[Openmp-commits] [PATCH] D149490: [OpenMP] In libomptarget, assume alignment at powers of two

[Joel E. Denny via Phabricator via Openmp-commits]

jdenny marked an inline comment as done.
jdenny added inline comments.


================
Comment at: openmp/libomptarget/src/omptarget.cpp:112
+  auto BaseAlignment = MaxAlignment;
+  while (reinterpret_cast<uintptr_t>(HstPtrBase) % BaseAlignment)
+    BaseAlignment /= 2;
----------------
jhuber6 wrote:
> Shouldn't this just be looking for the first non zero bit? Like `__builtin_ffsl` or similar?
That does seem like it would be more efficient.  I've not used it before, but it behaves in my CI.  Does the new code look reasonable to you?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D149490/new/

https://reviews.llvm.org/D149490