[Mlir-commits] [mlir] [mlir][x86vector] AVX2 I8 Dot Op (PR #147908)

Adam Siemieniuk llvmlistbot at llvm.org
Fri Jul 11 03:46:05 PDT 2025


================
@@ -86,6 +86,28 @@ x86vector::DotOp::getIntrinsicOperands(ArrayRef<Value> operands,
   return intrinsicOperands;
 }
 
+SmallVector<Value> x86vector::DotInt8Op::getIntrinsicOperands(
+    ArrayRef<Value> operands, const LLVMTypeConverter &typeConverter,
+    RewriterBase &rewriter) {
+  SmallVector<Value, 3> intrinsicOprnds;
+  intrinsicOprnds.push_back(operands[0]);
----------------
adam-smnk wrote:

Let's wrap the `operands` in the op adaptor and use its named getters for `W`, `A`, and `B` (see other examples in this file).

https://github.com/llvm/llvm-project/pull/147908


More information about the Mlir-commits mailing list