Announcement

Collapse
No announcement yet.

More Details On The AMD GCN Back-End For GCC That's Expected To Merge For GCC 9

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

  • #11
    Originally posted by coder View Post
    This seems like a big investment in an ISA that should soon be superseded.
    how do you know how much different will be superseding isa?

    Comment


    • #12
      Originally posted by pal666 View Post
      how do you know how much different will be superseding isa?
      I obviously don't, but it seems to me that the ISA should probably change rather significantly, if a big jump in performance or efficiency is to be achieved.

      Comment


      • #13
        Originally posted by bridgman View Post
        The hard work has been getting dGPU support for KFD upstream and then re-aligning the userspace components around the upstream user/kernel interface. That work has pretty much been done, so now it becomes possible to start working with distros to have them include more of the userspace packages.
        Good to know, I wish you best

        Comment


        • #14
          Originally posted by coder View Post
          I obviously don't, but it seems to me that the ISA should probably change rather significantly, if a big jump in performance or efficiency is to be achieved.
          What? Why?

          Comment


          • #15
            Originally posted by coder View Post
            I obviously don't, but it seems to me that the ISA should probably change rather significantly, if a big jump in performance or efficiency is to be achieved.
            like i386 to i486?

            Comment


            • #16
              Originally posted by juno View Post
              What? Why?
              Look at how many generations of GCN there have been, and yet it seems to have hit a wall on instructions-per-clock level efficiency. In order to make the kind of gains needed to catch Nvidia, I'm just imagining they'd have to break instruction-level compatibility with GCN. Plus, there's not much GCN assembly language code or tools support floating around, so why not take advantage of the fact that you can break compatibility, especially if doing so unlocks significant benefits?

              Comment


              • #17
                Originally posted by pal666 View Post
                like i386 to i486?
                Maybe something like that.

                Comment

                Working...
                X