Announcement

Collapse
No announcement yet.

AMD Radeon RX 480 On Linux

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

  • #41
    Originally posted by bridgman View Post

    Correct - with pretty much every new chip the first round of cards all follow a standard formula then (after more time to test & tweak) more heavily customized cards start to appear.



    Excellent. Hope you enjoy it !
    Thank you for the hard work!
    I very much like what you guys do, I even applied to the Markham office recently.
    I hope you guys are looking for junior engineers; low level Linux development would be my dream job.

    Comment


    • #42
      This is a really sad day for AMD. All the hype aside, this card isn't anywhere near the efficiency gains AMD promised, and does not perform good enough either. It's only selling point is performance is performance per dollar, which even the old GTX 970 is able to match.

      We knew this chip never could cope with GTX 1070, but I was hoping the performance gap would be more like 30% rather than 50%. Just as important, the performance per watt(from TechPowerup) is on par with Maxwell on 28nm. Even with Nvidia pushing their GTX 1070/1080 really high, they still achieve 70-80% better efficiency. At similar clocks Nvidia will be close to twice the efficiency of Polaris, meaning the efficiency gap has not decreased. This is pretty extreme considering Polaris achieves the efficiency of Maxwell on an inferior node. When Nvidia get's their GP106 and GP107 out AMD will be crushed, both in the low-end desktop and in the laptop segments.

      There is no "all open" driver support for this product. As every other GPU it requires proprietary firmware.

      Comment


      • #43
        For those complaining about perf/power values compared to GTX 970, the first part of the equation (perf) is usually in bad shape compared to Nvidia, this lowers the final ratio.

        In Windows where the AMD driver is more competitive with Nvidia you'll see better results.

        One test that is more apple-to-apples comparison are the Unigine ones, where the cards have similar results to their relative counterpart of AMD/Nvidia

        Comment


        • #44
          Originally posted by efikkan View Post
          There is no "all open" driver support for this product. As every other GPU it requires proprietary firmware.
          So you are so concerned about using a card with closed firmware that you're going to use an Nvidia card with completely closed drivers instead. Great logic there.

          Comment


          • #45
            Originally posted by Herem View Post

            So you are so concerned about using a card with closed firmware that you're going to use an Nvidia card with completely closed drivers instead. Great logic there.
            I think he/she was trying to say "As with every other GPU, it requires proprietary firmware."

            Comment


            • #46
              Originally posted by efikkan View Post
              There is no "all open" driver support for this product. As every other GPU it requires proprietary firmware.
              The *hardware* requires microcode (as does pretty much every modern chip) whether it be built into the die, loaded by the chip from flash, loaded by BIOS from flash, or loaded by the driver. The drivers are fully open source and would work the same way no matter how the microcode was loaded.
              Last edited by bridgman; 29 June 2016, 12:17 PM.
              Test signature

              Comment


              • #47
                Originally posted by Qaridarium
                so a person who is for freedom is a Dictator ?
                No, but they often have beards and/or mustaches
                Test signature

                Comment


                • #48
                  I really wish there were higher quality and more representative compute/OpenCL benchmarks. That's a pretty solid chunk of the Linux audience, I'd say and the current set of benchmarks are little and ill conceived IMHO.

                  - This GpuTest is pretty outdated OpenGL compute shader stuff, I doubt it's very representative for compute.
                  - Mixbench is, well, as its author describes: "The purpose of this benchmark tool is to evaluate performance bounds of GPUs on mixed operational intensity kernels. The executed kernel is customized on a range of different operational intensity values."
                  Without seeing a _curve_ of flops/byte vs performance, not even the flops/byte at which the tests are run, those numbers are meaningless. Even a simple peak flop rate (with the arithmetic intensity at which this is reached) would be more meaningful than just a single number.
                  - Last, the SHOC MD5 is frankly a niche benchmark that's heavy on integer and bitwise ops so it's special and that load is frankly not too interesting for most compute use-cases.

                  I'm baffled why aren't any of the standard/classic computational/HPC benchmarks included that are the de-facto tools to characterize the performance of a processor/compute architecture. These are in fact a the foundation of SHOC, included in the "Level 0" (device memory bw, max flops, etc.) and "Level 1" tests (FFT, scan, reduction, gemm, sort, triad, etc.): https://github.com/vetter/shoc/wiki

                  Michael, can you please take this feedback and consider it seriously. These GPUs are not anymore gamers-only anymore an I'm sure there are plenty of people in the technical computing/HPC community interested (I know at least a dozen ) in more thorough and representative compute benchmarks.
                  Last edited by pszilard; 29 June 2016, 12:27 PM.

                  Comment


                  • #49
                    Two questions came to my mind:
                    1. What about using a RX 480 on a mainboard that only offers PCI Express 2.0? Will it work and if it does how much impact on performance if any will the PCIe 2.0-interface have?
                    2. Is there any AMD APU available yet that profits from the same open source driver model?

                    Comment


                    • #50
                      Originally posted by Qaridarium

                      so a person who is for freedom is a Dictator ?

                      "This lifted a long standing damned and cursed illness of computer industries for to long we all suffer from closed source walled garden implementations Monopolizing the market to only benefit Nvidia und on the CPU side: Intel. But Now the Curse is broken we are no longer damned!
                      We are now free! Free to ~~~~choose~~~~ a free world over profit maximising anti social and anti consumer Dictatorship.
                      KILL THE CLOSED SOURCE OPENGL DRIVER WITH FIRE!!!!!!! KILL IT !!! END IT!!! NOW !!!! FOR THE GOOD OF ALL!!!!"
                      That is so close to his speeches you know, especially the end... And yes he heavily used the word freedom, ironically!

                      But hey, I hope you enjoy a little joke, or you should shave your mustache... for real ( to quote bridgman )
                      Last edited by Passso; 29 June 2016, 12:22 PM.

                      Comment

                      Working...
                      X