Announcement

Collapse
No announcement yet.

KDE On Wayland: "The Biggest Thing Needed Now Is Adoption By 3rd Party Apps"

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

  • Originally posted by MrCooper View Post

    Excellent question! I thought I could help you understand some basic concepts of X & Wayland. It's clearly hopeless though, so time to bow out for real.
    You did, thank you, and apologies for that shortness, but really, compilation time is over (for both of us).

    Comment


    • Originally posted by mSparks View Post
      I know it does, but only in an idle or moderate use state.
      That wrong that its only at idle or moderate use state.
      Everything under control with the TUXEDO Control Center: Augsburg (Germany), 20th of April 2020. For a long time TUXEDO Computers has been fine-tuning its own Control Center - now the time has come: In order for customers to be able to make their own ...


      Above all, further energy or power settings should be made possible: The user should be able to control the maximum power consumption of both the CPU and the graphics card, in addition to the currently existing settings for the number of active CPU cores and their clock speed. The fan control will be extended so that users can create their own fan curves in addition to the currently preset, selectable profiles.
      Tuxedo are not your normal laptops. Not being in performance mode limits your CPU and GPU max clock speed because it caps the maximum power consumption and the max clock speeds.

      Performance mode has real meaning on Tuxedo laptops that the one your pointed to use to claim 10% difference. Yes Tuxedo without using Tuxedo control center the system with be locked in low fan noise mode instead of performance and that really does a number on the max clockspeed of the GPU and CPU.


      Fun fact about owning that type of laptop right mSparks you are on Linux you have to install vendor software or you don't get max performance.

      Comment


      • Originally posted by oiaohm View Post


        Tuxedo are not your normal laptops. Not being in performance mode limits your CPU and GPU max clock speed because it caps the maximum power consumption and the max clock speeds.
        the performance mode of nobara (assuming nobara ships with cpupower and tuned rather than something else) has exactly the same max cpu and gpu power limits as the powersave mode.
        Last edited by mSparks; 25 September 2023, 11:56 AM.

        Comment


        • Originally posted by mSparks View Post
          the performance mode of nobara (assuming nobara ships with cpupower and tuned rather than something else) has exactly the same max cpu and gpu power limits as the powersave mode.
          Remember he said it was not in performance mode.when he had the 10% worse. He was referring to the platform performance mode. So that mistake alone he had to re run the tests. Its kind of annoying he did not re run the Wayland tests with performance mode on to get the correct Wayland hit percentage.

          Fedora he had turned performance mode on so turn on running at max fan speed and max power limit and max clock speed. By his prior tests that 7% with Fedora between the two setting when he reviewed the laptop he got..

          Comment


          • Originally posted by oiaohm View Post

            Remember he said it was not in performance mode.when he had the 10% worse. He was referring to the platform performance mode. So that mistake alone he had to re run the tests. Its kind of annoying he did not re run the Wayland tests with performance mode on to get the correct Wayland hit percentage.

            Fedora he had turned performance mode on so turn on running at max fan speed and max power limit and max clock speed. By his prior tests that 7% with Fedora between the two setting when he reviewed the laptop he got..
            it was in wayland and not performance mode.
            However, the cpupower modes do not effect game performance. (limiting max cpu would, but cpupower does not limit max cpu, just when max cpu is called for), at least in default fedora.

            Comment


            • Originally posted by mSparks View Post
              However, the cpupower modes do not effect game performance. (limiting max cpu would, but cpupower does not limit max cpu, just when max cpu is called for), at least in default fedora.
              Not the case on all device and you are wrong how it works.​

              https://access.redhat.com/documentat...s_and_performa nce/tuning-cpu-frequency-to-optimize-energy-consumption_monitoring-and-managing-system-status-and-performance
              Active mode with hardware-managed P-states​
              • performance: With the performance governor, the driver instructs internal CPU logic to be performance-oriented. The range of allowed P-states is restricted to the upper boundary of the range that the driver is allowed to use.
              • powersave: With the powersave governor, the driver instructs internal CPU logic to be powersave-oriented.
              This is the problem here you are wrong on how it works. The laptop involved has P-states in the CPU so not being in performance has turned on hardware limitations in the CPU. The laptop in this case is made by a vendor who ACPI setting do have a direct effect on number of watts the CPU and GPU come into play with as well.

              cpupower setting can in fact limit you max CPU clock so it hits the power targets by hell of a lot either by p-state or by acpi settings.

              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

              Yes scary enough this difference can get very extreme. Like the above where 100watts more power usage on performance for 9 to 30% performance boost. Yes the performance difference in the video you showed is in the range of being possible powersave and the latter tests being run with performance being on as the only cause. This is why I said I would have liked to see wayland with performance on. Yes the performance mode change is most likely the biggest part of the 10%.

              mSparks not all systems are created equal with ACPI and CPU types. Some are like how you describe where cpupower does not limit max cpu but the laptop in the video is not one of those.

              Never try to draw a performance result from something where user changed two settings. Some cases the second setting changed is the real cause of difference.

              Changing to performance mode with Intel CPU in laptops its not strange to see 20 percent uplift in performance and the fans of the laptop go from being basically never running to be stuck on constant with the same workloads. 10%+ difference that is fully in the range to be nothing more than the cpupower switch.

              But knowing the laptop there does look to be a little wayland overhead and I would expect that with how much I know Nvidia drivers are broken. Note I said a little not 10% worth. Of course I would like to have correct numbers.

              Also thinking he used the laptop though all the other trying out of the os in Wayland mode and did not notice that the performance was down that in it self question for general users is the difference is even important. Maybe lack of tearing so reduced eyestrain for general usage is more important to your general users. General users think people doing business email/word documents and so on as those using make up about 80% of the PC in use.

              Comment


              • Originally posted by oiaohm View Post

                Yes scary enough this difference can get very extreme. Like the above where 100watts more power usage on performance for 9 to 30% performance boost. .
                I seem to recall we had this discussion before and you still don't seem to get it.
                limiting CPU freq very rarely improves efficiency - or all the CPUs and GPUs would have 150000 cores running at 1Mhz. Efficiency management is about how the CPUfreq is scaled based on load. CPUs are most efficient at max load. Any deviation from that will increase power consumption/decrease battery life.
                Powersave favours keeping the CPU at close to max load, performance favours keeping the CPU at max frequency regardless of the load.
                when you hit max load at max frequency there is nothing more you can do, and Powersave and performance modes are identical.

                Comment


                • Originally posted by mSparks View Post
                  Powersave favours keeping the CPU at close to max load, performance favours keeping the CPU at max frequency regardless of the load.
                  when you hit max load at max frequency there is nothing more you can do, and Powersave and performance modes are identical.
                  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

                  With the P-State powersave governor the highest peak frequency recorded during the entire duration of benchmarks was 5.07GHz but with the P-State performance governor was 5.3GHz (or even a few reported readings at 5.439GHz). The Core i9 11900K has an advertised Turbo Boost (2.0) frequency of 5.1GHz, which seems to be what is hitting with the powersave governor. While the i9-11900K also has Turbo Boost Max 3.0 and Thermal Velocity Boost where the peak is at 5.3GHz... It seems that 5.3GHz rating on Linux is only being hit if changing over to the performance governor.
                  Notice here this chip 2021 the Powersave and performance max frequency don't match.

                  mSparks again horrible out of date. I pointed to this 2021 review for a reason. Powersave and performance mode have not been identical performance on intel hardware for anything built after Jan 2021. It anywhere between 5% to 40% performance difference.

                  This is why the Wayland with wrong powermode being slow you cannot tell anything. His laptop is not that old.

                  Since jan 2021 Intel Powersave has the habit of locking you to the max clock speed on the spec sheet and Intel Performance allows boosting to what ever the thermal limit is that is based on how good the chips cooling is and the silicon lottery. Very different now numbers.
                  Last edited by oiaohm; 26 September 2023, 10:21 AM.

                  Comment


                  • Originally posted by oiaohm View Post

                    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


                    Notice here this chip 2021 the Powersave and performance max frequency don't match.

                    mSparks again horrible out of date. I pointed to this 2021 review for a reason. Powersave and performance mode have not been identical performance on intel hardware for anything built after Jan 2021. It anywhere between 5% to 40% performance difference.

                    This is why the Wayland with wrong powermode being slow you cannot tell anything. His laptop is not that old.

                    Since jan 2021 Intel Powersave has the habit of locking you to the max clock speed on the spec sheet and Intel Performance allows boosting to what ever the thermal limit is that is based on how good the chips cooling is and the silicon lottery. Very different now numbers.
                    That phoronix link doesn't disagree with that I said. In fact if you read it again you will see it 100% agrees with me.

                    All performance mode generally does is use much more power to achieve the same benchmark results.

                    performance mode is useful for the rare cases when you need a high frequency while the CPU isnt under high load, and you dont care about power consumption- and not much else.

                    You absolutely still hit the highest CPU states in powersave mode, just only when the CPU is heavily loaded.
                    Last edited by mSparks; 26 September 2023, 12:49 PM.

                    Comment


                    • Originally posted by mSparks View Post
                      That phoronix link doesn't disagree with that I said. In fact if you read it again you will see it 100% agrees with me.
                      With the P-State powersave governor the highest peak frequency recorded during the entire duration of benchmarks was 5.07GHz but with the P-State performance governor was 5.3GHz (or even a few reported readings at 5.439GHz). The Core i9 11900K has an advertised Turbo Boost (2.0) frequency of 5.1GHz, which seems to be what is hitting with the powersave governor. While the i9-11900K also has Turbo Boost Max 3.0 and Thermal Velocity Boost where the peak is at 5.3GHz... It seems that 5.3GHz rating on Linux is only being hit if changing over to the performance governor.
                      I see mSparks cannot read.

                      How is CPU running in powersave with a max of 5.07Ghz going to compete with performance with CPU maxing out at 5.3 Gpz.

                      Or even on Intel Xeon Scalable governor tests going from P-State powersave to performance was 11% faster performance while the CPU power consumption on average went up by 7% and no difference to the CPU peak power consumption. That jives with our P-State/CPUFreq testing over the years where in general changing the governor can offer up to a few percent faster performance at the cost of similarly a few percent higher power consumption on average.​
                      That write up also notes that this difference is not only restricted to the Intel Core i9-11900K. Xeon Scalable power usage stays the same but the Clock speed goes up because you are in performance. Yes notice that a 11% faster for the Xeon in 2020.

                      No the phoronix link does not support you claim that performance and powersave are the same.

                      Also note the first sentence of that write up.

                      The P-State and CPUFreq "performance" governors on Linux with desktop Intel/AMD processors can be of help for gaming and other select workloads by tending to keep the CPU clock frequencies higher than the default ondemand (CPUFreq) or powersave (P-State) governors used by nearly all Linux distributions.​
                      Yes the first sentence totally disagrees with you. Note gaming you get higher clock speeds so higher performance when you have performance governor on. Depending on chip how much this is. Yes the distribution default ondemard and powersave CPU do not clock as high. It use to be there was no difference between ondemard and performance but that has not been the case since at least 2020 in server chips in consumer cpus since 2021.

                      Workloads that hold CPU load constant that games are not you see the difference between performance and powersave basically evaporate. Performance CPU will boost higher resulting in 5-30% extra performance.

                      10% performance difference like it or not is in the range to totally be power management choice.

                      Comment

                      Working...
                      X