[Mlir-commits] [mlir] [mlir][AMDGPU] Add support for AMD f16 math library calls (PR #108809)

[Nirvedh Meshram]

================
@@ -89,7 +91,14 @@ struct OpToFuncCallLowering : public ConvertOpToLLVMPattern<SourceOp> {
 private:
   Value maybeCast(Value operand, PatternRewriter &rewriter) const {
     Type type = operand.getType();
-    if (!isa<Float16Type>(type))
+    if (!isa<FloatType>(type))
+      return operand;
+
+    // if there's a f16 function, no need to cast f16 values
+    if (!f16Func.empty() && isa<Float16Type>(type))
+      return operand;
+
+    if (isa<Float64Type>(type) || isa<Float32Type>(type))
----------------
nirvedhmeshram wrote:

That would be different functionality then what the function currently does but yes seems reasonable to do, probably want to add a test for it though. showing the casting happening for bf16 for example.

https://github.com/llvm/llvm-project/pull/108809