Announcement

Collapse
No announcement yet.

RadeonSI Polaris: Mesa 12.0 vs. 13.0 vs. 17.0 vs. 17.1 Git

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

  • RadeonSI Polaris: Mesa 12.0 vs. 13.0 vs. 17.0 vs. 17.1 Git

    Phoronix: RadeonSI Polaris: Mesa 12.0 vs. 13.0 vs. 17.0 vs. 17.1 Git

    With Mesa 17.1 branching this weekend I figured it would be a fun Easter running benchmarks of Mesa Git compared to previous branches with a Radeon RX 470 Polaris graphics card. Here are these Mesa 17.1 benchmarks while other tests and on more GPUs is forthcoming.

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    WTF! i didnt absolutely see this coming. all your other tests show increasing performance so this is a big disappointment

    Comment


    • #3
      Huh, results for Deus Ex: Mankind Divided seem to show some hidden message when rotated. ;-)
      Well, hopefully this will improve before final release, or in some point release.

      Comment


      • #4
        No reason to feel disappointed. This is exactly the reason why you can't extrapolate benchmark results between hardware. It's not like we didn't already know that by like 50 years ago.

        Comment


        • #5
          Originally posted by davidbepo View Post
          WTF! i didnt absolutely see this coming. all your other tests show increasing performance so this is a big disappointment
          Well all others were on Ubuntu 16.10 and this is 17.04... different env you know, X is newer, etc...

          Also for 13 llvm matches 3.9, 17.0 have 4.0... and 17.1 is current development and things breaks in git, devs use llvm 5 git etc... It is normal that things breaks during development.

          Comment


          • #6
            Well, yeah, it is disappointing to see regressions, but it is not "stable" point release. So no problem there, good that someone test and find regressions

            Comment


            • #7
              This article shows that 480 performs very well with Mesa 17.1:
              Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite


              470 shouldn't be much different. I guess evictions due to CPU access to VRAM might explain the slowdown.

              Comment


              • #8
                Michael, you can test Tomb Raider with the maximum settings with radeonsi, no graphical glitches at all. Here, with a i7 3770k and a RX470 identical to yours, I get 58fps @ 1080p in the ultimate settings, and that includes the TressFX hair effect.

                Comment


                • #9
                  There is definitely a regression with Deux Ex. I using Padoka's PPA and, while me result is a little better, it's worse than before.

                  And Euro Truck Simulator 2 is running like crap with the lastest Mesa. Using Ubuntu 17.04 default (Mesa 17 and kernel 4.10) the game stutter and have slowdowns like there is no tomorrow. Only using Kernel 4.11RC6 and Mesa 17.1 with the gl_thread thingy enabled, things become tolerable. Also, I got a used GTX460 from a friend and I discovered that it get slightly better fps than my RX470 in this game using the Nvidia proprietary driver. Or the game is heavily optimized for the Nvidia driver, or radeonsi is not playing well with this game.

                  Comment


                  • #10
                    Originally posted by M@GOid View Post
                    Michael, you can test Tomb Raider with the maximum settings with radeonsi, no graphical glitches at all. Here, with a i7 3770k and a RX470 identical to yours, I get 58fps @ 1080p in the ultimate settings, and that includes the TressFX hair effect.
                    That Tomb Raider is weird thing, as last time Michael did benches just CPUs from different vendors behave much differently... Ryzen was much faster there vs Kaby which was several times slower than Ryzen

                    Maybe Michael should do extensive test of various drivers and various hardware combination all with that Tomb Raider, so we can conclude and mybe to point out on which part of the driver stack or maybe an app is the problem actually
                    Last edited by dungeon; 16 April 2017, 03:38 PM.

                    Comment

                    Working...
                    X