Announcement

Collapse
No announcement yet.

AMD Linux Driver's LRU Bulk Moves Can Be A Big Help For Demanding Linux Games

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

  • aufkrawall
    replied
    It's btw. enabled again in 5.5 wip branch.

    Leave a comment:


  • ernstp
    replied
    Originally posted by ms178 View Post
    bridgman When speaking of a big help for performance, could you please forward this to the concerned team which should get in touch with the glibc developers (Red Hat Engineer Partner Management or contact Florian Weimer directly) as it seems that glibc does enable a fast path only for capable Intel CPUs because they need more insights into AMD's architectures: https://sourceware.org/bugzilla/show_bug.cgi?id=24979

    As this issue came up recently, here is another Mailing list entry of interest to justify the current behaviour: https://sourceware.org/ml/libc-alpha.../msg00155.html
    I pinged about it on Twitter, they know: https://twitter.com/zappor/status/1172377033181020165

    Leave a comment:


  • ernstp
    replied
    Originally posted by M@GOid View Post
    Unfortunately, Padoka's PPA is not updated for about 2 months. Anyone knows another PPA that is not oibaf's, since he updates daily and that gets old fast.
    I made my own PPAs if anyones interested.

    Release candidates:

    Release candidates and release branches of Linux graphics related things. The ppa is controlled by these scripts, so if you have an idea you can send me a pull request! You can also use the GitHub repo to report issue or feature requests. https://github.com/ernstp/ppa-mesarc If you want to support what I do you can use this to buy me a coffee! https://www.paypal.com/donate/?business=LMKYKSFTXGP2A


    Mesa ACO:

    NEWS: * Now with new NVK driver nouveau_experimental * Now with LLVM 16! The PPA originally started to enable the AMD ACO feature, but now it's simply bleeding edge Mesa. The PPA is controlled by these scripts, so if you have an idea you can send me a pull request! https://github.com/ernstp/ppa-mesarc I'm running the scripts manually though. Report any issues straight here: https://gitlab.freedesktop.org/mesa/mesa/-/issues

    Leave a comment:


  • [TV]
    replied
    Originally posted by M@GOid View Post
    Unfortunately, Padoka's PPA is not updated for about 2 months. Anyone knows another PPA that is not oibaf's, since he updates daily and that gets old fast.
    I recently switched from Padoka's PPA back to "Ubuntu-X" team PPA: team: https://launchpad.net/~ubuntu-x-swat...ubuntu/updates

    Leave a comment:


  • pal666
    replied
    Originally posted by AsuMagic View Post
    To help putting things into perspective, to hit 60FPS you need to be below 16.6ms/frame.
    they submit on separate thread, so it doesn't add to other per-frame work time if you have enough cpu cores. still, for 240 fps you need 4 ms and it was not possible without fix

    Leave a comment:


  • ms178
    replied
    bridgman When speaking of a big help for performance, could you please forward this to the concerned team which should get in touch with the glibc developers (Red Hat Engineer Partner Management or contact Florian Weimer directly) as it seems that glibc does enable a fast path only for capable Intel CPUs because they need more insights into AMD's architectures: https://sourceware.org/bugzilla/show_bug.cgi?id=24979

    As this issue came up recently, here is another Mailing list entry of interest to justify the current behaviour: https://sourceware.org/ml/libc-alpha.../msg00155.html
    Last edited by ms178; 18 September 2019, 09:33 AM.

    Leave a comment:


  • AsuMagic
    replied
    To help putting things into perspective, to hit 60FPS you need to be below 16.6ms/frame.
    I don't know how long the rest of the frametime in these tests was, but this may be a very significant improvement in this scenario.

    Leave a comment:


  • skeevy420
    replied
    Originally posted by Venemo View Post
    How do you enable the LRU bulk moves if you want to compile a kernel that has this feature?
    It patches into 5.2 cleanly....

    Code:
    diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
    index 48349e4f0701..fd3fbaa73fa3 100644
    --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
    +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
    @@ -603,14 +603,12 @@ void amdgpu_vm_move_to_lru_tail(struct amdgpu_device *adev,
         struct ttm_bo_global *glob = adev->mman.bdev.glob;
         struct amdgpu_vm_bo_base *bo_base;
     
    -#if 0
         if (vm->bulk_moveable) {
             spin_lock(&glob->lru_lock);
             ttm_bo_bulk_move_lru_tail(&vm->lru_bulk_move);
             spin_unlock(&glob->lru_lock);
             return;
         }
    -#endif
     
         memset(&vm->lru_bulk_move, 0, sizeof(vm->lru_bulk_move));
     
    --
    2.17.1

    Leave a comment:


  • M@GOid
    replied
    Unfortunately, Padoka's PPA is not updated for about 2 months. Anyone knows another PPA that is not oibaf's, since he updates daily and that gets old fast.

    Leave a comment:


  • Venemo
    replied
    How do you enable the LRU bulk moves if you want to compile a kernel that has this feature?

    Leave a comment:

Working...
X