Announcement

Collapse
No announcement yet.

AMD Publishes Video To Explain The Radeon Open Compute Stack (ROCm)

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

  • AMD Publishes Video To Explain The Radeon Open Compute Stack (ROCm)

    Phoronix: AMD Publishes Video To Explain The Radeon Open Compute Stack (ROCm)

    AMD has shared with us that they have published a video to explain in basic terms for the audience at large "What is ROCm?", a.k.a. the Radeon Open Compute stack...

    http://www.phoronix.com/scan.php?pag...o-What-Is-ROCm

  • #2
    Too bad blender crashes with it on my GPU (Radeon VII) ... this is what I get as soon as I switch cycles to GPU Compute:
    mesa: CommandLine Error: Option 'help-list' registered more than once!
    LLVM ERROR: inconsistency in registered CommandLine options

    And yes, there is a bug reported for this since ages.

    Comment


    • #3
      ROCm nice idea, good marketing, horrible implementation. Sadly on the GPU Compute side CUDA is literally the only option.

      I'm guessing atm Radeon Group don't see much chance to eat market share here from nVidia so they basically kinda glue togheter ROCm to have something marketing wise but not much else.

      I'm really hoping the compute side of NIR get finalized by karol group in the near future since their conception seem to have way better chance to succeed and it will give the option to support CUDA as well as an state tracker

      Comment


      • #4
        ROCm is a terrible name, marketing-wise. It just doesn't roll.

        Comment


        • #5
          Originally posted by yariv View Post
          ROCm is a terrible name, marketing-wise. It just doesn't roll.
          Yes, I agree. Something like ROCA (Radeon Open Computing Architecture) would be better.

          Comment


          • #6
            Originally posted by xception View Post
            Too bad blender crashes with it on my GPU (Radeon VII) ... this is what I get as soon as I switch cycles to GPU Compute:
            mesa: CommandLine Error: Option 'help-list' registered more than once!
            LLVM ERROR: inconsistency in registered CommandLine options

            And yes, there is a bug reported for this since ages.
            Yes, sadly ROCm isn't usable with mesa. Maybe ROCm is god with servers, but unusable on workstations/desktops.

            Comment


            • #7
              The only interesting part of ROCm is their OpenCL version. The current iteration is slower on applications than the amdpgu-pro's especially on both desktop and mobile APU.

              Comment


              • #8
                Originally posted by PapadakosPanagiotis View Post

                Yes, I agree. Something like ROCA (Radeon Open Computing Architecture) would be better.
                Or SHIT (Standard Heterogeneous Idle-free Tasks).

                Comment


                • #9
                  Any sign of them supporting all their cards with this? I think I've missed out on 4 generations of "compute support" with AMD, it worked partially (though badly) with Clover on my 6770M, my M395X is just about EOL without ever having compute support, the Ryzen 3400G doesn't work with it and I'm pretty sure the 5700M in the latest laptops is NAVI and not supported either

                  Comment


                  • #10
                    bridgman mentioned that there will be Navi support for ROCm. But why bother, it's unusable for consumer/enduser purposes.
                    No idea how well the proprietary driver works with Navi these days. At least it should be fully compatible with upstream and is easy to install on Arch (not AMD's merit though):
                    https://aur.archlinux.org/packages/opencl-amd/

                    Comment

                    Working...
                    X