[CIR][LLVMLowering] Add LLVM lowering support for _Float16 (#716)

This PR adds LLVM lowering support for `_Float16` type. The only change we need to make here is adding a new type converter to the LLVM lowering pass. The majority of this PR is tests that check the generated LLVM IR. Later I'll add another separate PR that adds LLVM lowering support for the `__bf16` type. This PR is already big enough.
llvm · Jul 2, 2024 · a9eb565 · a9eb565
1 parent c8891bd
commit a9eb565
Show file tree

Hide file tree

Showing 2 changed files with 655 additions and 2 deletions.
diff --git a/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp b/clang/lib/CIR/Lowering/DirectToLLVM/LowerToLLVM.cpp
@@ -3418,6 +3418,9 @@ void prepareTypeConverter(mlir::LLVMTypeConverter &converter,
   converter.addConversion([&](mlir::cir::LongDoubleType type) -> mlir::Type {
     return converter.convertType(type.getUnderlying());
   });
+  converter.addConversion([&](mlir::cir::FP16Type type) -> mlir::Type {
+    return mlir::FloatType::getF16(type.getContext());
+  });
   converter.addConversion([&](mlir::cir::FuncType type) -> mlir::Type {
     auto result = converter.convertType(type.getReturnType());
     llvm::SmallVector<mlir::Type> arguments;