Announcement

Collapse
No announcement yet.

Zink OpenGL-Over-Vulkan Now Exposes GLSL 1.30

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

  • #11
    This is great news.
    Maybe it will take Zink to finally make OpenGL properly standardized.
    Also hardware manufacturers only need to make Vulkan drivers. That means cheaper hardware with less bugs.
    They should put more resources into this project.

    Comment


    • #12
      Depending on the hardware/driver, Zink can be faster than a native OpenGL implementation, when it will be mature. Right now Zink is nowhere near optimized, and that is natural. But the "can't be faster than a native implementation" thing some people are saying is wrong. Vulkan is so low level and can utilize multiple cores much better, that if really optimized it can do wonders. We are seeing this with DXVK even, some games perform better than on Windows!

      Comment


      • #13
        Originally posted by pal666 View Post
        everything can be faster than it is today. but zinc can't be faster than no zinc
        I dont think that is strictly true. It depends on the openGL implementation.

        In mesa, the Zink and the non-Zink paths could share the majority of their code via gallium with the next effect being little difference other than code paths that have been micromanaged by one implementation or the other for speed (though which once assumptions change over time may end up being slower that without the optimisation).

        Comment


        • #14
          Originally posted by zboszor View Post

          And if coupled with the copper from the Commodore Amiga, you get brass. ;-)
          You know what would be amazing.
          Easy raster bars for modern PCs. This way developers could debug timings synchronized to the display.

          Comment


          • #15
            Since the Pi 4 has a Vulkan implementation, I wonder if this will allow opengl 3.0 on it.... Especially since OpenGL 3.0 is not doable in that hardware, but Vulkan is.

            Comment


            • #16
              Originally posted by tildearrow View Post

              You know what would be amazing.
              Easy raster bars for modern PCs. This way developers could debug timings synchronized to the display.
              The ZX Spectrum Next is a modern PC (as in: Personal Computer), released early this year and it does have copper like the Amiga but it also has border effects with plain port I/O for debugging timings. Not something you see on an IBM compatible nowadays.

              Comment


              • #17
                Originally posted by zboszor View Post

                The ZX Spectrum Next is a modern PC (as in: Personal Computer), released early this year and it does have copper like the Amiga but it also has border effects with plain port I/O for debugging timings. Not something you see on an IBM compatible nowadays.
                Nice but of course can I run Linux on it?

                Comment


                • #18
                  Originally posted by tildearrow View Post

                  Nice but of course can I run Linux on it?
                  Believe it or not, there is an RPi0 accelerator inside running Linux that works in parallel with the main Z80 CPU implemented in FPGA.

                  Comment


                  • #19
                    Originally posted by You- View Post
                    I dont think that is strictly true. It depends on the openGL implementation.
                    no, it's not
                    Originally posted by You- View Post
                    In mesa, the Zink and the non-Zink paths could share the majority of their code via gallium
                    except after gallium non-zink paths drive hardware, while zink path produces sequence of vulkan calls which will do bunch of work and then do same thing as non-zink paths

                    Comment


                    • #20
                      Originally posted by TemplarGR View Post
                      But the "can't be faster than a native implementation" thing some people are saying is wrong. Vulkan is so low level and can utilize multiple cores much better, that if really optimized it can do wonders.
                      this is so much bullshit. when you already have opengl application, you can't make it faster by showing vulkan on top of it, because vulkan can't do anything which opengl driver can't do. vulkan can help only when there is no opengil in the first place - for vulkan programs. it's not "opengl driver is slow", it's "opengl is slow". and you already have all your slow opengl in zink, by adding vulkan on top you can't make it faster, only slower
                      Last edited by pal666; 19 June 2020, 01:49 AM.

                      Comment

                      Working...
                      X