Announcement

Collapse
No announcement yet.

Civilization VI Now Available For Linux

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

  • #41
    Originally posted by randomsalad View Post
    Getting further than me atleast. On Debian Stretch with LLVM 4.0 & Mesa git it just crashes immediately with:
    Code:
    /home/user/.local/share/Steam/steamapps/common/Sid Meier's Civilization VI/./Civ6: Symbol `_ZTVN10__cxxabiv120__si_class_type_infoE' has different size in shared object, consider re-linking
    /home/user/.local/share/Steam/steamapps/common/Sid Meier's Civilization VI/./Civ6: Symbol `_ZTVN10__cxxabiv117__class_type_infoE' has different size in shared object, consider re-linking
    Segmentation fault
    run with same or { newer and backwards compatible } libstdc++ which was used to build(means gcc version).
    Last edited by pal666; 10 February 2017, 12:16 AM.

    Comment


    • #42
      Originally posted by pal666 View Post
      factually it doesn't
      because amd drivers are linux drivers, while nvidia drivers are windows drivers with linux shim
      Factually it does. My computer is a working example. I think you did not get the difference between KMS and Wayland. Nvidia driver does not do KMS using GBM, but it does run Wayland compositors like gnome but with a terrible approach - eglstreams instead of GBM. And that pisses developers off because they have to implement that in their compositors. But your statement that it does not is 100% wrong.

      Comment


      • #43
        Originally posted by randomsalad View Post

        Getting further than me atleast. On Debian Stretch with LLVM 4.0 & Mesa git it just crashes immediately with:
        Code:
        /home/user/.local/share/Steam/steamapps/common/Sid Meier's Civilization VI/./Civ6: Symbol `_ZTVN10__cxxabiv120__si_class_type_infoE' has different size in shared object, consider re-linking
        /home/user/.local/share/Steam/steamapps/common/Sid Meier's Civilization VI/./Civ6: Symbol `_ZTVN10__cxxabiv117__class_type_infoE' has different size in shared object, consider re-linking
        Segmentation fault
        Had the same problem on Arch. I then installed the "libstdc++5" and "lib32-libstdc++5" packages, put "LD_PRELOAD='./libcxxrt.so:/usr/$LIB/libstdc++.so.6' %command%" into the game launch properties and now it runs fine.

        Alternative to the last part, you can just launch Steam in Native mode instead of the Runtime.


        As for my experience, I'm running it on Medium details @ 1080p because, to be honest, the game doesn't look any different to me on Ultra, even though it runs worse. Maybe AMDGPU + Mesa just doesn't render the improved details for me to see, I don't know. Anyway I cranked up the MSAA on Medium and it looks fine.
        Last edited by ElderSnake; 10 February 2017, 02:26 AM.

        Comment


        • #44
          Originally posted by FireBurn View Post
          So I'm using the latest LLVM & Mesa from git, the intro video plays but then the game errors out with:

          Civ6: /var/tmp/portage/sys-devel/llvm-9999/work/llvm-9999/lib/Target/AMDGPU/SIMachineScheduler.cpp:1584: void llvm::SIScheduleBlockScheduler::blockScheduled(llv m::SIScheduleBlock*): Assertion `LiveRegsConsumers[RegP.first] == 0' failed.
          So it was R600_DEBUG=sisched that was causing the issue with LLVM from git, it worked fine with LLVM 4.0.0-rc1 - I've raised a bug

          Comment


          • #45
            My windows vs mesa benchmark (~3 times slower)

            http://www.gearsongallium.com/?p=3736

            Comment


            • #46
              Originally posted by Pontostroy View Post
              My windows vs mesa benchmark (~3 times slower)

              http://www.gearsongallium.com/?p=3736
              Uh, boy... eagerly waiting for Michael's unsupported slideshow results

              Comment


              • #47
                Originally posted by funfunctor View Post
                Long long story short;

                GL is old, big, fat, complex, stateful, single threaded, sometimes ambiguous and generally shit for today.
                I'd take a look at this: https://www.khronos.org/assets/uploa...-GDC-Mar14.pdf

                Comment


                • #48
                  Originally posted by FishPls View Post

                  You shouldn't be mad at Aspyr for the general worse state of the AMD drivers... It's not their job to fix the drivers too.
                  That would be true if it wasn't for the fact that, in the past, Aspyr has been very awesome at supporting AMD on launch. I guess they have their reasons, right?

                  Comment


                  • #49
                    Guess I'll have to buy a new computer to play this...

                    Comment


                    • #50
                      Originally posted by bug77 View Post
                      Imagine being a game developer, working on porting stuff to a platform having 1% market share. And then seeing messages like yours above. Would that wntice you into pouring even more money into that platform?
                      You don't like it, you don't buy it. But don't whine like the Universe in indebted to you.
                      Well, for me Aspyr statement is clear:
                      More of the latter stuff. We can't vouch for solid performance and a high quality bar, so it also means our player support won't be taking tickets for AMD-specific issues.
                      http://steamcommunity.com/app/289070...08489633235476
                      So they don't really care. Why should we?

                      Comment

                      Working...
                      X