Announcement

Collapse
No announcement yet.

AMD HSA Offloading Support Dropped From The GCC Compiler

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

  • AMD HSA Offloading Support Dropped From The GCC Compiler

    Phoronix: AMD HSA Offloading Support Dropped From The GCC Compiler

    There didn't appear to be much usage ever out of the AMD HSA (Heterogeneous System Architecture) support within the GCC compiler and hadn't been maintained in a while so now has been wiped out of the GNU Compiler Collection...

    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
    I don't understand how are they able to support CUDA but not this.

    Comment


    • #3
      I'm starting to see a pattern if things being added to GCC and then quickly dropped

      Comment


      • #4
        rip HSA.
        shame, on the paper it looked great.

        Comment


        • #5
          Originally posted by tildearrow View Post
          I don't understand how are they able to support CUDA but not this.
          CUDA is actually used by downstream so there are maintainers for it.

          HSA on the other hand...

          Comment


          • #6
            Originally posted by starshipeleven View Post
            CUDA is actually used by downstream so there are maintainers for it.

            HSA on the other hand...
            HSA is used as well, but not so much with GCC, it would seem. AMD's own SDK is clang-based. I'm using it with tensorflow; works fine.

            Comment


            • #7
              Originally posted by oleid View Post

              HSA is used as well, but not so much with GCC, it would seem. AMD's own SDK is clang-based. I'm using it with tensorflow; works fine.
              Good to know it's not dead.

              Comment


              • #8
                i never realised it was ready for use...
                It came with the flaw of being paired with the wrong CPUs... and notebook manufacturers using single RAM builds.
                Was it so hard to write an HSA-capable program?

                Comment


                • #9
                  Originally posted by oleid View Post

                  HSA is used as well, but not so much with GCC, it would seem. AMD's own SDK is clang-based. I'm using it with tensorflow; works fine.
                  You have to distinguish between HSAIL (HSA Intermediate Language) and HSA run-time (and there are other HSA specifications too). HSAIL generation has been dropped from GCC, GCC continues to be able to utilize HSA run-time to offload to GCN devices. To my best knowledge LLVM never produced HSAIL and their "use of HSA" is also just using the run-time to run GCN ISA kernels.

                  Comment


                  • #10
                    I don't have a crystal ball but what about future HPC capable APUs?! Heterogenious Computing (and HSA) might make a comeback with them - it would be great to have some sort of compiler support for exploiting all of their capabilities and the HSA stack could be a base. But maybe with oneAPU Level Zero (or other through other means), there are better ways to start with having better chances of succeeding in the market?!

                    I haven't heard of AMD's software strategy for heterogenious computing for a long time (in fact their HSA summit in 2012 was the last big event around it). And with all of their recent design wins in the server market it would be great to hear what they do plan for the future to make better use of their capabilities from the software side (on Linux and Windows, please).

                    Comment

                    Working...
                    X