[PATCH] D104126: [MicroBenchmarks] Add initial SLP vectorization benchmarks.

[Michael Kruse via Phabricator via llvm-commits]

Meinersbur added a comment.

Does this expect on loop unrolling to happened? Did you consider adding `#pragma clang loop unroll(full)` to make sure of it?



================
Comment at: MicroBenchmarks/SLPVectorization/Versioning.cpp:17
+template <unsigned N, typename T>
+__attribute__((noinline)) void do_xor_reduction(T *A, T *B) {
+  for (int i = 0; i < N; i++) {
----------------
Where does the "reduction" in the name come from?


Repository:
  rT test-suite

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D104126/new/

https://reviews.llvm.org/D104126