Announcement

Collapse
No announcement yet.

Intel Skylake HD Graphics 530 Performance On Linux

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

  • Intel Skylake HD Graphics 530 Performance On Linux

    Phoronix: Intel Skylake HD Graphics 530 Performance On Linux

    Intel's Core i5 6600K and i7 6700K processors released earlier this month feature HD Graphics 530 as the first Skylake graphics processor. Given that Intel's Open-Source Technology Center has been working on open-source Linux graphics driver support for over a year for Skylake, I've been quite excited to see how the Linux performance compares for Haswell and Broadwell as well as AMD's APUs on Linux. In this article is the first of these OpenGL benchmarks comparing the Core i5 6600K to other offerings from Intel and AMD.

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Looking at just the integrated graphics, 250 bucks is a bit too expensive. Makes you think though, if the gpu on the i5 has this level of performance then the i3 will be a pain in the ass for AMD apus

    Comment


    • #3
      strange - again p-state powersave for the intel - why? is this fedoras default?
      and some APU results with fglrx and fast RAM would be nice.

      Nevertheless - thanks michael for all the hard work!

      Comment


      • #4
        Originally posted by tomtomme View Post
        strange - again p-state powersave for the intel - why? is this fedoras default?
        and some APU results with fglrx and fast RAM would be nice.

        Nevertheless - thanks michael for all the hard work!
        Yes, default driver/governor used.

        The RAM in each of the systems matched the potential for the CPU/APU, thus for the Kaveri setups was DDR3-2133.

        Fglrx results probably when I run some Ubuntu-based tests, hadn't tried installing Catalyst lately on F22 but attempts in its development cycle earlier were yielding issues.
        Michael Larabel
        https://www.michaellarabel.com/

        Comment


        • #5
          Originally posted by tomtomme View Post
          strange - again p-state powersave for the intel - why?
          You do realize intel_pstate powersave works like acpi_cpufreq ondemand, right? intel_pstate powersave does *not* keep the frequency at its lowest, it scales dynamically as needed. It is exactly what *should* be used on modern Intel CPUs.

          Comment


          • #6
            Where's that hypocritical guy who always screams about any AMD vs. Nvidia review where you use Catalyst & Nvidia's drivers instead of AMDs supposedly "amazing" open source drivers?

            I want to see his post where he screams at you for using AMD's open source drivers instead of Catalyst in this review.

            Comment


            • #7
              Originally posted by chuckula View Post
              Where's that hypocritical guy who always screams about any AMD vs. Nvidia review where you use Catalyst & Nvidia's drivers instead of AMDs supposedly "amazing" open source drivers?

              I want to see his post where he screams at you for using AMD's open source drivers instead of Catalyst in this review.
              well to be honest i don't see the problem with radeonsi here or catalyst in that sense, this apu's are really weak in the CPU side and is quite hard for those cores to keep the GPU feed, especially if there is CPU activity in another thread(this is shown in many windows benchmarks too where hard physics can plummet the FPS specially in A6 and A8 APUs), on the other hand intel skylake for 100$+ bucks more offer a beast CPU more than powerful enough to keep that IGP super feed to death regardless of any other threads work and we all know intel drivers are very good(not perfect but the same is true in windows).

              So my point here is that is blatantly obvious that a full scale 300$ skylake CPU with a decent IGP is gonna crush to death any AMD APU offer any day regardless the operative system/driver, even when in paper the r7 should be faster as an GPU. So if you could put the APU GPU inside the Intel CPU probably things will go lot better for the R7 GPU

              mmm anadtech shows in DX some wins for the APU but only for few frames, broadwell still beat the living shit out of both tho but in CPU is a total massacre for the APU
              Last edited by jrch2k8; 18 August 2015, 06:56 PM.

              Comment


              • #8
                Well, I'm certain at least some existing games will get updates for DX12 or Vulkan support. At which point I do expect AMD APU's will look much better than now.

                Comment


                • #9
                  Originally posted by chuckula View Post
                  Where's that hypocritical guy who always screams about any AMD vs. Nvidia review where you use Catalyst & Nvidia's drivers instead of AMDs supposedly "amazing" open source drivers?

                  I want to see his post where he screams at you for using AMD's open source drivers instead of Catalyst in this review.
                  You sound just like you're making things up based on your own bias. AMD's OSS drivers are amazing, but they aren't magical.

                  Comment


                  • #10
                    Originally posted by Gusar View Post
                    You do realize intel_pstate powersave works like acpi_cpufreq ondemand, right? intel_pstate powersave does *not* keep the frequency at its lowest, it scales dynamically as needed. It is exactly what *should* be used on modern Intel CPUs.
                    I know. But i think i saw a performance penalty with p-state powersave in some cases in Michaels last bench. Maybe i confused this with cpufreq

                    Comment

                    Working...
                    X