Announcement

Collapse
No announcement yet.

Haswell-Tuned Binaries Are Still Sometimes Faster On AMD Ryzen Than Znver1 Tuning

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

  • Haswell-Tuned Binaries Are Still Sometimes Faster On AMD Ryzen Than Znver1 Tuning

    Phoronix: Haswell-Tuned Binaries Are Still Sometimes Faster On AMD Ryzen Than Znver1 Tuning

    Yesterday I posted benchmarks of Clang 4 vs. Clang 5 with AMD Ryzen now that LLVM 5 has received the initial AMD Zen "znver1" scheduler model. That scheduler model and improvements to LLVM/Clang 5.0 yielded some performance improvements, but still there are cases where using "-march=haswell" for tuning and running those generated binaries on Ryzen leads to better performance...

    http://www.phoronix.com/scan.php?pag...s-Znver1-GCC-5

  • #2
    Thanks for running those tests like I suggested. It's not much of a surprise considering how much of Haswell's design made it into Zen minus some AVX hardware.

    Comment


    • #3
      It looks like Clang devs just added a dummy switch and called it a day.

      Comment


      • #4
        Originally posted by eydee View Post
        It looks like Clang devs just added a dummy switch and called it a day.
        It makes things worse, I wouldn't call it a dummy switch

        Comment


        • #5
          I don't see this as a bad thing at all. Maybe it's screwed up that AMD's own optimizations aren't very well optimized, but it is great thing that Intels optimizations are performing so good. I mean we all know that Intel has spent literally billions of dollars convincing proprietary software developers to screw AMD and optimize their binaries only for Intel.... So this is fantastic news.....
          Last edited by duby229; 07-21-2017, 11:21 PM.

          Comment

          Working...
          X