Announcement

Collapse
No announcement yet.

GCC vs. LLVM/Clang On The AMD Richland APU

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #11
    Actually, they do: for someone who does not follow LLVM development, it's informational to know it still doesn't support it.

    Then again, this could be just mentioned in the text too

    Comment


    • #12
      Originally posted by curaga View Post
      Actually, they do: for someone who does not follow LLVM development, it's informational to know it still doesn't support it.

      Then again, this could be just mentioned in the text too
      yes,exactly. “we did not include openmp tests, because LLVM does not support it”.

      Comment


      • #13
        Originally posted by ArneBab View Post
        Please have a look at

        * Timed MAFFT alignment,
        * BLAKE2,
        * Botan MAC,
        * Himeno and
        * C-Ray (please call this one “not by much” again…)

        LLVM seems to be very fast at matrix multiplication, though.

        So my summary of the results would be:

        * LLVM is great at Successive Jacobi Relaxation.
        * GCC is great at C-Ray.
        * LLVM has no OpenMP support, so don’t even try to use it for scientific code, except if you want to go all the way and use explicit MPI (which makes the SciMark test somewhat less useful).
        I said just about any time, not any time. Yes GCC spanked LLVM/Clang in C-Ray.

        Comment


        • #14
          Originally posted by ArneBab View Post
          yes,exactly. “we did not include openmp tests, because LLVM does not support it”.
          As these forums have proven time and time again, nobody reads the text. You might as well just skip the openMP tests without explanation if that's what you're going to do.

          Comment

          Working...
          X