[Mlir-commits] [mlir] [mlir][LLVMIR] Extend FP array-splat constant lowering (PR #192378)

[llvmlistbot at llvm.org]

================
@@ -683,6 +682,26 @@ llvm::Constant *mlir::LLVM::detail::getLLVMConstant(
             }
           }
         }
+        if (elementType->isFloatingPointTy()) {
+          if (llvm::ConstantFP *cfp = dyn_cast<llvm::ConstantFP>(child)) {
+            APInt bitPattern = cfp->getValueAPF().bitcastToAPInt();
+            uint64_t value = bitPattern.getZExtValue();
+            // TODO: This code only handles 16, 32, and 64 bit floats. Handle
----------------
zackc6 wrote:

I didn’t find corresponding apis, either.
I think there fp8 and fp4 support is not so mature (comparing to f32) in the code base. That’s why I left TODO here.

https://github.com/llvm/llvm-project/pull/192378