[Mlir-commits] [mlir] [MLIR][NVVM] Add S2G and Reduce override NVVM Dialect ops (PR #216481)

Guray Ozen llvmlistbot at llvm.org
Wed Aug 19 00:18:10 PDT 2026


================
@@ -120,6 +120,32 @@ static LogicalResult cpAsyncBulkTensorCommonVerifier(size_t tensorDims,
   return success();
 }
 
+LogicalResult CpAsyncBulkTensorOverrideAddrCommonVerifier(
+    OperandRange coordinates, OperandRange tensorSize, OperandRange lowerStride,
+    Value upperStride, bool isTile, Location loc) {
+  LogicalResult res = success();
+  if (!tensorSize.empty() && coordinates.size() != tensorSize.size())
+    res =
+        emitError(loc, "Expected coordinates size to be equal to tensor size");
+
+  if (!lowerStride.empty() && lowerStride.size() != tensorSize.size() - 1)
+    res = emitError(
+        loc,
+        "Expected lower_stride size to be equal to one less than tensor size");
----------------
grypp wrote:

maybe add {}

https://github.com/llvm/llvm-project/pull/216481


More information about the Mlir-commits mailing list