Announcement

Collapse
No announcement yet.

Mesa 17.0 Delayed To Allow For Ivy Bridge OpenGL 4.0

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

  • Mesa 17.0 Delayed To Allow For Ivy Bridge OpenGL 4.0

    Phoronix: Mesa 17.0 Delayed To Allow For Ivy Bridge OpenGL 4.0

    Mesa 17.0 (formerly known as Mesa 13.1) was supposed to enter its feature freeze last weekend, but that milestone and branching of the code-base didn't happen due to last minute feature work...

    http://www.phoronix.com/scan.php?pag...17.0-IVB-Delay

  • #2
    or not....
    https://lists.freedesktop.org/archiv...ry/141151.html

    although it'd be quite cool if it got included!

    Comment


    • #3
      Next will be r200 delaying release as hyperz is still not there, someone removed that during KMS switch and never returned feature back
      Last edited by dungeon; 01-17-2017, 04:37 PM.

      Comment


      • #4
        Why do this?

        IMHO dependable release schedule is much more important than including last minute features. Having deadlines helps distros in syncronizing their work.

        Comment


        • #5
          Originally posted by dungeon View Post
          Next will be r200 delaying release as hyperz is still not there, someone removed that during KMS switch and never returned feature back
          A 16 year old gpu vs the IGP in a CPU that is just as fast as the current released chips, hmm... more likely they would just drop the r200 support altogether.

          2012 - Ivy Bridge
          2013 - Haswell
          2014 - Broadwell - (no real desktop parts)
          2015 - Skylake
          2016 - Kaby Lake - (desktop chips released this past week)

          There has been effectively no speed increase over the past 5 years or so from Intel, so people are using those older chips a lot longer than previously.

          And no I don't have a Ivy Bridge system.
          Last edited by calc; 01-17-2017, 06:21 PM.

          Comment


          • #6
            And here I am sitting on a Sandybridge system feeling left behind.

            Comment


            • #7
              Originally posted by calc View Post
              A 16 year old gpu vs the IGP in a CPU that is just as fast as the current released chips, hmm... more likely they would just drop the r200 support altogether.
              Ha, ha, and some cards supported by r200 are faster than RPi3 GPU... so maybe we should drop vc4 first

              Originally posted by calc View Post
              There has been effectively no speed increase over the past 5 years or so from Intel, so people are using those older chips a lot longer than previously.
              Eh, there was speed increase about 3.5 times effectively See there fastest IB was 330 GFLOPS and current fastest is Iris Pro 580 with 1150 GFLOPS

              https://en.wikipedia.org/wiki/Intel_..._Iris_Graphics
              Last edited by dungeon; 01-17-2017, 07:26 PM.

              Comment


              • #8
                Originally posted by dungeon View Post

                Ha, ha, and some cards supported by r200 are faster than RPi3 GPU... so maybe we should drop vc4 first



                Eh, there was speed increase about 3.5 times effectively See there fastest IB was 330 GFLOPS and current fastest is Iris Pro 580 with 1150 GFLOPS

                https://en.wikipedia.org/wiki/Intel_..._Iris_Graphics
                I was referring to the CPU itself not the IGP. At least the IGP has improved some over the past 5+ years, heh. Very few people have access to Iris Pro chips though.

                Comparing just the high end desktop models you get the following:
                Intel i7-3770K HD Graphics 4000 GT2 332.8 GFLOPS
                Intel i7-4770K HD Graphics 4600 GT2 400.0 GFLOPS
                Intel i7-6700K HD Graphics 530 GT2 441.6 GFLOPS
                Intel i7-7700K HD Graphics 630 GT2 ~ 515 GFLOPS
                So somewhere around a 55% graphics GFLOPS speed increase in 5 years, not that great really.
                Last edited by calc; 01-18-2017, 12:45 AM.

                Comment


                • #9
                  Originally posted by dh04000 View Post
                  And here I am sitting on a Sandybridge system feeling left behind.
                  Does the hardware support more than OpenGL 3.3?

                  http://www.intel.com/content/www/us/...000005524.html -> OpenGL 3.1 on Windows
                  https://lists.freedesktop.org/archiv...er/067982.html -> OpenGL 3.3 on Linux

                  You should be happy
                  Last edited by -MacNuke-; 01-18-2017, 06:53 AM.

                  Comment


                  • #10
                    Originally posted by calc View Post
                    So somewhere around a 55% graphics GFLOPS speed increase in 5 years, not that great really.
                    He, he, it would be even less diff if number of EUs were the same It goes from 16, 20 and 24 EU there for these GT2... basically more EUs more perf. GT1>GT2>GT3>GT4 with 12>24>48>72 EUs - bigger GPUs more perf as always and same size hardest to improve on the same node
                    Last edited by dungeon; 01-18-2017, 10:16 AM.

                    Comment

                    Working...
                    X