[Mlir-commits] [mlir] Introduce `arith.scaling_extf` and `arith.scaling_truncf` (PR #141965)

Umang Yadav llvmlistbot at llvm.org
Fri Jun 6 16:41:25 PDT 2025


================
@@ -1280,6 +1333,64 @@ def Arith_TruncFOp :
                           attr-dict `:` type($in) `to` type($out) }];
 }
 
+//===----------------------------------------------------------------------===//
+// Scaling TruncFOp
+//===----------------------------------------------------------------------===//
+
+def Arith_ScalingTruncFOp
+    : Arith_Op<"scaling_truncf",
+               [Pure, SameInputOutputTensorDims,
+                DeclareOpInterfaceMethods<ArithRoundingModeInterface>,
+                DeclareOpInterfaceMethods<ArithFastMathInterface>,
+                DeclareOpInterfaceMethods<CastOpInterface>]>,
+      Arguments<(ins FloatLike:$in, FloatLike:$scale,
+          OptionalAttr<Arith_RoundingModeAttr>:$roundingmode,
+          OptionalAttr<Arith_FastMathAttr>:$fastmath)>,
+      Results<(outs FloatLike:$out)> {
+  let summary =
+      "Downcasts input floating point values using provided scales values following OCP MXFP Spec";
+  let description = [{
+    This operation quantizes input using the provided scale values. It expects 
----------------
umangyadav wrote:

Replaced with `downcast` word

https://github.com/llvm/llvm-project/pull/141965


More information about the Mlir-commits mailing list