[PATCH] D74171: [mlir] [LLVMIR] add all vector reduction intrinsics to LLVM IR dialect

[Nicolas Vasilache via Phabricator via llvm-commits]

nicolasvasilache added inline comments.


================
Comment at: mlir/test/Target/llvmir-intrinsics.mlir:98
 
+// CHECK-LABEL: @vector_reductions
+llvm.func @vector_reductions(%arg0: !llvm.float, %arg1: !llvm<"<8 x float>">, %arg2: !llvm<"<8 x i32>">) {
----------------
nicolasvasilache wrote:
> Did you try to pipe these through `mlir-translate --mlir-to-llvmir` ?
> I've had surprises on my end with more complex masked operations.
I would also argue that we should have an extra test in this file that does the piping through `mlir-translate --mlir-to-llvmir` (no need to filecheck the result just that it doesn't crash).
However @ftynse repeatedly made the case against it.
But I don't buy the separability of tests argument here :) esp in light of how easy it is to get crashes when trying to go all the way down to actual LLVMIR.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D74171/new/

https://reviews.llvm.org/D74171