Announcement

Collapse
No announcement yet.

4K AMD/NVIDIA High-End GPU Comparison On SteamOS Linux

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

  • #31
    Originally posted by Vassili View Post

    If I might give a recommendation, make these graphs:
    http://openbenchmarking.org/embed.ph...ha=2dc7523&p=2

    But then instead of power consumption, with minimal FPS, 50% average FPS, average FPS and maximal FPS.
    Yeah, I can picture it easily. I'm definitely not a programmer in any way, but I've looked alot at how pts profiles work and I'm sure that's doable.

    Michael, does this already exist? Is there a starting point you would recommend to someone if they wanted to try and implement this?

    EDIT: Ok, so it seems the game has to be able to report the right information and that seems to be rare in a lot of games. Still a few games could be made to output a graph in this manner I think.

    EDIT: Also it seems if the output format is different, then each game will need it's own profile for this, is that right? I'm just trying to figure out how to make that work. If it's as simple as I think it is, then I think I can make the profiles for it.
    Last edited by duby229; 26 October 2015, 04:53 PM.

    Comment


    • #32
      Master5000 http://arstechnica.com/gaming/2015/0...nt-for-nvidia/
      Yeah, they're totally sucking

      Comment


      • #33
        Originally posted by duby229 View Post

        AMD has already said that gamers should using the OSS drivers [...]
        Ok, not a long time ago Catalyst was said to be for gamers. Now if OSS is for gamers and desktop, then, for whom the Catalyst is

        Comment


        • #34
          This is the kind of article that drew me into Phoronix in the beginning, I bet overclock.net would love these kinds of results as they chase after performance like no one else.

          I myself am running 4k on a 4k 43" Vizio M43-C1 and I can say the technology is almost there, however my GTX 970 outputs HDMI 2.0 on the HDMI-1 port with some trouble as turning the screen off will require I reboot as I get "No Signal" unless dropping to TTY2 and back.

          If anyone here gets a 4k, make sure it can run @ 60hz - looking back I might have got a LG or Samsung, Vizio firmware seems trashy & updates feel like they break the 60hz @ 4k for me as I had to replace this TV before.

          Looking forward to more benchmarks, maybe my next build will be AMD & I'll swap out my 970 into my PC.

          Comment


          • #35
            Originally posted by Vassili View Post
            But then instead of power consumption, with minimal FPS, 50% average FPS, average FPS and maximal FPS.
            It's entirely limited based upon what the game engine exposes... For a majority of games, not much info beyond the average FPS. In others, I can show frame time graphs and more or at least min and max too.

            The "universal" approaches to that like libframetime haven't been exactly universal and have got differing results, hence relying upon the game engine.
            Michael Larabel
            https://www.michaellarabel.com/

            Comment


            • #36
              Originally posted by duby229 View Post

              Yeah, I can picture it easily. I'm definitely not a programmer in any way, but I've looked alot at how pts profiles work and I'm sure that's doable.

              Michael, does this already exist? Is there a starting point you would recommend to someone if they wanted to try and implement this?

              EDIT: Ok, so it seems the game has to be able to report the right information and that seems to be rare in a lot of games. Still a few games could be made to output a graph in this manner I think.

              EDIT: Also it seems if the output format is different, then each game will need it's own profile for this, is that right? I'm just trying to figure out how to make that work. If it's as simple as I think it is, then I think I can make the profiles for it.
              It's entirely limited based upon what the game engine exposes... For a majority of games, not much info beyond the average FPS. In others, I can show frame time graphs and more or at least min and max too.

              The "universal" approaches to that like libframetime haven't been exactly universal and have got differing results, hence relying upon the game engine.

              You can see openarean, xonotic, and a few others for different result parsing options based upon what's available.
              Michael Larabel
              https://www.michaellarabel.com/

              Comment


              • #37
                Hey, Valve just pushed a huge update! New drivers should change the results quite a bit! Interested to see if they actually do. http://steamcommunity.com/groups/ste...3938443060773/

                Comment


                • #38
                  Yep, it's good valve are getting things working for their OS. Be interesting to see what they do with Kernel 4.3 and 4.4

                  Comment


                  • #39
                    Hi Michael,

                    thanks for the 4K review of Tuxian gaming, from ATI/AMD vs nVida.

                    Having used both a GTX680 & now a R390x I can say that the ATI/AMD drivers are pants & I would not dream of gaming in 4K on Tux. Pity really, but ATI/AMD CBF doing drivers properly....

                    In Windows 10, this card is all gravy & kudos to ATI/AMD. Maybe one day they will get there shit together on Linux, but I have been saying that for over 10 years and it is frankly boring waiting or hearing excuses. Though I really must hat tip the idea, that there are no resourcing issues with the dev of Linux drivers. Pure insanity.

                    Keep up the good work Michael & keep the Green & Red teams honest.

                    GreekGeek :-)

                    Comment


                    • #40
                      Originally posted by faldżip View Post
                      Most of the games require high single-core performance so there is a chance that the dual core Pentium G3258 overclocked to something like 4.3-4.5GHz performs better then this Haswell-E.
                      You are right, G3258 beats it Haswell-E in most games.
                      Until games will really be massively multi-threaded it will remain the best gaming CPU for money

                      (I OC mine to 4.Ghz without changing anything else than the multiplicator, but I am sure it can go far higher)

                      Comment

                      Working...
                      X