Announcement

Collapse
No announcement yet.

Radeonsi is awesome, beats Catalyst!

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

  • #21
    Fortunately it seems someone else already did it: https://launchpad.net/~nzatkovich/+a...eediergraphics
    ## VGA ##
    AMD: X1950XTX, HD3870, HD5870
    Intel: GMA45, HD3000 (Core i5 2500K)

    Comment


    • #22
      Don't kill the goose that lays the golden eggs!

      Originally posted by oibaf View Post
      I have a Radeon X1600 and an Intel cards, both on netobooks, and I am not planning to upgrade to a newer GPU since I currently just don't need it. Unfortunately I have no more time to devote to the PPA than what I can currently do now.
      When people contribute code or packaging work in their spare time, please everyone don't be so demanding! Work done on-demand is the province of the pay software people, putting pressure on hobbyists is a good way to make people quit and keep their code for their own use only. Lots of us benefit from 0ibaf's work, I'd hate to see him driven away

      Comment


      • #23
        Originally posted by darkbasic View Post
        Fortunately it seems someone else already did it: https://launchpad.net/~nzatkovich/+a...eediergraphics
        This is 4 weeks old, it's like a lifetime :/

        Originally posted by Luke View Post
        When people contribute code or packaging work in their spare time, please everyone don't be so demanding! Work done on-demand is the province of the pay software people, putting pressure on hobbyists is a good way to make people quit and keep their code for their own use only. Lots of us benefit from 0ibaf's work, I'd hate to see him driven away
        Yep, Oibaf is doing awesome job, but I don't think asking a question about a feature does any harm. Ofc as long as it's not in demanding tone.
        Tbh until this thread I've thought Oibaf is maintaining that PPA on behalf of AMD, but since it's just a free time project, that's even more impressive

        Thanks Oibaf!
        Last edited by gutigen; 28 May 2014, 07:46 AM.

        Comment


        • #24
          Originally posted by darkbasic View Post
          Oibaf OpenGL 3.3 is really the least important benefit, llvm *GIT* is much MUCH more important for a thousand of reasons for radeonsi users so sitting down waiting for LLVM 3.4 will not help: we still need LLVM git. .
          Those are old, but even llvm git is not enough, you want spill fixes & stuff too do you .

          They just need to settle those llvm things fine first, so because of that situation i simply currently considered opensource radeonsi support as alpha support . Until at least clean gits started working fine, and all asics worked (hi Hawaii ) it is just there - alpha state .

          So right now i simply does not care anymore about all that radeonsi brokeness (i filled my bugs on fdo, so i am there if someone respond), but right now i am using fglrx until at least clean llvm git become usable . And for that matter maybe they will just abandon that llvm thing, you'll never knows .

          Comment


          • #25
            spill-fixes will get merged sooner or later but we will still need llvm master if you want "the latest and greatest" because every shader compiler optimization will go into master.
            ## VGA ##
            AMD: X1950XTX, HD3870, HD5870
            Intel: GMA45, HD3000 (Core i5 2500K)

            Comment


            • #26
              Originally posted by darkbasic View Post
              spill-fixes will get merged sooner or later but we will still need llvm master if you want "the latest and greatest" because every shader compiler optimization will go into master.
              I know, but oibaf trying to support most of it, even stable llvm ftbs right now because of arm, newest mesa as i see ftbs right now in egl_gallium... that ppa is PITA to support daily on fixed distro like trusty and we will be soon with xserver 1.16 and glamor from it so oohhoo . that is much work for him i think even without llvm considered it doing the work in his spare time ...

              So it much better to just build your own if you want to try latest and gratest .
              Last edited by dungeon; 29 May 2014, 03:01 PM.

              Comment


              • #27
                Originally posted by dungeon View Post
                I know, but oibaf trying to support most of it, even stable llvm ftbs right now because of arm, newest mesa as i see ftbs right now in egl_gallium... that ppa is PITA to support daily on fixed distro like trusty and we will be soon with xserver 1.16 and glamor from it so oohhoo . that is much work for him i think even without llvm considered it doing the work in his spare time ...

                So it much better to just build your own if you want to try latest and gratest .
                Any good, newbie friendly and up to date guides how to properly compile and install mesa git on Ubuntu?

                Comment


                • #28
                  Originally posted by darkbasic View Post
                  spill-fixes will get merged sooner or later but we will still need llvm master if you want "the latest and greatest" because every shader compiler optimization will go into master.
                  This guy is keeping it at least a little bit updated every few days or so: https://github.com/darkbasic/llvm/co...spill-fixes-v4

                  And can somebody tell me why debian developers put up with their overcomplicated packaging? I mean, look at archlinux's makepkg. Isn't it time to write a wrapper around dh_make and whatever is necessary to make it like that?

                  Comment


                  • #29
                    Originally posted by ChrisXY View Post
                    This guy is keeping it at least a little bit updated every few days or so: https://github.com/darkbasic/llvm/co...spill-fixes-v4
                    I know, I'm the guy you're talking about
                    ## VGA ##
                    AMD: X1950XTX, HD3870, HD5870
                    Intel: GMA45, HD3000 (Core i5 2500K)

                    Comment


                    • #30
                      Damn, I did it again. I really thought I was replying to someone else.

                      Comment

                      Working...
                      X