[Mlir-commits] [mlir] [MLIR][NVGPU] Add convert.fpext and convert.fptrunc Ops (PR #199700)

[Durgadoss R]

================
@@ -1709,6 +1711,669 @@ struct NVGPURcpOpLowering : public ConvertOpToLLVMPattern<nvgpu::RcpOp> {
         rewriter);
   }
 };
+
+//===----------------------------------------------------------------------===//
+// FPTruncOp Lowering
+//===----------------------------------------------------------------------===//
+
+/// Conversion op identifier for nvgpu.convert.fptrunc lowering dispatch table.
+enum class TruncConvOp {
----------------
durga4github wrote:

let us prefix it with "FP".. same for both `Kind` enums below...

https://github.com/llvm/llvm-project/pull/199700