[PATCH] D50530: Added LLVM metadata to generated IR to increase vectorization width

[Matt Morehouse via Phabricator via llvm-commits]

morehouse added inline comments.


================
Comment at: clang/tools/clang-fuzzer/proto-to-llvm/loop_proto_to_llvm.cpp:141
+            << "!1 = !{!\"llvm.loop.vectorize.enable\", i1 true}\n"
+            << "!2 = !{!\"llvm.loop.vectorize.width\", i32 " << kArraySize
+            << "}\n";
----------------
I'm not sure `kArraySize` is what you want here.  Does "width" refer to array size or the SIMD width?

Any problem with letting the vectorizer determine this automatically?


================
Comment at: clang/tools/clang-fuzzer/proto-to-llvm/loop_proto_to_llvm.cpp:142
+            << "!2 = !{!\"llvm.loop.vectorize.width\", i32 " << kArraySize
+            << "}\n";
 }
----------------
Does this metadata change coverage in the vectorizer?


================
Comment at: clang/tools/clang-fuzzer/proto-to-llvm/loop_proto_to_llvm.h:18
 
+#include "../handle-llvm/input_arrays.h"
+
----------------
This include should be in the cpp file.


Repository:
  rC Clang

https://reviews.llvm.org/D50530