@Srinivas It doesn't look too good on thermald 1.9 until dptfxtract is used.
* Result of thermald 1.9 only: - https://docs.google.com/spreadsheets/d/1sSZe2gqfM8NWqaDyUfXx-Q2iGUkKvrREjJsobHfIZiQ/edit#gid=980670338 * Result of thermald 1.9 + dptfxtract 1.4.2: - https://docs.google.com/spreadsheets/d/1sSZe2gqfM8NWqaDyUfXx-Q2iGUkKvrREjJsobHfIZiQ/edit#gid=1205697583 -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to thermald in Ubuntu. https://bugs.launchpad.net/bugs/1811730 Title: Thermald does not set max CPU after reseting the voltage using RAPL Status in thermald package in Ubuntu: Confirmed Status in thermald source package in Bionic: Fix Committed Bug description: Hi, I was using Ubuntu 18.10 thermald package, but I noted that, after few seconds at max CPU usage (max temp), thermald send the signal to RALP to reduce the voltage of the CPU. It set the freq to minimum (800MHz in my case). But when the CPU is idle and temp is lowered (35-40ºC) it did not send the signal to resume normal operation of the CPU. I compiled the latest version of thermald from git (https://github.com/intel/thermal_daemon) and now everything works fine. I started to thought that the problem was the hardware or BIOS problem, as I disabled the CPU scaling on the BIOS, but intel_pstate continued doing freq scaling (I think for Turbo mode). But the real problem was Thermald. The latest version from git works really fine and it automatically disables and enable the Intel Turbo state and balance the freqs and fan control fine. My hardware is a Lenovo Thinkpad P52 with i7-8850H. I tested it using Ubuntu kernel and Kernel 4.20 optimized for i7 processor but with Ubuntu default config (except processor family="Core2/newer Xeon", Preemption Model="Preemptible Kernel (Low- Latency Desktop)" and Timer frequency="1000HZ". (Only changed those settings from make oldconfig and make deb-pkg). ProblemType: Bug DistroRelease: Ubuntu 18.10 Package: thermald (not installed) ProcVersionSignature: Ubuntu 4.18.0-13.14-generic 4.18.17 Uname: Linux 4.18.0-13-generic x86_64 NonfreeKernelModules: nvidia_modeset nvidia ApportVersion: 2.20.10-0ubuntu13.1 Architecture: amd64 CurrentDesktop: ubuntu:GNOME Date: Mon Jan 14 23:45:01 2019 InstallationDate: Installed on 2018-12-11 (34 days ago) InstallationMedia: Ubuntu 18.10 "Cosmic Cuttlefish" - Release amd64 (20181017.3) SourcePackage: thermald UpgradeStatus: No upgrade log present (probably fresh install) ----------------------------------------- SRU Justification ============== [Impact] * As described by the original bug reporter, CPU usage of Lenovo P52 is sub-optimal under heavy load. * My observation is the machine exhibits a sharp drop of power usage and CPU frequency and takes time to slowly ramp up again (refer to the chart at https://bit.ly/2OJphB8) * Fixed by bisecting and backporting fixes from thermald project. [Test Case] * One can stress the CPU load of the machine and collect the CPU frequency and power usage over time to check for any anamoly. The script at https://people.canonical.com/~ypwong/p52_test_cpu.sh can help with this. * With the fix, the behaviour should be like this: https://bit.ly/2KA9EXB, a consistent power usage can be maintained. [Regression Potential] * Medium. The fix consists of 7 commits cherry-picked from upstream, these changes will affect any machines using RAPL cooling device. The impact may not be obvious in normal daily usage but will be manifested during heavy load, suggest a longer verification period so that more people can discover any adverse effect. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/thermald/+bug/1811730/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
