Announcement

Collapse
No announcement yet.

Zink's Mike Blumenkrantz Tackling Mesa WSI Inefficiencies

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

  • Zink's Mike Blumenkrantz Tackling Mesa WSI Inefficiencies

    Phoronix: Zink's Mike Blumenkrantz Tackling Mesa WSI Inefficiencies

    Earlier this month I ran some fresh benchmarks of Zink OpenGL-on-Vulkan against RadeonSI. While Zink in general is already quite speedy and in good shape for most workloads, those tests uncovered some troubled spots and Zink lead developer Mike Blumenkrantz of Valve has been diving into some of those issues with fixes. Another merge request is pending to deal with inefficiencies in the Mesa Vulkan windowing system integration (WSI) code...

    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

    What's unethical is you won't let me in through the window.
    Making me walk around the house is inefficient.

    Comment


    • #3
      Before more people fixate on "unethical":
      Unethical Gains
      These changes are still a bit experimental. I’ve tested them locally and gotten good results without issues in all scenarios, but that doesn’t mean they’re “correct”. VK WSI is used by a lot of applications in a lot of ways, so it’s possible there’s scenarios I haven’t considered.

      It's only a reference to whether or not his fix is the correct solution to the problem, or if it will blow some other program up. In other words: fix or oops?

      Comment


      • #4
        I don't know how much we really need to care about this. This seems to be a CPU-bound test, and I don't see how Zink is ever supposed to do better in such cases.

        Comment


        • #5
          I think that next mesa release will be kind of “zink 1.0” stage.
          Great to see that phoronix benchmarks help fixings edge cases bottlenecks.

          Comment


          • #6
            The fact that explicit sync is one reason for zink being slower should blow a couple of minds here.

            Comment


            • #7
              Originally posted by MrCooper View Post
              The fact that explicit sync is one reason for zink being slower should blow a couple of minds here.
              Interesting, could You elaborate on that, please?

              Comment


              • #8
                Originally posted by Linuxxx View Post

                Interesting, could You elaborate on that, please?
                Read the blog post. He says that implicit sync allows other mesa drivers to cheat a bit in ways he can't since Vulkan is explicit sync.

                Comment

                Working...
                X