Announcement

Collapse
No announcement yet.

AMD Radeon RX 5500 XT Linux Performance

Collapse
This topic is closed.
X
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #51
    Originally posted by bridgman View Post

    For clarity, I'm not saying it will not be do-able on APU - if anything APUs tend to pick up new display engines faster than dGPUs - but I can't talk about APU plans without getting into that "unreleased products" thing.

    Talking about future products is a marketing/business team responsibility; we're not going to be doing it from inside the engineering teams.
    Thanks for this clarification. As being conservative I at least had to assume it is not safe that a Navi APU would be 8k capable. But of cause I am aware of your situation and I won't demand something I would not provide if I would be in your position.
    So if a Navi APU is released before I had to buy a PC, this might be an option.

    Concerning worldwide energy concerns I think a Zen2+Navi with 35 to 65 TDP would be a great system - for most office and even for workstation usage.
    And if I haven't made wrong guesses Raven Ridge (and even a 35-65 W Icelake - which doesn't fully reach Raven Ridge performance) would be enough to work with 8k (desktop publishing, image manipulation etc.) and to play games in FullHD and even some in 4k.
    Thus 120 W is a lot - but if it is not used it consumes 7 W (I read it for Windows - and it is about the threshold in the graph `power consumption monitor' of this article) so it would not hurt that much even concerning my workflow.

    And I think that Nvidia is wrong that iGPUs/APUs are just there to be replaced by a dGPU - which of cause is neccessary to play high quality/graphically demanding games.
    So having the fastest CPU/GPU out first is OK - but I would wish for an APU at least half a year after the architecture (RDNA Navi in this case) was introduced and to be announced the day the 1st Navi product is released (which was done in the past).

    I think this should have advantages even visible for marketing/business team ... while announcing things longer (as was common in the past) could cause problems if competitors don't make such early announcements for the future and thus could react to be superior (concerning marketing at least).
    Maybe the users had more power or must be pleased in the past - at least business customers do have access to such roadmaps (even not under NDA, just my personal experience).

    Comment


    • #52
      Originally posted by atomsymbol View Post

      My experience with MSI Radeon RX 570 ARMOR in Linux:

      Zero RPM does not work in Linux. By default the fans are rotating at about 1000 RPM in idle and at 2000+ RPM under load in case of this GPU. 2000+ RPM is very audible (unless wearing headphones). If the user wants zero RPM with this GPU, he/she will have to find (or develop) a small app watching the GPUs temperature and adjusting RPM via files in the directory "/sys/class/drm/card?/device/hwmon/hwmon?". Setting fan rotation below 250 RPM will physically stop the fans (zero RPM).

      The idle power consumption is about 8 Watts with amdgpu.ppfeaturemask=0xffffffff on kernel command-line, and about twice as much Watts without the ppfeaturemask. However, using ppfeaturemask might result in heavy display flickering when multiple displays are connected to the GPU, which can be solved by manually (or via a small script) fixing the GPU's memory clock when multiple displays are connected and disconnected.

      The RX 570 GPU halts the system about 20% of the time when the system is resumed from suspend-to-memory, which basically means that you won't be able to suspend your machine ever if you want to avoid the risk of the machine failing to resume.
      This is very bad and one of the main things that make me feel disappointed about AMD.
      I assume the case with Navi it's the same.
      I wonder if AMD will ever release a control panel where we can change these things like human beings...

      Thank you very much for sharing this!

      Comment


      • #53
        Originally posted by dimko View Post

        just to let you know, HDMI does not support freesync. Not on Linux: https://www.amd.com/en/support/kb/faq/gpu-754
        • FreeSync via HDMI is not supported for this feature. Only DP FreeSync displays will work.

        Better buy display port cables
        I know that HDMI does not support Freesync.
        I am more interested into High freame rate at 4K, anything over 60 Hz, let's say 120 Hz.
        And DisplayPort is good, but it's mostly for monitors only, I haven't seen yet a tv with it, just with HDMI

        Comment


        • #54
          Originally posted by F.Ultra View Post

          Why would they spend the money on HDMI 2.1 for a 1920x1080 card? The only difference between 2.0b and 2.1 is the added bandwidth for new 4K and 10K resolutions.
          Who says it's a 1920x1080 card ?
          Who says you can't hook it up to 4K monitor or TV and watch a movie on it ?
          I honestly don't buy GPUs just for games.
          I want to be able to watch movies with them with the latest quality features turned on like HDR (High dynamic range) and HFR (High frame rate).
          Also I want in the future to buy a [email protected] or more when they will become available.
          Why should I need a new GPU just for this just because they chose the cheaper path of putting a very old version of the HDMI instead of a somwhat newer one?

          Comment


          • #55
            Originally posted by atomsymbol View Post

            RTX 2060 has 10.8e9 transistors at 12nm in 445mm². RX 5700 has 10.3e9 transistors at 7nm in 251mm². But the expression sqrt(445)/sqrt(251)*7nm yields 9.32nm.
            Why are you dividing the square root of a 12nm RTX 2060 (445mm²) by the square root of the 7nm RX 5700 (251mm²) then multiplying by a 7nm process node to give you 9.32nm, that just doesn't make any sense and appears to be mathematical gobbledygook. I think you need to go and visit a basic formula for fitting carpet tiles and see where you have gone wrong and why.
            Last edited by Slartifartblast; 14 December 2019, 08:59 AM.

            Comment


            • #56
              Originally posted by Slartifartblast View Post
              Why are you dividing the square root of a 12nm RTX 2060 (445mm²) by the square root of the 7nm RX 5700 (251mm²) then multiplying by a 7nm process node to give you 9.32nm, that just doesn't make any sense and appears to be mathematical gobbledygook. I think you need to go and visit a basic formula for fitting carpet tiles and see where you have gone wrong and why.
              It works like this: You (or somebody else) will have to suggest a better math expression with improved accuracy than the expression submitted by me, or (xor) you will have to accept my computation as the best one so far in this particular discussion.

              Comment


              • #57
                Originally posted by atomsymbol View Post

                It works like this: You (or somebody else) will have to suggest a better math expression with improved accuracy than the expression submitted by me, or (xor) you will have to accept my computation as the best one so far in this particular discussion.
                My suggestion is you return to kindergarten maths class, I just can't be arsed doing it for you. It's not just inaccurate it's complete crap. 0/10, try harder.

                Have a nice day

                Comment


                • #58
                  Originally posted by Slartifartblast View Post
                  My suggestion is you return to kindergarten maths class, I just can't be arsed doing it for you. It's not just inaccurate it's complete crap. 0/10, try harder.

                  Have a nice day
                  This is called trolling.

                  Do you have a university degree?

                  Comment


                  • #59
                    Originally posted by atomsymbol View Post

                    This is called trolling.

                    Do you have a university degree?
                    Yes thank you, joint honours in Chemistry and Biochemistry. Now there will be no more correspondence on my part as this has degenerated into a rather tedious melodrama that I have no intention of participating in any further.

                    Good day and goodbye.

                    Comment


                    • #60
                      Originally posted by Slartifartblast View Post
                      Yes thank you, joint honours in Chemistry and Biochemistry. Now there will be no more correspondence on my part as this has degenerated into a rather tedious melodrama that I have no intention of participating in any further.
                      The melodrama was in most part your own doing.

                      If you have joint honours in Chemistry and Biochemistry (which I am unable to verify), then I don't understand why you didn't post a math expression that is more accurate than mine - but instead resorted to terminology such as "mathematical gobbledygook", "arsed", "complete crap" or "tedious melodrama".

                      I would like to see an expression that is more accurate. So please post it here.
                      Last edited by atomsymbol; 14 December 2019, 02:21 PM.

                      Comment

                      Working...
                      X