor_Op<"values", [NoSideEffect]>,
     the `bufferization.to_memref` operation in the sense that it provides a bridge
     between a tensor world view and a bufferized world view. Unlike the
     `bufferization.to_memref` operation, however, this sparse operation actually
-    lowers into a call into a support library to obtain access to the
-    values array.
+    lowers into code that extracts the values array from the sparse storage
+    scheme (either by calling a support library or through direct code).
 
     Example:
 
@@ -195,8 +195,9 @@ def SparseTensor_ConcatenateOp : SparseTensor_Op<"concatenate", []>,
 // Sparse Tensor Management Operations. These operations are "impure" in the
 // sense that they do not properly operate on SSA values. Instead, the behavior
 // is solely defined by side-effects. These operations provide a bridge between
-// the code generator and the support library. The semantics of these operations
-// may be refined over time as our sparse abstractions evolve.
+// "sparsification" on one hand and a support library or actual code generation
+// on the other hand. The semantics of these operations may be refined over time
+// as our sparse abstractions evolve.
 //===----------------------------------------------------------------------===//
 
 def SparseTensor_LexInsertOp : SparseTensor_Op<"lex_insert", []>,
@@ -675,5 +676,4 @@ def SparseTensor_StorageSetOp : SparseTensor_Op<"storage_set", []>,
   let hasVerifier = 1;
 }
 
-
 #endif // SPARSETENSOR_OPS