Announcement

Collapse
No announcement yet.

Don't Keep Dreaming For Linux Video Improvements On Intel's Old Platforms

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

  • #31
    Originally posted by ferry View Post
    vaapi is only very important on slow CPU's like atom based. So for gma500 (FitPC) this is a problem that will never be solved. But for E3815 (NUC) vaapi is working fine for 1080p playback with MPV. Not good enough with VLC unfortunately, but that is because VLC feels the need to 'improve' video beyond repair. NUC 720p will work though.
    i don't agree with smplayer using mpv vaapi is good for decoding fullhd+ videos. ofc vlc sucks, the vaapi never works like it should

    Comment


    • #32
      Originally posted by chuckula
      Assuming you have a Nehalem-era desktop system with these graphics, I would recommend upgrading to even a very low-end discrete GPU that doesn't even require an external power connector. For example, a GTX-1050.
      Well, duh, but the main issue here is mobile. BTW, as Tomin pointed out, a GTX 1050 isn't what most people would call a low end card, but rather it's currently the lowest discrete card Nvidia is willing to make. A GTX 1050 would be the cheapest, most future-proof, new Nvidia card one could buy right now. However, for h.264 playback, a cheap, used RadeonHD 4350 or something like a Geforce 200-series would get the job done.

      if you have that attitude about Intel then I suggest never buying an AMD product since support will be dropped way way faster.
      False. Proprietary Catalyst support was dropped fairly quickly, but open source support for video playback is in good shape for AMD cards that are of the same vintage or older as Ironlake (thanks mainly to Christian Koenig and AMD for supporting for him). I would take any AMD notebook supporting UVD2 or later (e.g. RadeonHD 4250) for desktop use and video playback over an Ironlake any day.

      Note: I'm guessing poor Ironlake video accel support is due to poor hardware or programming model design, rather than Intel's attitude towards supporting older hardware in Linux or some kind of conspiracy to force people to buy newer hardware.

      Incidentally, this article should point out that you can certainly play back video on the hardware, you just don't get acceleration bells & whistles.
      ..and lowered fan noise and good battery life...

      I have a nearly 10 year old Core 2 notebook that's even older than these parts and it can playback non-crazy resolution video (including H.264 video) over HDMI just fine.
      Yeah, but I'm going to guess you have a higher end CPU. It's not really consolation to those with lower end CPU's that would have a reasonable expectation to get some help from the GPU when the GPU is supposedly capable.

      Comment


      • #33
        I have a core i5 520m based laptop with an ips screen that I use for web browsing and watching videos.
        It is still a good machine.

        With the right settings, it can handle 1080p 60fps h264 videos in smplayer/vlc and in firefox/chromium with software decoding.
        Even though it would have been great to have hardware decoding, especially since it was announced years ago.

        Comment


        • #34
          Originally posted by andre30correia View Post
          i don't agree with smplayer using mpv vaapi is good for decoding fullhd+ videos.
          smplayer and mpv is a *really* bad combination. smplayer does not use mpv properly, it pretends mpv is mplayer, which it most definitely is not. So who knows what crappy settings smplayer drives mpv with. Use mpv directly, with --hwdec=vaapi --vo=opengl or --vo=vaapi. If using opengl, a new enough mesa is recommended (version 12 I think), to get the significantly better VAAPI/EGL interop, the older VAAPI/GLX interop is less good. Or just use the vaapi output, it's the most efficient, but less feature rich than the opengl output.

          If you insist on a GUI, give baka-mplayer a try, Contrary to its name, it's actually a mpv frontend, and interfaces with mpv as is proper - through the libmpv client API.

          Comment


          • #35
            Originally posted by chuckula View Post
            Assuming you have a Nehalem-era desktop system with these graphics, I would recommend upgrading to even a very low-end discrete GPU that doesn't even require an external power connector. For example, a GTX-1050 if you want to be modern would provide all the video playback capabilities you want and a massive improvement in graphics performance too.

            As for mobile, if you are still rocking a Nehalem-era notebook that came out years before AMD even introduced the marketing term "APU", I would strongly suggest an upgrade. Kaby Lake is an extremely strong mobile platform as just one example.
            I don't think a GTX-1050 goes in an AGP port or any other port on motherboard with that old a CPU.

            Comment


            • #36
              Originally posted by c117152 View Post
              My 2010 GTX460(Fermi) TDR-locks* under Windows 10 and doesn't have linux framebuffer support with the property driver (while the Nouveau driver does).
              Seems to me your issue in Windows is more of a hardware or OS related problem, not a driver problem. Unless you've confirmed many people are experiencing the same issue, I don't think this is a driver/nvidia issue.
              Also, I don't understand who would ever need a GTX GPU with a framebuffer. That's like putting a fancy new engine in an old, heavy, rusty piece of junk car. If you really need framebuffer, just get the vesa driver.
              Intel is failing on tech, not marketing: A 2010 i7 is clocked roughly 1.5gigahertz below a 2016 i7 and has 10 times less the transistors (600million versus 6billion) but is, at most, 5 times slower compute.
              If they won't drop old hardware support, they'll end up making no sales at all.
              I agree. I never mentioned anything about marketing; you are basically confirming what I said but phrasing it differently.

              Comment


              • #37
                I have an Acer Extensa laptop that I bought in 2009 I guess. It has a 4500MHD and it still does what it suppose to do. I watch movies, tv shows on 720p because 1080p is pointless since the resolution is 1280x800; I occasionally play games on it; I often get my work done using this notebook. It's fine. Sure, it's nowhere near as powerful as my desktop, but it still is functional. - I expect an electronic to do what it should do without a major problem for 5-10 years, so of course I'm going to buy another one sooner or later. However reading such news does help me determine whether I should support Intel or not.

                Intel's incompetence has been a big problem, their lack of enthusiasm for supporting a product for more than 4 years is ridiculous. I don't have any problems with their GNU/Linux support, because I'm used to do my own dirty work when it comes to tweak my expectations; however their Windows 10 support, for instance, does tell us how can't-be-bothered they really are. Did you know they dropped their Graphics Properties in Windows 8? I used compatibility layer with Windows 7 drivers in order to get to those configuration options. On Windows 10, they didn't even work. They used an old driver version on Windows Update thing, whatever they call it. - So, I use Linux mainly, what's the big deal; big deal is that they don't even bother try to support their older hardware on a popular system for General Purpose. Sure, it works out of the box but software still is old technology.

                Who cares, right? I do. - Intel spends enormous amount of their budget on bullshit campaigns, millions of dollars to accomplish absolutely nothing.

                The point is, yes they have to sell newer products one way or the other. They have to force you to buy something newer, eventually. But they have the fucking budget to fund for at least 5 years of support for their older hardware. They just choose to focus on crap.

                Comment


                • #38
                  Originally posted by schmidtbag View Post
                  Seems to me your issue in Windows is more of a hardware or OS related problem, not a driver problem. Unless you've confirmed many people are experiencing the same issue, I don't think this is a driver/nvidia issue.
                  They are. It's very common. Google it.

                  Originally posted by schmidtbag View Post
                  Also, I don't understand who would ever need a GTX GPU with a framebuffer.
                  Plymouth and kms \ kmscon. Pretty annoying when you need to debug x or wayland from a 640x480 console.

                  Regardless, my point was that nVidia doesn't maintain drivers as well as they're being credited for. I had identical issues with multiple 2012 Kepler cards over various workstations & desktops so considering win10 was released in July15, that's around 3 years of product life which is what everyone else is doing.

                  Comment


                  • #39
                    Yea it's called Lifecycle Management. Exactly 5 years after the introduction of Ironlake, the support for Gen1 HD Graphics silently vanishes, compare 2015Q3 with 2015Q4 release notes (scroll to "Supported Hardware" at the bottom):
                    https://01.org/linuxgraphics/downloa...tack-release-0
                    https://01.org/linuxgraphics/downloa...-stack-release

                    The thing is, those Gen1 chips were massively popular due to successful marketing with the invention of "HD Graphics" and the cute i3/5/7 labeling. A LOT of chips were ordered. And they were quite good in fact - a meager Core i3 M was good enough for a Full HD media PC, which, unfortunately for Intel, is still adequate for a home theater PC, that is unless you need 4K, which a lot of us in 2017 still don't.

                    Intel simply cannot afford to support this chip, doing so could put a big dent in Gen7 sales figures.

                    Comment

                    Working...
                    X