Announcement

Collapse
No announcement yet.

AMD's GPU Performance API Library 3.5 Drops ROCm/HSA Support

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

  • AMD's GPU Performance API Library 3.5 Drops ROCm/HSA Support

    Phoronix: AMD's GPU Performance API Library 3.5 Drops ROCm/HSA Support

    Released on Friday was a new version of AMD's GPU Performance API "GPUPerfAPI" project under the GPUOpen umbrella. This is the AMD library used by CodeXL, Radeon Compute Profiler, and others for tapping GPU performance counters and to help in analyzing performance/execution characteristics for Radeon hardware. But this new GPUPerfAPI 3.5 release comes with a rather surprising change...

    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
    System Requirements

    • An AMD Radeon GPU or APU based on Graphics IP version 8 and newer.
    • Windows: Radeon Software Adrenaline 2020 Edition 19.12.2 or later (Driver Packaging Version 19.50 or later).
    • Linux: Radeon Software for Linux Revision 19.50 or later.
    • Radeon GPUs or APUs based on Graphics IP version 6 and 7 are no longer supported by GPUPerfAPI. Please use an older version (3.3) with older hardware.
    • Windows 7, 8.1, and 10.
    • Ubuntu (16.04 and later) and CentOS/RHEL (7 and later) distributions.
    I guess a 19.50 drop soon?

    Comment


    • #3
      Maybe AMD will hop onto Intel's oneAPI solution soon. It wouldn't surprise me, their own HSA efforts did not find any meaningful traction and Intel has still its industry weight behind it and it is designed to be open to other vendors.

      Comment


      • #4
        I think that they are preparing for ROCm/OpenMP (gcc and llvm) and ROCm/OpenACC (gcc).

        Comment


        • #5
          Originally posted by Setif View Post
          I think that they are preparing for ROCm/OpenMP (gcc and llvm) and ROCm/OpenACC (gcc).
          ROCm/OpenMP/OpenACC are still needed even if they join the oneAPI/SYCL/DPC++ route as it is complementary, see: https://github.com/intel/intel-graph...iler/issues/62

          Comment


          • #6
            I expect this is because it was broken and didn't actually work. I reported this a while ago and didn't get any response from AMD.

            Comment


            • #7
              Originally posted by Madgemade View Post
              I expect this is because it was broken and didn't actually work. I reported this a while ago and didn't get any response from AMD.
              yeah. reporting problems to amd is exactly just as equal as not reporting anything at all.

              but of course it's really easy to talk to their oss devs on irc, so if you really want to get help it's always better to skip amd altogether and go straight to their oss irc feeds.
              Last edited by duby229; 15 December 2019, 07:39 AM.

              Comment


              • #8
                Originally posted by duby229 View Post
                yeah. reporting problems to amd is exactly just as equal as not reporting anything at all.
                Check out the ROCm Github issues. A year ago or so you used to see AMD devs responding to almost every issue, definitely every valid one. Now they are nowhere to be seen. There are some big issues with ROCm, such as false advertising for many broken features, devices which are not supported etc. Only silence from AMD, it's a shame really.

                Comment


                • #9
                  Please AMD fix OpenCL >= 1.2 implementation in Mesa and everyone is happy and buy more Radeon stuff. In the moment people who buying some navi based graphics cards (like me) instantly wish they had thrown their money on something that will provide them with Optix.

                  Comment


                  • #10
                    I still have no idea what AMD's Compute plan is, none of their open offerings are complete, wish they'd get their finger out and commit to one of them

                    Comment

                    Working...
                    X