Announcement

Collapse
No announcement yet.

AMD Radeon HD 8670D: Gallium3D vs. Catalyst

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

  • #31
    Originally posted by Michael View Post
    It would likely require Linux kernel changes first to expose the DMI as non-root, otherwise it isn't too useful to begin with since most users aren't running as root. Just checked the code and in conditions where there is accurate DMI and PTS is run as root it will read the operating frequency, DIMM configuration, and RAM manufacturer / model (though not many systems get the manufacturer/model) but yeah in terms of memory latencies and other details, that would be nice but isn't a big ticket item and no commercial customers have requested such support either, but any patches are welcome.
    I'm not fixing your program to make you more money. If you want to play the part of the hardware reviewer then put the effort in to your site that everyone else in your field does.

    Comment


    • #32
      Originally posted by Kivada View Post
      I'm not fixing your program to make you more money. If you want to play the part of the hardware reviewer then put the effort in to your site that everyone else in your field does.
      The Phoronix Test Suite is open-source and not licensed. I nor any commercial customers are the ones asking for RAM latency requests, thus not commercial related. For Phoronix readers, if anything more advanced system tables would just be served to non-AdBlock users if I were to add the support, all the others can just use the existing approaches: ask in the forums or Twitter.
      Michael Larabel
      http://www.michaellarabel.com/

      Comment


      • #33
        Originally posted by Michael View Post
        The Phoronix Test Suite is open-source and not licensed. I nor any commercial customers are the ones asking for RAM latency requests, thus not commercial related. For Phoronix readers, if anything more advanced system tables would just be served to non-AdBlock users if I were to add the support, all the others can just use the existing approaches: ask in the forums or Twitter.
        You may not sell the software but you derive income from the ads on the reviews you post on your own site, without PTS you'd be in a similar position to the Windows review sites. Given the quality of your reviews the mantra of "Patches welcome" isn't all that compelling.

        Meh, go ahead and try blocking adblock users, we will find a way around, we always do. In the end it'll be the same more trouble then it's worth to try and stop us from killing your internet kittens.

        As I've said before though, I have no aversion to paying for a site, provided it's content is actually worth the expense. Yet so far I've seen nothing posted here that has shown to be worth the expense and your apparent lack of ability to give a damn when your existing paid readers where seeing ads they where paying not to see since you got to charge them and they had to give you ad hits allowing you to make yet more money off of it.

        Comment


        • #34
          Originally posted by Michael View Post
          PTS doesn't break, it just sits there waiting for the process to be killed. Could add a kill command to the test profile but there isn't any sane way to do that unless just having it sit and wait for an arbitrary amount of time to kill the process, hoping the test completed in that amount of time.
          Please do it, or please just don't use PTS. But *please* bench Left 4 Dead 2.
          ## VGA ##
          AMD: X1950XTX, HD3870, HD5870
          Intel: GMA45, HD3000 (Core i5 2500K)

          Comment


          • #35
            STC2 textures - on or off? -

            With open source drivers in AMD GPUs you cannot play Steam games but ....
            I
            As STC2 and 3 texture compresion is copyrighted until 2015 at some countries and not at others, if you are at other countries or at your own risk, you can inistall

            STC2 package to make it work.

            I think that this test must include Gallium 3d and Gallium 3d + STC2, and we will read the improvement that this STC2 package does and what can be expected when in 2015 it will be copyright expired and as free software be able to be improved.

            Comment


            • #36
              not everything should be done in kernel

              Originally posted by Michael View Post
              It would likely require Linux kernel changes first to expose the DMI as non-root, otherwise it isn't too useful to begin with since most users aren't running as root. Just checked the code and in conditions where there is accurate DMI and PTS is run as root it will read the operating frequency, DIMM configuration, and RAM manufacturer / model (though not many systems get the manufacturer/model) but yeah in terms of memory latencies and other details, that would be nice but isn't a big ticket item and no commercial customers have requested such support either, but any patches are welcome.
              having a non-root access would classify as information leak security bug (at least dmidecode requires access to /dev/mem).
              you know you can have suid binary that would read and dump the data (or just run dmidecode) and let the rest of PTS work with that dump

              Comment


              • #37
                Originally posted by Michael View Post
                It's at 2133MHz as controlled from the BIOS/UEFI but the Catalyst driver as was reported to the result table reports 1066MHz.
                I doubt it states 2133 MHz in the Bios, most BIOS I've seen just stated it DDR3-2133 or just 2133, never with MHz.

                And not only Catalyst reports it as 1066 MHz, every ramtool in the world will report DDD3 2133 with 1066 MHz, because that's the bus speed of the RAM.
                http://en.wikipedia.org/wiki/DDR3_SDRAM

                Comment


                • #38
                  Originally posted by Kivada View Post
                  I'm not fixing your program to make you more money. If you want to play the part of the hardware reviewer then put the effort in to your site that everyone else in your field does.
                  Name one existing hardware review site that contributes to the linux kernel.

                  Comment


                  • #39
                    Honestly

                    Actually gallium in confront of amd drivers is good like "shit".

                    Sorry but bench feedback is shocking. Is it possible open programmers are not able to optimize our currently devices? Morover several devices looses owner supports in newest linux kernel.... and OS distros.

                    so problems are still great, men!!! the truth is that kernel and open cannot optimize hardware.
                    Last edited by Azrael5; 07-06-2013, 09:49 AM.

                    Comment


                    • #40
                      Originally posted by Azrael5 View Post
                      Actually gallium in confront of amd drivers is good like "shit".

                      Sorry but bench feedback is shocking. Is it possible open programmers are not able to optimize our currently devices? Morover several devices looses owner supports in newest linux kernel.... and OS distros.

                      so problems are still great, men!!! the truth is that kernel and open cannot optimize hardware.
                      Are you retard or something?

                      Comment


                      • #41
                        Originally posted by Drago View Post
                        Are you retard or something?
                        are you idiot or moron? see bench to understand better if you can.

                        open (mesa) drivers aren't able to optimize amd vga. so difficult to understand?
                        Last edited by Azrael5; 07-06-2013, 10:15 AM.

                        Comment


                        • #42
                          Originally posted by Azrael5 View Post
                          Sorry but bench feedback is shocking. Is it possible open programmers are not able to optimize our currently devices?
                          The open source drivers are running with default clocks, which are very low on APUs and on newer discrete GPUs. Dynamic power management code was recently pulled into the 3.11 kernel, allowing the clocks to run at full speed automatically (with correspondingly higher performance) but these benchmarks used the 3.9 kernel from Fedora 19. That's the primary reason results were slow.

                          Originally posted by Azrael5 View Post
                          Morover several devices looses owner supports in newest linux kernel.... and OS distros.
                          Don't understand this, can you provide more details ?

                          Originally posted by Azrael5 View Post
                          so problems are still great, men!!! the truth is that kernel and open cannot optimize hardware.
                          See above...

                          Comment


                          • #43
                            Originally posted by Azrael5 View Post
                            are you idiot or moron? see bench to understand better if you can.

                            open (mesa) drivers aren't able to optimize amd vga. so difficult to understand?
                            Personal tests with a HD6670 and a Core2Duo (1920x1080)

                            Heaven 3.0:
                            radeon r600sb 15 fps
                            catalyst 13 fps

                            Lost Coast:
                            radeon r600sb 80 fps
                            catalyst 77 fps

                            Comment


                            • #44
                              Originally posted by bridgman View Post
                              The open source drivers are running with default clocks, which are very low on APUs and on newer discrete GPUs. Dynamic power management code was recently pulled into the 3.11 kernel, allowing the clocks to run at full speed automatically (with correspondingly higher performance) but these benchmarks used the 3.9 kernel from Fedora 19. That's the primary reason results were slow.



                              Don't understand this, can you provide more details ?
                              older graphic cards cannot uses owner drivers. Ex. HD3850 has not amd driver, so user has to use mesa.

                              Problems of linux kernel is that it's too much rigid because of dependencies.

                              Ex. in microsoft system user can install older or newer driver according to his hardware structure... linux doesn't let user to choose how to manage its own devices. On XP for example, User is able to manage system drivers, devices drivers, control panel features... in simple way. That's the deprecable issue on linux system.


                              Other thought: sound and hardware acceleration on video streaming, and video in general: this capability is not fully supported on linux distros. If we consult for example dxdiag panel in XP, we can see that this features are present (API) and that drivers must interface with this api to use them. In linux system I don't know if these features are working during video and audio activity neither cpu is on charge or drivers enable these possibilities. So CPu becomes fundamentals too reach good performance in the whole system.


                              linux has to adopt methods to maximize hardware system to spread itself. once got this aim users will leave microsft rewarding a better system

                              Comment


                              • #45
                                Originally posted by Azrael5 View Post
                                older graphic cards cannot uses owner drivers. Ex. HD3850 has not amd driver, so user has to use mesa.

                                Problems of linux kernel is that it's too much rigid because of dependencies.

                                Ex. in microsoft system user can install older or newer driver according to his hardware structure... linux doesn't let user to choose how to manage its own devices. On XP for example, User is able to manage system drivers, devices drivers, control panel features... in simple way. That's the deprecable issue on linux system.


                                Other thought: sound and hardware acceleration on video streaming, and video in general: this capability is not fully supported on linux distros. If we consult for example dxdiag panel in XP, we can see that this features are present (API) and that drivers must interface with this api to use them. In linux system I don't know if these features are working during video and audio activity neither cpu is on charge or drivers enable these possibilities. So CPu becomes fundamentals too reach good performance in the whole system.


                                linux has to adopt methods to maximize hardware system to spread itself. once got this aim users will leave microsft rewarding a better system
                                Man go install XP, make one million restarts, Install hundreds of drivers for everything, eat spam and viruses and let us here on Linux front being deprecated!
                                I 'm tired of all those smart bright guys coming from the black box window$ archaic BSOD os!!
                                DxDiag? Feel happy to see AGP acceleration enabled there with a PCI express card and 3GB of RAM max!

                                Comment

                                Working...
                                X