Announcement

Collapse
No announcement yet.

Microsoft & AMD Release C++ AMP Compiler With Linux Support

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

  • Microsoft & AMD Release C++ AMP Compiler With Linux Support

    Phoronix: Microsoft & AMD Release C++ AMP Compiler With Linux Support

    AMD and Microsoft jointly released C++ AMP version 1.2 compiler that supports Linux alongside Windows. Ubuntu is the officially supported distribution at this time...

    http://www.phoronix.com/vr.php?view=MTc3Nzg

  • #2
    mmm, seems a bad idea for some reason, i'll wait for opencl in mesa better.

    Microsoft and AMD software sounds like a very bad idea waiting to kill kittens

    Comment


    • #3
      Originally posted by jrch2k8 View Post
      mmm, seems a bad idea for some reason, i'll wait for opencl in mesa better.

      Microsoft and AMD software sounds like a very bad idea waiting to kill kittens
      In fact, I have not seen anywhere that this compiler is related to Microsoft or AMD: it seems to come from multicoreware only.
      And personnally I think that it would be great, it seems to be higher-level than OpenCL.

      Comment


      • #4
        Originally posted by Ziple View Post
        In fact, I have not seen anywhere that this compiler is related to Microsoft or AMD: it seems to come from multicoreware only.
        And personnally I think that it would be great, it seems to be higher-level than OpenCL.
        I haven't been studying this AMP much, but isn't it pretty much like OpenACC which NVIDIA supports?

        Comment


        • #5
          yes

          Originally posted by caligula View Post
          I haven't been studying this AMP much, but isn't it pretty much like OpenACC which NVIDIA supports?

          Comment


          • #6
            Originally posted by caligula View Post
            I haven't been studying this AMP much, but isn't it pretty much like OpenACC which NVIDIA supports?
            Not at all. More like OpenCL, except based on C++11, with inline kernel definitions (as lambdas). It's somewhat similar to SyCL, actually.

            Comment


            • #7
              IMO it would be much better to see working OpenCL in MESA. This AMP stuff sounds really overcomplicated. And Microsoft around? Hmm, no, thanks.

              Comment


              • #8
                Originally posted by 0xBADCODE View Post
                IMO it would be much better to see working OpenCL in MESA. This AMP stuff sounds really overcomplicated. And Microsoft around? Hmm, no, thanks.
                OpenCL in Mesa is shaping up slowly but steadily. I think that in one year the OpenCL support for Radeon cards will be really correct (OpenCL 1.2).
                There are quite capable people at Microsoft you know. And quite often their APIs are a model: OpenGL NG will be just like D3D12, Gallium has been designed with the D3D11 concepts etc. Like always, marketing decisions are to blame, not engineers.

                Comment


                • #9
                  Originally posted by Ziple View Post
                  OpenCL in Mesa is shaping up slowly but steadily. I think that in one year the OpenCL support for Radeon cards will be really correct (OpenCL 1.2).
                  There are quite capable people at Microsoft you know. And quite often their APIs are a model: OpenGL NG will be just like D3D12, Gallium has been designed with the D3D11 concepts etc. Like always, marketing decisions are to blame, not engineers.
                  Psst. This is an open source enthusiast site. Therefore Microsoft can't ever do anything right.

                  In all seriousness, Microsoft's engineering practices have come a long way since Windows 95 and I would honestly say they're pretty good. If I distrust a piece of software, operating system, programming tool, or programming language that comes out of Redmond it's only because of the history of the company's business practices and their tendency to attack open source with patent lawsuits and FUD. The tool itself is often quite good.

                  Comment

                  Working...
                  X