Announcement

Collapse
No announcement yet.

Radeon Rays 4.0 Released - Adds Vulkan While Dropping OpenCL, No Longer Open-Source

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

  • Radeon Rays 4.0 Released - Adds Vulkan While Dropping OpenCL, No Longer Open-Source

    Phoronix: Radeon Rays 4.0 Released - Adds Vulkan While Dropping OpenCL, No Longer Open-Source

    Continuing with AMD's relaunch of GPUOpen and introducing new software releases all week, out this morning is Radeon Rays 4.0. It takes another step forward while taking a step back in terms of no longer being open-source...

    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
    The github project is still marked as MIT and the LICENSE file in their top level directory is a MIT license. So, technically, this is still open source, but probably will get corrected soon. I'd very much like to hear a reason for changing it to a closed software. A big disappointment, AMD :-/

    Comment


    • #3
      This isn't very surprising to me. Nvidia showed AMD that you can lock down your technology with proprietary tech and people still buy their products.

      Comment


      • #4
        I bet this has something to do with game consoles.

        Comment


        • #5
          Originally posted by profoundWHALE View Post
          This isn't very surprising to me. Nvidia showed AMD that you can lock down your technology with proprietary tech and people still buy their products.
          This. I don't like it, but AMD's actions of goodwill are never rewarded financially and people still buy Nvidia despite AMD's focus on open standards. At some point they have to benefit from their investments. Nvidia never properly opensource their software either. I suspect these closed libraries will be used extensively in next gen console ports and will offer a performance benefit to AMD gpus. Time for Nvidia to taste their "gameworks" medicine for once.

          Still, i would prefer it to be opensource. But since those libs are mainly used in games, it doesn't matter. Games are typically closed source anyway.

          Comment


          • #6
            Originally posted by kparal View Post
            The github project is still marked as MIT and the LICENSE file in their top level directory is a MIT license. So, technically, this is still open source, but probably will get corrected soon. I'd very much like to hear a reason for changing it to a closed software. A big disappointment, AMD :-/
            Disappointment could be overcome with excitment for Vulkan, as I would expect.

            Comment


            • #7
              Originally posted by TemplarGR View Post

              This. I don't like it, but AMD's actions of goodwill are never rewarded financially and people still buy Nvidia despite AMD's focus on open standards. At some point they have to benefit from their investments. Nvidia never properly opensource their software either. I suspect these closed libraries will be used extensively in next gen console ports and will offer a performance benefit to AMD gpus. Time for Nvidia to taste their "gameworks" medicine for once.

              Still, i would prefer it to be opensource. But since those libs are mainly used in games, it doesn't matter. Games are typically closed source anyway.
              People still think that NVIDIA is "good technology" and AMD "bad technology", even with the proven scheduled obsolescence of NVIDIA's products (see Vulkan and DirectX 12) and its lack support for open standards.

              In the Linux area, I still find people who say "I don't buy AMD graphics because their drivers are very bad". They don't know that the situation has changed a lot in the last four years.

              AMD has a very bad brand image. In CPU it improved a lot thanks to Ryzen, but in GPU still has a very bad brand image, especially because AMD has a lot of problems to compete in high end segment.

              Comment


              • #8
                It is extreme disapointment for me as wall as state of art for



                AMD To Support Coreboot On All Future CPUs
                Written by Michael Larabel in Coreboot on 9 May 2011


                https://www.phoronix.com/scan.php?pa...item&px=OTQyMQ

                Last edited by Peter Fodrek; 13 May 2020, 11:56 AM.

                Comment


                • #9
                  Originally posted by Peter Fodrek View Post
                  It is extreme disapointment for me as wall as state of art for



                  AMD To Support Coreboot On All Future CPUs
                  Written by Michael Larabel in Coreboot on 9 May 2011


                  https://www.phoronix.com/scan.php?pa...item&px=OTQyMQ
                  Coreboot only gets support from Purism and Google (for Chromebooks), and Google's implementation is horribly closed to user-installable OS's. That's pretty much the sum of it. If you look at ARM too, they're not even approaching it. It's either DTB and a janky custom bootloader, or SBBR/SBSA which is essentially what we have on x86 now: UEFI (+ACPI). Take a look at the new Raspberry Pi endeavours: it's backed by big tech providers to push UEFI into broader standardization on ARM. Coreboot doesn't really solve anything over and above what the EDK2 does - both are open source, but only to a certain extent, and both still require closed blobs from ARM, Intel and AMD to complete full device initialization.

                  Comment


                  • #10
                    Originally posted by ColdDistance View Post

                    People still think that NVIDIA is "good technology" and AMD "bad technology", even with the proven scheduled obsolescence of NVIDIA's products (see Vulkan and DirectX 12) and its lack support for open standards.

                    In the Linux area, I still find people who say "I don't buy AMD graphics because their drivers are very bad". They don't know that the situation has changed a lot in the last four years.

                    AMD has a very bad brand image. In CPU it improved a lot thanks to Ryzen, but in GPU still has a very bad brand image, especially because AMD has a lot of problems to compete in high end segment.
                    Why are you suprised it is a thing? Drivers for AMD GPUs on linux are heavy delayed (like day 1 support doesn't exist even on GIT kernel), practically you need to wait quite a lot, while on Windows there are still complains about them being unstable. Obsolescence is quite questionable because maybe AMD gains some performance, but you have to remember EOL AMD's GPUs reach much earlier, and Nvidia instead gives you technologies like DLSS 2.0 which suddenly flips benchmarks performance wise and makes AMD look like obsolence. Like there is Directx 12 ultimate... and much older Nvidia's GPUs are fully supporting it while AMD's GPU lack not just raytracing but other features too.

                    Honestly AMD looks good just in 3 cases, 1. you care about open source drivers, 2. you strictly care about price to performance while not caring at all about any of Nvidia's features there AMD just barerly wins, 3. you need VAAPI decoding.

                    Ironically Nvidia's 12 nm is more energy efficient with all those features then 7nm of AMD's...

                    Comment

                    Working...
                    X