FIXED. Chat & Screens & Benchs: Various Intel Gen CPUs & X570 + Ryzen 3rd Gen

[olejon]

• make fails on Linux 4.15.0-52-generic, current latest kernel for Ubuntu 18.04 LTS 64-bit, with Vanilla GNOME (used as developer and main workstation + some server services)
• The package intel-microcode was also updated recently
• Straight forward unlocked SkyLake Workstation CPU: Intel Core i5-6600K
• Worked like a charm first time I installed CoreFreq
• Logically did not work after next kernel update, since CF kernel module was not compiled for it
• Compiled again and worked fine again
• BUT now, with Linux 4.15.0-52-generic, it fails with the error below
• Installed libelf-dev since make suggested to do so, and that warning goes away, but still build fails
• Has not bothered with adding it to dkms since make takes like 5 seconds on my system (seems like that was a good decision)
• Overclocked simply enabling Intel XMP in UEFI, which enables all cores to go to max TurboBoost ("sync all cores", 3.9 GHz under constant all cores max stress = temperature fine) + XMP automatically detects my XMP compatible 3200 MHz DDR4 RAM, and such sets correct values

$ make
...
make -j1 -C /lib/modules/4.15.0-52-generic/build M=/home/username/Downloads/CoreFreq modules
make[1]: Entering directory '/usr/src/linux-headers-4.15.0-52-generic'
  CC [M]  /home/username/Downloads/CoreFreq/corefreqk.o
/home/username/Downloads/CoreFreq/corefreqk.c: In function ‘SMBIOS_Collect’:
/home/username/Downloads/CoreFreq/corefreqk.c:9107:5: error: ‘DMI_PRODUCT_SKU’ undeclared (first use in this function); did you mean ‘DMI_PRODUCT_UUID’?
   { DMI_PRODUCT_SKU, Proc->SMB.Product.SKU },
     ^~~~~~~~~~~~~~~
     DMI_PRODUCT_UUID
/home/username/Downloads/CoreFreq/corefreqk.c:9107:5: note: each undeclared identifier is reported only once for each function it appears in
scripts/Makefile.build:337: recipe for target '/home/username/Downloads/CoreFreq/corefreqk.o' failed
make[2]: *** [/home/username/Downloads/CoreFreq/corefreqk.o] Error 1
Makefile:1552: recipe for target '_module_/home/username/Downloads/CoreFreq' failed
make[1]: *** [_module_/home/username/Downloads/CoreFreq] Error 2
make[1]: Leaving directory '/usr/src/linux-headers-4.15.0-52-generic'
Makefile:70: recipe for target 'all' failed
make: *** [all] Error 2

I see you did varios SMBIOS 6 days ago. That is what failing it seems.

I haven't installed/purged or changed anything that should have anything to do with building from source code.

[cyring]

Will fix asap.
Some dmi might be available with latest kernel 5.x

[olejon]

Great :-)

[ss1966]

I need to login hp laptop that has bitlocker, I have no idea how to go about this

[cyring]

I need to login hp laptop that has bitlocker, I have no idea how to go about this

Create another issue if yours concerns CoreFreq ?

[cyring]

Just checked (w/ Smartphone) the minimum Kernel version is 4.18.x for those missing definitions (DMI_PRODUCT_SKU,...)

I will refactor code using conditions to let previous versions build through.

Also thinking about writing myself parts of the DMI parsing to offer the remaining strings

[olejon]

I usually install the Ubuntu LTS Enablement Stacks once they are out (newer kernel and Xorg/Wayland/Mesa/etc for desktops and just newer kernel for servers).

Pretty sure the one's in the repo now is > 4.18. On my phone now.

https://wiki.ubuntu.com/Kernel/LTSEnablementStack

The 18.04.2 and newer point releases will ship with an updated kernel and X stack by default for the desktop. Server installations will default to the GA kernel and provide the enablement kernel as optional. The 18.04 HWE Stacks will follow a new Rolling Update Model as documented at the following location: https://wiki.ubuntu.com/Kernel/RollingLTSEnablementStack Installing the HWE stack is simple: DESKTOP sudo apt-get install --install-recommends linux-generic-hwe-18.04 xserver-xorg-hwe-18.04

ANYWAY most Ubuntu LTS users don't do this manually.

And my installation has been totally stable. The open shipped with amdgpu driver, which AMD maintains together with the community, is actually faster than installing their latest open one from their repo for exactly Ubuntu 18.04 so Ubuntu maybe optimize more or regression in later Mesa packages or other libs it updates, as it updates a ton of packages as well from the AMD repo, including Xorg stuff IIRC. The Pro driver is constantly beaten by the open one, also. Really don't need the Pro one anyway unless you do some specific stuff. I don't even need a dedicated GPU as I don't play games or use GPU-heavy software. Most of that does my Raspberry Pi's GPU in form of HW acceleration of my Terabytes of video stored on this exact workstation. Well the Android Emulator is one example, but the iGPU from Intel is waaaay good enough. Only difference I see is in glmark2 :P And of course vulkan-smoketest.

Basically bought it intrigued by reading about how open and good AMD GPUs have become on Linux (especially the first one meaning it'll only get better), with everything plug and play, making it open enough to ship with the kernel, and actually fully working with the community, which NVIDIA does not (just randomly sometimes, the open driver is a mess if you want to use that, gotta check that table with supporting this but not that for everything - the closed really good and faster still, but must be installed as a separate package, and Wayland compatibility well, dunno).

Basically supporting AMD :-)

And Intel since all their HW is dead stable out of the box on Linux, including iGPU, much easier basically than Windows, and they actively contribute to the kernel for people to make tools such as your own, and even have created an own super fast distro, not for your daily driver but if supported by your CPU fun to check out! It has its own package system, more and more packages, including full GNOME desktop, all vanilla packages compiled for max speed on Intel cipset MB and CPU. Rolling release IIRC,

[olejon]

BTW will install the above and check if I get to it before you do :) Don't worry, no rush for me.

Just gotta say CoreFreq, although praised in the article I found it in, was way above my expectations! You should try to get it into the Debian/Ubuntu repos (can use dkms to allow updates to latest tested stable version of your code - or fixed version is probably preferred by distro maintainers, updated when distro is, as long as you make a well tested deb package from latest code, as it is already feature rich as heck, and probably systemd service to get approval I suppose but that isn't that hard when FOSS...).

EDIT: Found CoreFreq here: https://medium.com/@datamove/over-clocking-under-linux-67d04cfb0974

[cyring]

Hello, for your tests, fix is committed.

[olejon]

Thanks for responding so fast and fixing. I will test soon, with the kernel mentioned in the issue.

[olejon]

Confirmed building as before and working on Linux 4.15.0-52-generic without installing the LTS Enablement Stacks.

• Ubuntu 18.04 LTS 64-bit with latest default desktop kernel and intel-microcode package

Thanks again!

I have a suggestion. Basically toggling NMI Watchdog on demand, because... NMI Watchdog is there for a reason, as far as I have read I do not want it off permanently on such an important workstation + server.

My solution:

I have created a script like this as /usr/local/bin/cf. I have added /usr/local/bin/cf to the whitelist using visudo so it does not require a password to run, so I can simply run sudo cf.

#!/bin/bash

### CF SCRIPT IS RUN

bash -c "echo 0 > /proc/sys/kernel/nmi_watchdog"

sleep 1

insmod /opt/corefreq/corefreqk.ko

sleep 1

/opt/corefreq/corefreqd &

sleep 2

su -l -c "/opt/corefreq/corefreq-cli" MYDESKTOPUSERNAME

### CTRL+C IS PRESSED TO STOP THE CLI

sleep 4

rmmod corefreqk.ko

sleep 1

bash -c "echo 1 > /proc/sys/kernel/nmi_watchdog"

The sleeps are necessary to give some time before starting and stopping. I have added a little extra because you never know. Sometimes it may fail to rmmod the module if only 2 seconds after quitting the cli, for instance. As you can see I copy the 3 necessary files to /opt/corefreq/ after make.

BTW: The reason for bash -c is that running echo X > Y does not work with sudo in these cases. Pages often suggest running sudo su first, but that is ugly, not for scripts made to avoid prompts, and just plain wrong. Those using sudo su should run sudo -s or sudo -i, depending on need. That is the right way, NEVER sudo su (some exceptions may exist, usually the first time doing something then never again).

Clue is NMI Watchdog toggle - on-demand - like:

When starting, first run:

bash -c "echo 0 > /proc/sys/kernel/nmi_watchdog"

When everything is stopped/unloaded, run as last command:

bash -c "echo 1 > /proc/sys/kernel/nmi_watchdog"

Kernel confirmes in dmesg and journalctl -b --no-pager that NMI Watchdog is enabled again with:

kernel: NMI watchdog: Enabled

Pretty sure you can deactivate/reactivate NMI Watchdog in corefreqd daemon on start/stop? At least it can be an option? People should at least know the consequences of NMI Watchdog off permanently, and that it can be toggled on-demand instead of kernel parameter.

Add it to README.md?

[cyring]

Thanks, I will process your returns.
Meanwhile, nmi_watchdog is not the source of the incompatibility issue; it just rewrites periodically the fixed counter register. And that biases the readings. Evidence can be found in the Kernel source code.

As a workaround you can also build CoreFreq with directives to use other registers: APERF and MPERF
Just type make help

You can then keep the nmi_watchdog while corefreqk.ko reads MSR_IA32_APERF and MSR_IA32_MPERF

[olejon]

Thanks. Just adding a little information about this to README.md would seem enough. It's a tool for power users anyway. Maybe a link to what the NMI Watchdog does. You can maybe add my script to the repo as an example if you want to. My solution works just fine so I won't touch the build process. I don't use CF that often. It does replace a lot of other BASH aliases I've created for frequencies and temperature, updated every X seconds. I use them still to verify (frequencies polled from the kernel and temperature from sensors). Also having the GUI PSensor program is nice. Lots of flexibility. Uses lm-sensors, AMD/NVIDIA APIs if found and udisks or udev for disks temperature, no need for the hd monitor daemon package. One nice thing is that you can set alarms if a certain part, say the CPU package or individual core goes above a certain temperature. A secure way to avoid parts being destroyed if HW and kernel modules built in protection fails and a heavy process has freezed and causes 100 % usage, especially if over clocked and basic protection like voltage regulator is disabled in UEFI. That heats up the CPU extremely fast under stress -c numcores. Just set command to run, if above X °C, run bash -c reboot for instance. Could be something for this tool, the daemon at least. Well AFAIK CF doesn't write any configuration, so would need that. htop does. Nice since I always change the settings. Basically the only basic monitoring tool CF can't replace, but CF isn't really meant to either I suppose.

BTW I can probably just fine host CF as a deb in my repo (only for packages that never are in the Debian/Ubuntu repos). Build it from latest source, (tested by me first on Debian and Ubuntu) on first install and add to dkms, or simply ask user to apt install --reinstall corefreq after kernel update. I saw a PPA it seems you made but it's totally out of date, and no link to it from here. But could use it as a base for my deb. Of course only if you're interested, like having a link to an up to date CF repo for debs.

[cyring]

Getting your support for distributions repo will be a great help. Indeed I want to focus on R&D.

CoreFreq is monitoring thermal thresholds, those which are Processor internal such as thermal throatling. For instance if temp reaches tjmax, HOT will light on. Press the H key to clear the associated event register bit.

I don't think about programming what other monitoring software are doing much better. The goal of CoreFreq is more to offer features which are not pretty centralized in the Linux space: Processor + BIOS + OS + Performance
Although I'm noticing that CoreFreq is used in areas I was not expecting to see it: Trading, Production,...; CoreFreq remains an experimental open source subject and I have tons of idea not developed yet.

[cyring]

1. In your script I don't see when the daemon is requested to close. It may explain why the driver has shutdown issue. Shared memory is not cleanly released.

Please give a look into corefreqd.service : the daemon is listening to signals.
Daemon is then sending a shutdown signal to the client ==> It's a cascade procedure.

2. Disabling the nmi_watchdog from the user-space won't guarantee the counter register is freed by kernel. That's why I recommend to blacklist the counter usage straight from the Kernel boot command line. However, as stated previously, you can build CoreFreq with APERF/MPERF counters and let nmi_watchdog runs.

3. Give a look into the other monitoring GUI tools documentation and source code for hardware probing incompatibilities. I have started a list of CoreFreq conflicts into the Wiki - Software incompatibility

Regards
Cyril

[olejon]

First of all, with this script, the daemon, for some reason, shuts itself down like immediately after pressing Ctrl+C, BUT only when the commands are run in a script like that, even when the CLI program is run as the desktop user. Easy fix to be 100 % sure on all systems anyway if you read below.

EDIT: Without using any *.service file, just those 3 files. No signals then I suppose.

HOWEVER, when the commands are run separately one by one, all commands as root except the CLI program which is run by the desktop user either using su -l -c or simply in a desktop terminal, it does NOT shut itself down. Interesting. But running this script on my system, this way, always makes the daemon shut itself down, before the module is unloaded, because it won't unload of course if corefreqd is still running, it then refuses to rmmod the module (still in use), but that never happens when run as a script this way on my system, only when the commands are run one-by-one in 2 terminals, one for the root commands and one for the CLI program, or using su -l -c in same terminal after running the daemon.

Easy to fix to be 100 % sure. Before rmmod, add command:

pkill corefreqd

Note: pkill does not give any output if the wanted-to-kill command is not running either so no need to send unnecessary output to /dev/null or anything, but one can use pkill || command or pkill && command and such, or combine with pgrep.

This I discovered and have now confirmed, WITHOUT the pkill command, by running ps aux | grep coref and also lsmod, which gives that both the CLI program and the daemon is not running anymore when closing the CLI using Ctrl+C, NONE of them, SO fast that I can not even switch to the other terminal using Alt+Tab to run ps and see it - it is ALREADY shut down. The module as per the script not unloaded before 4 seconds, so it is a "safety buffer" to 100 % avoid any error. lsmod confirmes it is unloaded. It will say still in use when the daemon is running so...

I can not explain why, but that is why I have not added a command to shut down the daemon - because it does it itself on my system. The sleep time of 4 seconds was because a few times just got the message that the module was still in use with a sleep time of just 1 second for instance. But 4 seconds is a sure "safety buffer", more than enough seconds, just added to NEVER ever get that the module is still in use.

Now my script is like this and seems to work just fine:

#!/bin/bash

bash -c "echo 0 > /proc/sys/kernel/nmi_watchdog"

insmod /opt/corefreq/corefreqk.ko

/opt/corefreq/corefreqd &

sleep 2

su -l -c "/opt/corefreq/corefreq-cli" MYDESKTOPUSERNAME

pkill corefreqd

sleep 2

rmmod corefreqk.ko

bash -c "echo 1 > /proc/sys/kernel/nmi_watchdog"

Both ps and lsmod shows everything is killed and unloaded fine. No errors from kernel or terminal.

Only output from terminal (run and exit):

CoreFreq Daemon 1.57.0  Copyright (C) 2015-2019 CYRIL INGENIERIE
 ...terminated.

Only output from dmesg (run and exit):

CoreFreq(2:-1): Processor [ 06_5E] Architecture [Skylake/S] CPU [4/4]
CoreFreq: Unload
NMI watchdog: Enabled [again]. Permanently consumes one hw-PMU counter.

Note 1: Kernel does not report when NMI Watchdog is disabled, only when enabled, but can be verified disabled by checking value of /proc/sys/kernel/nmi_watchdog of course, which is also 0 when using the kernel parameter.

Note 2: As you can see I have removed some unnecessary sleep commands (on my system at least), and shortened sleep time to CLI to 2 seconds and exit to 2 seconds only. Not annoying for such a tool. Now, this should maybe be tested on a slow system, slow in every way, not sure if 100 % enough to tweak UEFI and kernel to "make a slow system", and since CoreFreq is logically not made for VMs, it is harder then. Basically having both more sleep commands and sleep seconds makes sure it should work on slow systems as well.

Does it work correctly on 32-bit? If so I may test on a really old laptop... Maybe CoreFreq does not support that CPU even, like not enough "info and toggles" in the kernel for CoreFreq to use, or CoreFreq is aimed at newer CPUs than that one.

Regarding NMI Watchdog: I was not aware of that, hmm... And do not have this high level of understanding you have of CPUs and the Linux kernel. Just seen many tutorials, blog posts, wikis etc around monitoring tools and overclocking tends to use on-demand toggles, like my command, or disabling NMI Watchdog in sysctl or something in a way that makes it reversible without a reboot, not as a kernel boot parameter.

In MY case I do not see any difference if it is enabled or disabled. I will double check. Especially interesting then to compare CoreFreq values when NMI Watchdog is disabled as a kernel boot parameter compared to not disabled in any way.

Regarding deb package: Have a pretty good idea. Instead of in a repo, you simply host it here on GitHub, since updating it will basically never be needed after it is well tested (2 of the 4 dependency packages in question may change names on newer versions of Debian/Ubuntu, but not like every year, so, and easy peasy to fix then, because you can add packages with alternative names).

• Basic control file, we only need 4 small dependencies on Debian/Ubuntu derivatives. One is build-essential which is a metapackage that depends on basic build from source tools, but does not bloat the system or use much bandwidth or space
• The package only puts 2 files into right place (script to run CoreFreq in /usr/bin or /usr/sbin, and the copyright file standard for debs)
• The postinst script does all the magic, including acting as an upgrade script if CoreFreq already found to be installed (easy check), cloning latest from GitHub and compiling into suitable folder like /opt/corefreq
• If package is deleted, the prerm script simply deletes /opt/corefreq. The 2 other files are automatically deleted as they are part of the package.
• The script to run CoreFreq will check if NMI Watchdog is enabled, and if so, give a warning and a link to https://github.com/cyring/CoreFreq
• It will also check if the kernel has changed since last run, give a warning, and offer [Y/n] to download the deb from this repo and reinstall it, and such it will compile it again against the new kernel headers/source, which has the advantage that users will then be using your latest code when they run CoreFreq after a kernel update and accept to reinstall it, cloning the latest code

Note: This is EXACTLY what the VirtualBox Guest Additions install script for VMs under VirtualBox does, puts files in /opt, compiles Guest Modules, I think it now adds them to DKMS, BUT when a new version of the VB Guest Additions are available, always when a new version of VirtualBox is out, then you get a warning to install the new VB Guest Additions (and reboot the VM). I always use the CLI way on Linux VMs, run the script directly as root and reboot. IDK if running the software you are asked to run under certain Desktop Environments, when inserting the virtual CD with the guest additions, actually work, and if so, I am sure it just opens the default terminal and runs the script.

Not like on Windows, where it has a standard GUI installer with "Next > Next > Bla bla > Compiling drivers > Windows reacts to new drivers > Allow > Bla bla > Next > Reboot", or something in that typical order. I do not use VirtualBox anymore in Linux, use KVM+QEMU+libvirt, which is way better of course. Just on macOS I use VB.

[olejon]

EDIT: Some changes, typos, formatting etc in post above, sure there are more typos :P

And yes, that's what great with CoreFreq, it does not try to do things other tools do as good as possible anyway. It pulls together things that otherwise needs various commands/scripts/tools.

I have seen myself small projects I made like a decade ago and put on the Internet being used in ways I never thought. Even pages where people offer > 500 USD to any freelance developer taking such jobs - to modify it to do something more or integration... Always fun to find, especially small web apps I have made and by searching finding it hosted on some university loooong away from Norway :P Although stated not supported in any way except read the docs and be aware that the code and the docs will become deprecated with newer libs, server software, web standards etc. People find me through all sorts of channels, and basically want "free help" to rewrite a ton of the code for just their use case... NOPE! And darn that Facebook Messenger allowing for non-friends in any way to send a super long "Message requests", just as an e-mail basically, but more intrusive/pushing IMO...

[olejon]

• My bad, regarding asking about 32-bit, didn't even read the main description - that it is made for 64-bit CPUs.

• I ran make so that NMI Watchdog can be on. At a glance it seems to show the same information as before. "SMBIOS data" shows "To be filled by O.E.M." but pretty sure it was like that before, and if I understood correctly could show the info if kernel is >= 4.18, which can be done just by installing LTS Enablement Stacks (one command). All other data shows, basically. Although my RAM is running at 3200 MHz (Intel [Certified] XMP mode), but shows standard DDR4 value - 2166 MHz.

Q: That data does not really matter for me anyway, but does it make CoreFreq more accurate if available?

[cyring]

Hello,

I'm working on TDP limits and I would appreciate testings from various Intel Atom, Core ix, Xeon.
Could you give a try to the develop branch ?

The UI Power[w] window or option -s will reveal those PL1 PL2 per domain. As specs, Package and Platform have both; other domains have only PL1

This is an experimental feature, CPU may crash.

Thanks for trying.

[olejon]

---

You're Interested in my old Intel NUC i3? If so say yes.

---

Just for fun, I checked Blender Opendata for Blender Benchmarks, and the conclusion is:

• If you run Linux, there's no reason to upgrade from a 3950X to a 5950X. The Net Gain for the scene most run, bmw27 (always shows up when sorting by rendering time and no scene filter), the 5950X only beats the 3950X by a lousy 3 seconds, and yes, it's heavy, just takes about ~1min20secs (or shorter when using the below)
• If you run Windows, as I only have added 50 results in the screenshot, you see you gotta go way down to see Windows results, and yes, for Windows the 5950X is the only one appearing. However some more digging and it's not a big deal on Windows either, but bigger than on Linux, yet Windows with the latest official AMD Chipset Drivers is always beaten by Linux, where the 3950X on Linux beats the 5950X by a whopping 10 seconds on the shortest render (again the 5950X only beats the 3950X by a max of 3 seconds here)
• The 5900X is NOT worth it if having a 3950X already. It's way down. It's not included on the screenshot for that reason
  If you have the money, well, it might be better, BUt mainly in games because of the frequencies, and what game can't the 3950X run perfectly with a good GPU? None. For AMD the 6800 series gets the best results for GPU benchmarks on Blender, however NVIDIA is still somewhat better and with a 6800 the price starts to become pretty equal with worse results, so (and let's face it the NVIDIA driver for Linux is excellent and their Control Panel is more modern than the Windows one, which still uses XP icons) and the PSU requirement is huge as with any high-end GPU - not for me! I still run a 500W PSU and no problems with my MSI RX 570 OC 8 GB which is supported by AMD's Pro driver IF I want it, but as OpenCL is "legacy" for my card, it only adds OpenCL to say Blender. And boom with Ubuntu 20.04.2 it stopped working. Just no Xorg. Now the driver says specifically "Official support for Ubuntu 20.04.2" (I was stupid and didn't read that it didn't support 20.04.2 yet, then the next version said Initial beta support for 20.04.2, and now it says it's Official - but then comes 20.04.3 and yeah - it's not something APT tells you - which of distros it supports alongside CentOS and RHEL - and hence partly openSUSE with some skills but it's listed. However, no Pro OpenCL available for me, usually the next kernel upgrade has all the goods in the open one for me, with better FPS and. No OpenCL shows up on Blender Benchmark and IIRC it doesn't work in Geekbench either without AMD's driver - IIRC doesn't have to be the Pro one in my case but you want Pro if you go that route... Docs are good and Release Notes, definitely read (lesson learned)
• The only benchmarks (no matter OS and all or some scenes) where someone beats my 3950X are mainly GPUs on Windows. However, the best are AMD Epyc (especially if running 2 sockets) and AMD Threadripper (3990X & sometimes but not always the 3970X). The 3990X is appealing but watch Linus Tech Tips (YouTube) on it and you see why it's kind of a bad deal, compared to Epyc. It's really for a great desktop but there's limits you don't want for so much money
• If you can trust your MB manufacturer, and well it gotta be a shitty one if you can't, you can, after some searching, securely enable PBO with Limits = Motherboard, as the FIT part of the CPU will make sure it never goes above 1.30V and therefore make no harm. You can easily check this with CoreFreq. Run something heavy and you'll see it runs around 1.29V, lower or sometimes exactly 1.30V, but never more, so basically forget the high Watts Reading. If it goes too hot it simply turns off, AS LONG AS YOU Leave Scalar at Auto or 1X (Same As Auto), as otherwise it might reduce your CPUs max turbo over time, but we're talking years for a 50 MHz decrease but you never know. But since Scalar above 1X doesn't give any performance gain, WHY put it above 1X? You might wanna set it to manual just to make sure it's set to 1X, and this under AMD Overclocking, as the MB one I've seen at mine use 2X as default IF set to Manual
• NEVER OVERCLOCK A RYZEN X570 CPU WITH MANUAL VOLTAGE ADJUSTMENTS, as that basically disables FIT and you're on your own, and you're likely to get more from PBO Limit = Motherboard so… And if not, still a bad idea. Basically, avoid manual OC. Use the built-ins. It's what AMD intended. Intel is another story but they're so behind (again watch Linus Tech Tips there for the latest Threadrippers - AMD has the luxury Intel once had aaaand, use it to sell CPUs great on specs but a closer look and really? You want to waste so much and be restrained by RAM limits and PCIe lanes limits? Likely not)
• To say, I've yet to get any problems with Limit set to Motherboard. I do stress using CoreFreq to monitor to check that Voltage doesn't go above 1.30V after say a BIOS or Kernel upgrade
• Once my journalctl -b --no-pager showed some CPU error after running a full Benchmark, with regular values, but I haven't been able to make it happen again. I guess it was simply some RAM corruption which after all does happen without ECC RAM. One bit is wrong and could cause nothing or a full system freeze. Here it just showed up as bad in the logs and I rebooted. Simple as that

[cyring]

Hello @olejon

Yes I'm particularly interested in your Intel NUC i3 testing results.
Be sure to test with latest develop branch where Power Clamping is implemented.
In your CLI screenshots, observe the sensors and max frequency whenever you're altering the power domain settings: PL1, PL2 limit, clamp, values (watt offset)

On the AMD side, unfortunately no specs available for such Power features. I love my 3950X but less the stability of BIOS firmware, and the lack of registers documentation. Intel is still the champ for Engineers.

Various codes have been arranged lately to provide those PL; please let me also know if you face regressions.

Btw, I've provided a new loop for Goldmont monitoring. But no more tests from User. Last screenshot, I noticed wrong Formula units in Voltage, Power. I'm looking for various Atom, Celeron, Pentium to complete that work.

Regards
Cyril

[olejon]

@cyring Alright. Here you go. Cheers 🍻

OS: Ubuntu Minimal (Netinstaller) 20.04.2 LTS fully updated (Linux 5.8.18)
USE: Headless Server
CPU: Intel(R) Core(TM) i3-4010U CPU @ 1.70GHz Haswell/Ultra Low TDP

BUILD + RUN:

$ make
$ sudo insmod ./corefreqk.ko
$ sudo ./corefreqd -q &

NOTE: Hyperthreading disabled in BIOS as this CPU is part of the Intel Core generations that suffers from some of these mainly Intel security bugs, and instead of just giving VMs the actual cores (not threads too), which fixes the bug, or use the Linux kernel parameter to boot it without HT, I just disabled it entirely, as it runs almost as fast without, and way lower temps.

REGRESSION?

Not really. It's not a bug in CoreFreq, more like was a "feature", that does nothing for the CPU's performance anyways. Just saying.

This is also in the Stable branch. As you explained certain CPUs from Intel, including this one, will always trigger HOT > Power Limitation at boot. Before I could clear it with H, but now the Stable branch clears it for like 1 second (grey not cyan) then it's back. The Develop branch doesn't show it greyed out at all (though that's more consistent). IDK if this is CoreFreq code change or e.g. Kernel or Microcode change (that NUC hasn't received a BIOS update since 2018 or something). Anyways, doesn't matter, but if the whole thing is broken for those where it does matter (like on my MacBook I can fool throttling very temporarily).

SCREENSHOT:

RESULT OF OUTPUT:

$ ./corefreq-cli -s

Processor                             [Intel(R) Core(TM) i3-4010U CPU @ 1.70GHz]
|- Architecture                                          [Haswell/Ultra Low TDP]
|- Vendor ID                                                      [GenuineIntel]
|- Microcode                                                        [0x00000026]
|- Signature                                                           [  06_45]
|- Stepping                                                            [      1]
|- Online CPU                                                          [  2/  2]
|- Base Clock                                                          [ 99.772]
|- Frequency            (MHz)                      Ratio                        
                 Min    798.18                    <   8 >                       
                 Max   1696.12                    <  17 >                       
|- Factory                                                             [100.000]
                       1700                       [  17 ]                       
|- Performance                                                                  
   |- P-State                                                                   
                 TGT    798.18                    <   8 >                       
|- Turbo Boost                                                         [ UNLOCK]
                  1C   1696.12                    <  17 >                       
                  2C   1696.12                    <  17 >                       
                  3C   1696.12                    <  17 >                       
                  4C   1696.12                    <  17 >                       
|- Uncore                                                              [   LOCK]
                 Min    798.18                    [   8 ]                       
                 Max   1696.12                    [  17 ]                       
|- TDP                                                           Level <  0:3  >
   |- Programmable                                                     [ UNLOCK]
   |- Configuration                                                    [ UNLOCK]
   |- Turbo Activation                                                 [ UNLOCK]
             Nominal   1696.12                    [  17 ]                       
              Level1    798.18                    [   8 ]                       
               Turbo   1596.35                    <  16 >                       
                                                                                
Instruction Set Extensions                                                      
|- 3DNow!/Ext [N/N]          ADX [N]          AES [Y]  AVX/AVX2 [Y/Y] 
|- AVX512-F     [N]    AVX512-DQ [N]  AVX512-IFMA [N]   AVX512-PF [N] 
|- AVX512-ER    [N]    AVX512-CD [N]    AVX512-BW [N]   AVX512-VL [N] 
|- AVX512-VBMI  [N] AVX512-VBMI2 [N]  AVX512-VNMI [N]  AVX512-ALG [N] 
|- AVX512-VPOP  [N] AVX512-VNNIW [N] AVX512-FMAPS [N] AVX512-VP2I [N] 
|- AVX512-BF16  [N]  BMI1/BMI2 [Y/Y]         CLWB [N] CLFLUSH/O [Y/N] 
|- CLAC-STAC    [N]         CMOV [Y]    CMPXCHG8B [Y]  CMPXCHG16B [Y] 
|- F16C         [Y]          FPU [Y]         FXSR [Y]   LAHF-SAHF [Y] 
|- MMX/Ext    [Y/N] MON/MWAITX [Y/N]        MOVBE [Y]   PCLMULQDQ [Y] 
|- POPCNT       [Y]       RDRAND [Y]       RDSEED [N]      RDTSCP [Y] 
|- SEP          [Y]          SHA [N]          SSE [Y]        SSE2 [Y] 
|- SSE3         [Y]        SSSE3 [Y]  SSE4.1/4A [Y/N]      SSE4.2 [Y] 
|- SERIALIZE    [N]      SYSCALL [Y]          SGX [N]       RDPID [N] 
                                                                                
Features                                                                        
|- 1 GB Pages Support                                      1GB-PAGES   [Capable]
|- Advanced Configuration & Power Interface                     ACPI   [Capable]
|- Advanced Programmable Interrupt Controller                   APIC   [Capable]
|- Core Multi-Processing                                  CMP Legacy   [Missing]
|- L1 Data Cache Context ID                                  CNXT-ID   [Missing]
|- Direct Cache Access                                           DCA   [Missing]
|- Debugging Extension                                            DE   [Capable]
|- Debug Store & Precise Event Based Sampling               DS, PEBS   [Capable]
|- CPL Qualified Debug Store                                  DS-CPL   [Capable]
|- 64-Bit Debug Store                                         DTES64   [Capable]
|- Fast-String Operation                                Fast-Strings   [Capable]
|- Fused Multiply Add                                     FMA | FMA4   [Capable]
|- Hardware Lock Elision                                         HLE   [Missing]
|- Instruction Based Sampling                                    IBS   [Missing]
|- Long Mode 64 bits                                       IA64 | LM   [Capable]
|- LightWeight Profiling                                         LWP   [Missing]
|- Machine-Check Architecture                                    MCA   [Capable]
|- Memory Protection Extensions                                  MPX   [Missing]
|- Model Specific Registers                                      MSR   [Capable]
|- Memory Type Range Registers                                  MTRR   [Capable]
|- OS-Enabled Ext. State Management                          OSXSAVE   [Capable]
|- Physical Address Extension                                    PAE   [Capable]
|- Page Attribute Table                                          PAT   [Capable]
|- Pending Break Enable                                          PBE   [Capable]
|- Process Context Identifiers                                  PCID   [Capable]
|- Perfmon and Debug Capability                                 PDCM   [Capable]
|- Page Global Enable                                            PGE   [Capable]
|- Page Size Extension                                           PSE   [Capable]
|- 36-bit Page Size Extension                                  PSE36   [Capable]
|- Processor Serial Number                                       PSN   [Missing]
|- Resource Director Technology/PQE                            RDT-A   [Missing]
|- Resource Director Technology/PQM                            RDT-M   [Missing]
|- Restricted Transactional Memory                               RTM   [Missing]
|- Safer Mode Extensions                                         SMX   [Missing]
|- Self-Snoop                                                     SS   [Capable]
|- Supervisor-Mode Access Prevention                            SMAP   [Missing]
|- Supervisor-Mode Execution Prevention                         SMEP   [Capable]
|- Time Stamp Counter                                            TSC [Invariant]
|- Time Stamp Counter Deadline                          TSC-DEADLINE   [Capable]
|- TSX Force Abort MSR Register                            TSX-ABORT   [Missing]
|- TSX Suspend Load Address Tracking                       TSX-LDTRK   [Missing]
|- User-Mode Instruction Prevention                             UMIP   [Missing]
|- Virtual Mode Extension                                        VME   [Capable]
|- Virtual Machine Extensions                                    VMX   [Capable]
|- Extended xAPIC Support                                     x2APIC   [ x2APIC]
|- Execution Disable Bit Support                              XD-Bit   [Capable]
|- XSAVE/XSTOR States                                          XSAVE   [Capable]
|- xTPR Update Control                                          xTPR   [Capable]
Mitigation mechanisms                                                           
|- Indirect Branch Restricted Speculation                       IBRS   [Capable]
|- Indirect Branch Prediction Barrier                           IBPB   [Capable]
|- Single Thread Indirect Branch Predictor                     STIBP   [Capable]
|- Speculative Store Bypass Disable                             SSBD   [Capable]
|- Writeback & invalidate the L1 data cache                L1D-FLUSH   [Capable]
|- Hypervisor - No flush L1D on VM entry            L1DFL_VMENTRY_NO   [Missing]
|- Architectural - Buffer Overwriting                       MD-CLEAR   [Capable]
|- Architectural - Rogue Data Cache Load                     RDCL_NO   [Missing]
|- Architectural - Enhanced IBRS                            IBRS_ALL   [Missing]
|- Architectural - Return Stack Buffer Alternate                RSBA   [Missing]
|- Architectural - Speculative Store Bypass                   SSB_NO   [Missing]
|- Architectural - Microarchitectural Data Sampling           MDS_NO   [Missing]
|- Architectural - TSX Asynchronous Abort                     TAA_NO   [Missing]
|- Architectural - Page Size Change MCE               PSCHANGE_MC_NO   [Missing]
|- Architectural - Split Locked Access Exception                SPLA   [Missing]
                                                                                
Technologies                                                                    
|- Data Cache Unit                                                              
   |- L1 Prefetcher                                                L1 HW   < ON>
   |- L1 IP Prefetcher                                          L1 HW IP   < ON>
   |- L2 Prefetcher                                                L2 HW   < ON>
   |- L2 Line Prefetcher                                        L2 HW CL   < ON>
|- System Management Mode                                       SMM-Dual   [ ON]
|- Hyper-Threading                                                   HTT   [OFF]
|- SpeedStep                                                        EIST   < ON>
|- Dynamic Acceleration                                              IDA   [OFF]
|- Turbo Boost                                                     TURBO   <OFF>
|- Energy Efficiency Optimization                                    EEO   <OFF>
|- Race To Halt Optimization                                         R2H   <OFF>
|- Virtualization                                                    VMX   [ ON]
   |- I/O MMU                                                       VT-d   [ ON]
   |- Version                                                     [         N/A]
   |- Hypervisor                                                           [OFF]
   |- Vendor ID                                                   [         N/A]
                                                                                
Performance Monitoring                                                          
|- Version                                                        PM       [  3]
|- Counters:          General                   Fixed                           
|                     8 x 48 bits             3 x 48 bits                       
|- Enhanced Halt State                                           C1E       <OFF>
|- C1 Auto Demotion                                              C1A       < ON>
|- C3 Auto Demotion                                              C3A       < ON>
|- C1 UnDemotion                                                 C1U       < ON>
|- C3 UnDemotion                                                 C3U       < ON>
|- C6 Core Demotion                                              CC6       <OFF>
|- C6 Module Demotion                                            MC6       <OFF>
|- Legacy Frequency ID control                                   FID       [OFF]
|- Legacy Voltage ID control                                     VID       [OFF]
|- P-State Hardware Coordination Feedback                MPERF/APERF       [ ON]
|- Hardware-Controlled Performance States                        HWP       [OFF]
|- Hardware Duty Cycling                                         HDC       [OFF]
|- Package C-States                                                             
   |- Configuration Control                                   CONFIG   [   LOCK]
   |- Lowest C-State                                           LIMIT   <    C10>
   |- I/O MWAIT Redirection                                  IOMWAIT   < Enable>
   |- Max C-State Inclusion                                    RANGE   <     C3>
|- Core C-States                                                                
   |- C-States Base Address                                      BAR   [ 0x1814]
|- MONITOR/MWAIT                                                                
   |- State index:    #0    #1    #2    #3    #4    #5    #6    #7              
   |- Sub C-State:     0     2     1     2     4     1     1     1              
|- Core Cycles                                                         [Capable]
|- Instructions Retired                                                [Capable]
|- Reference Cycles                                                    [Capable]
|- Last Level Cache References                                         [Capable]
|- Last Level Cache Misses                                             [Capable]
|- Branch Instructions Retired                                         [Capable]
|- Branch Mispredicts Retired                                          [Capable]
                                                                                
Power, Current & Thermal                                                        
|- Clock Modulation                                             ODCM   <Disable>
   |- DutyCycle                                                        [  0.00%]
|- Power Management                                         PWR MGMT   [   LOCK]
   |- Energy Policy                                        Bias Hint   <      5>
   |- Energy Policy                                          HWP EPP   [      0]
|- Junction Temperature                                        TjMax   [ 0:100C]
|- Digital Thermal Sensor                                        DTS   [Capable]
|- Power Limit Notification                                      PLN   [Capable]
|- Package Thermal Management                                    PTM   [Capable]
|- Thermal Monitor 1                                             TM1   [ Enable]
|- Thermal Monitor 2                                             TM2   [Capable]
|- Thermal Design Power                                          TDP   [   15 W]
   |- Minimum Power                                              Min   [Missing]
   |- Maximum Power                                              Max   [Missing]
|- Thermal Design Power                                      Package   < Enable>
   |- Power Limit (28 sec)                                       PL1   <   15 W>
   |- Power Limit ( 0 sec)                                       PL2   <   25 W>
|- Thermal Design Power                                         Core   <Disable>
   |- Power Limit                                                PL1   [Missing]
|- Thermal Design Power                                       Uncore   <Disable>
   |- Power Limit                                                PL1   [Missing]
|- Thermal Design Power                                         DRAM   <Disable>
   |- Power Limit                                                PL1   [Missing]
|- Thermal Design Power                                     Platform   <Disable>
   |- Power Limit                                                PL1   [Missing]
   |- Power Limit                                                PL2   [Missing]
|- Electrical Design Current                                     EDC   [Missing]
|- Thermal Design Current                                        TDC   [Missing]
|- Units                                                                        
   |- Power                                               watt   [  0.125000000]
   |- Energy                                             joule   [  0.000061035]
   |- Window                                            second   [  0.000976562]

---

STABILITY: So far stable now with 1 VM running using about 50% of the Host's CPU and a lot of I/O over NFS, with many open Internet connections (both TCP and UDP). Host has a Bridged network for VMs , so directly to LAN file server without any overhead. Zero bad stuff in log files nor any unexpected memory usage.

---

Well, I've just seen improvements in CoreFreq for X570 👍 👍 like I rarely wander into the more complex submenus, but e.g. I saw some weeks back my 3950X now shows Junction Temperature (max 95 Celcius), just like on my Intel NUC. So great stuff. Now running it all through DKMS + systemd too. Just modded the systemd service file a little. Remember there was a time the Memory Controller wasn't even available (greyed out)? 😃

Well, now it shows timers, capacity, slots etc. And well, as for Intel, they better come up with something new, because both on desktop and server (expect maybe super duper Enterprise), they are being totally crushed by AMD just re-using their i9 platform and their latest are slower than the i9-9900K for a lot of stuff, so. Gotta OC to get close to AMD when that shouldn't be necessary. Of course I hope they do. Competition is good. Although, I enjoy having chosen the best HW and as expected it lasts and is still at the very top, including the NVMes etc.

[cyring]

Very kind of you to provide these results.
Don't remember the order, but I'm fixing, optimizing, adding new features, and worst: refactoring, the way ideas come. Thus Intel HOT and AMD UMC happened among them.
Your NUC has VT-d; I should be able to program basic I/O MMU features, starting with its version structure.
Too much ideas, not enough time, you know the story.

Thanks again for providing me those outputs. I feel reassured I have not crashed your CPU.
But I'm not aware about the CoreFreq User experience with the many other processors.
I'm reading twice, many times, specs and datasheet, but oftenly I'm crossing fingers when pushing a new commit that it won't hurt CPUs, even if I simulate in my head how they will follow my instructions call flow.
A few weeks a go I had a bad time with Goldmont and User has been patient enough to let me program a full new loop for this architecture. But contribution stopped before Vcore implementation.
What else could I say about garage development & open-source ? It's just for the fun. 🤣

EDIT: Btw, have you successfully altered the PL1 or PL2 value ? Did you notice a change in max frequency or temperature or voltage core when limiting the power with lower offsets ?

[olejon]

Yeah well the HOT stuff doesn't matter as long as it's only happening where it "should" anyways, like that Intel CPU on the NUC. But no surprise the MacBook goes as close to 100 Celsius as possible before activating the fans. And hence activated HOT(s). But you rarely notice throttling. CoreFreq just shows it boots as fast as possible at the cost of getting pretty damn hot in there while keeping the fans silent as long as possible. Makes sense for an Air. Just when battery below 5%, if on macOS. Can't be turned off.

Oh, yeah I know 😄 Though, IMHO CoreFreq has come to a level where it the stable should be "markedet" rather than spending all times on development. I know it's boring, but seeing thousands of new users is fun (though that requires e.g. a voluntary "Send anonymous system data" question at first launch, in this case.

And by marketing, nothing would be more effective than getting it into the Debian repositories, or the Ubuntu ones if they are super idiots (the Debian ppl can be "difficult"). But hey, it's all FOSS, so. What gets into Debian gets into Ubuntu (however unlikely it'll be Canonical supported, but hey that's for Enterprise paying for it).

With the right description, containing all the keywords people would search for, yeah, it'll find its way to thousands of people. Maybe add a subreddit for discussion (making clear it isn't a support subreddit and that issues and code suggestions is for GitHub).

However, you might reach even more people by making it a Snap package. Snapd is supported on every major distro, and has plugins for basically all GUI software centers (as does flatpak). Just look at the stats on that website for various distros. You'll see e.g. VLC, Spotify etc are mainly downloaded as Snaps. Snaps has an even more flexible permission model than Android apps. Although, for this it probably required the same as say Visual Studio, very well supported by Microsoft there with excellent reviews for making FOSS apps (proprietary it is) and Android Studio (Open Source). IDEs and such, easily get accepted getting full system access. Your may not need full file system access. I haven't read all the Snap permissions docs. But say VLC is actually completely contained. Can read files from anywhere, and use Qt/GTK themes but that's it. E.g. saving from transcoding I'm pretty sure can only be saved in its "container" (each Snap is mounted as a separate mount point). Though I use the repo one as it's enough up to date for me, but say Windows users would prefer an always up-to-date VLC without having to even update it. It just happens (checked some ~4 times a day and the user doesn't have to do anything).

I mean, it's way more powerful than any CPU tool found in official distro repos today.

[olejon]

BTW can make my mom turn on my Monster Laptop when home again. Has the the first gen Intel Core released after Core2Duo, though Core i7, and fast as heck still. It's set to static IP and will be accessible through SSH ASAP as booted. So can check there too.

[olejon]

Oh, OK, I now got the Develop Branch one to also Grey Out the HOT indicator for 1 second, than it comes back on. As said, no biggie. It really should be on as I understood it. And it made no difference before anyways.

My NUC crashes like a couple of times a year. Always when I'm NOT logged in to it (meaning no SSH). It just freezes, but doesn't go total fan crazy for instance. I only notice it because it's my LAN's main DNS server (with reverse DNS for every device), so when devices must fall back to DNS server 2 (OpenDNS), you definitely notice it. Or if the first thing I do that day is to log into it. No contact. Logs show nothing, so it just froze. And FYI CoreFreq doesn't run when I don't actively use it so it's not that. I remember once you made one of my systems crash as soon as I loaded corefreqk.ko with a small code update (don't remember if NUC or Ryzen), but you fixed that immediately, so.

Those freezes might have been an unstable SATA SSD. As gradually I noticed more and more that if a VM had heavy I/O (the host never has so...), then the Host's log showed the SATA link being reset very often with a bunch of debug info I don't understand. Now I've taken the Corsair that was in my workstation for Windows, fully firmware updated, and put it in the NUC instead, and Zero SATA problems, so luckily not a Motherboard failure. Though, it may have crashed once since that too, typically in the middle of the night. Not sure. Time will tell. Gotta put that Kingston into my workstation and update the firmware (Windows required of course zZz). Maybe it was that + never OS. But I don't know where to use it anyways, my workstation is full with 18 TB disks. Oh yeah, my Monster Laptop. Yup. That needs one.

My 3950X system with the non-gaming Pro MB, has NEVER CRASHED. It's pretty incredible, considering all it's "been through" and has to run every day. Dead stable. Well also on Windows but that's like 2 hours every month. And I do update the BIOS to the latest stable version. Each time. No problems yet. And when those "IRQ errors" went away, it has the exact same "warnings/notices" during boot as any Intel system I've used. And it's faster than Windows. Maybe it uses more Watts in total? Haven't bothered to check with my wattmeter. But certainly not under load. k10temp shows GPU info, and Ryzen Master shows the same as CoreFreq (and when idle a little more MHz as you've pointed out). So I'm not desperately waiting for a pstate and idle driver. More because they allow so many parameters on Linux, rather than performance. Very, very well documented is intel_pstate, for sure.

BTW, regarding Pentium CPUs, are you aware that the cheapest Intel NUCs come with a Pentium Silver? (if not Gold, don't remember). Not just that, here in Norway the version that supports BOTH a SATA SSD and an M.2 NVMe is actually cheaper than the one only supporting an NVMe. It's just a little taller. Like mine. Back then NVMEs didn't exist so had to buy a special Intel chip if not buying a SATA model and those chips/disks had bad reviews so nopes.

I assume you have a SATA SSD or an NVMe lying around + a laptop RAM chip? If so you got what the NUC needs. Wi-Fi + Bluetooth module is optional of course. It's a pretty good deal then. IDK how well a Pentium performs. It's like, for me, "they still exist??". Well, apparently. Probably cheapest option to check a Pentium with CoreFreq, and after that have a decent server 😺

[cyring]

Multiple discussions above, but about HOT indicator if it's coming back, may be there's a persistent thermal issue logged in Processor event bits ?

[olejon]

Usually it doesn't turn off at all. Just once in a random blue moon does it turn off briefly. However, as you explained like in 2019/2020, all of these older Intel Ultra Low TDP CPUs activate "Power Limitation" on boot. Probably for some reason or another. 🤷‍♂️

However, it MIGHT be a result of me deactivating Hyperthreading in BIOS.... Just a guess. As I've never bothered to toggle that HOT indicator off after you said it's normal. And toggling it off obviously had no effect apart from the good feeling of not seeing HOT 😺

But I won't bother to activate Hyperthreading again. It's headless and then I need HDMI + keyboard + downtime. It runs BIND9 (DNS for my LAN including reverse lookups - PTR) and Squid + squidGuard (Proxy for my LAN with adblocking and tracker blocking, blocked and discarded before those even reach devices, much more efficient than adblocker extensions as they use quite a lot RAM and CPU to filter them out + it works in browsers not supporting extensions...), so.

IDK exactly what to make of the phrasing "Power Limitation", but the NUC Package Watt usage is BELOW 1 Watt (!) when idle and just slightly above 1 Watt when running idle VMs.

And the temperature with Hyperthreading off never reaches 55 Celsius. With it on however it can reach like 65 after a good while. But the main difference is idle temperature is somewhat lower and it takes way longer time to make it reach say 50 Celsius.

I guess you have the HW to check that the HOT code works as it should, so... Or? I only got my MacBook Air i7 to test that. Those Airs trigger minimum 2 at boot and like all of them minus one with usage. But that's normal.

Apple do a very good job at letting you get a fast experience without the fans ramping up. Most tasks people do on such laptops are short lived, so the CPU is allowed to quickly reach 90+ Celsius without ramping up the fans to an audible level, and rather throttle a little bit (if task takes more than X time) so the temp quickly goes down after say launching Chrome with a bunch of extensions and tabs. However, longer running tasks and games (CPU + iGPU) it seems to be configured to understand that you want max speed and rather let the fans ramp up. Like a benchmark, compiling, gaming, etc.

Most of this lives in the SMC or the NVRAM or whatever, as a bunch of the low level stuff (which I always clear after a macOS update as well, why not) but it seems like the macOS Intel driver does some extra trickery, as I can never get macOS completely idle. Just the Task Manager uses too much, even Apples top. But usually something's running, although I've disabled everything that do stuff in the background (e.g. Spotlight).

Yet macOS keeps quieter than Linux which I definitely get completely idle, but Linux is much more likely to ramp up the fans. Using the Intel Power Gadget app (Windows and macOS only, but it only shows values and pretty messy too, so who cares), shows the macOS throttling situation and compared to say CoreFreq on Fedora on that MacBook it shows Linux basically doesn't have this "don't ramp up fans if short task, rather give the pleasure of plenty fast enough speed + much longer battery life".

That CPU wasn't exactly meant for such a thin laptop (they failed with the Air 2011 model - it was so noisy even iSheep girls in my class hated it - but Apple totally succeeded with the mid-summer 2013 model - and MacBook sales has dropped since 2015 so...).

Reason I bought it was that absolutely no laptop at the time was even CLOSE to that performance + minimum 12h battery life, that with a simple update got up to 15h. Ideal when starting as a student. Practically everyone had it. Used it a lot for Android Studio too. Still could code and compile for hours and basically never thought about the battery before hitting 5% where it throttles like crazy so you get some extra time to finish your work I guess, but the throttling is so heavy even basic web browsing sucks. I think it happens in Linux too but it was maybe there I could disable all the HOT indicators and it would go back to normal speed. Don't quite remember.

But yeah, that's the HW I have if you want to know if there is some bug on the HOT indicators code because it's like possible to activate all of them (minus one IIRC). Dunno if you're interested in the PL output from that one too? My guess is the same result as the NUC. Both are Intel 2013 Ultra Low TDP CPU's so... Haven't got Fedora installed now (with latest macOS I decided on a clean install from recovery mode after the upgrade, it still gets latest macOS!), but it can work just fine with a USB stick and an Android phone using USB tethering (Network Manager notice that immediately and connect through your phone, sweet when Wi-Fi doesn't work out of the box). Maybe Ubuntu come with that Broadcom driver, but for Fedora you need RPM fusion to get the Wi-Fi driver (and possibly reboot meaning USB stick won't work).

[olejon]

So yeah, TL;DR I can check my MacBook for the PL values of you want to (although both it and the NUC have 2013 Ultra Low TDP CPUs, difference is i3 vs i7 mainly - maybe the MacBook's CPU is one generation newer - IIRC). And also if the HOT stuff act as before on the MacBook as that is the only HW I have to test HOT triggers/toggles on and, well, last time I used CoreFreq on it (2019-2020) it is ideal as it can trigger all but one (IIRC), and toggle them off again, as long as I start with the right one.

[cyring]

Probably BD_PROCHOT

Could you read this register:

modprobe msr
rdmsr -ax 0x1fc

This undocumented bit should be at position zero
If set then clear it by and-ing the above output value with 0xfffffffffffffffe

Write back the result VALUE into MSR POWER CONTROL then read again that it has been acknowledged

wrmsr -a 0x1fc 0xVALUE

rdmsr -ax 0x1fc

Check CoreFreq for any HOT improvement while setting or clearing BD_PROCHOT bit

Note: this thermal bit and other SKL events will be soon part of CoreFreq

[olejon]

As said, I don't care, TBH. Not sure if I did it correctly, the write part, as I get the same value afterwards. This right?

sudo modprobe msr
sudo rdmsr -ax 0x1fc
4005d
4005d
sudo wrmsr -a 0x1fc 0x4005d
sudo rdmsr -ax 0x1fc
4005d
4005d

And no, no change in CoreFreq (develop). Still can't toggle off Power Limitation. We're not talking the PROCHOT# Agent here. That one is never on. On my MacBook however, it did trigger and was the first I had to clear to clear the rest, or the last. Don't quite remember. Or had to do it twice.

[cyring]

As said, I don't care, TBH. Not sure if I did it correctly, the write part, as I get the same value afterwards. This right?

sudo modprobe msr
sudo rdmsr -ax 0x1fc
4005d
4005d
sudo wrmsr -a 0x1fc 0x4005d
sudo rdmsr -ax 0x1fc
4005d
4005d

And no, no change in CoreFreq (develop). Still can't toggle off Power Limitation. We're not talking the PROCHOT# Agent here. That one is never on. On my MacBook however, it did trigger and was the first I had to clear to clear the rest, or the last. Don't quite remember. Or had to do it twice.

4005d = 1000000000001011101

Flip first bit

1000000000001011100 = 4005c

wrmsr -a 0x1fc 0x4005c
rdmsr -ax 0x1fc

[olejon]

Maybe you meant this?

sudo wrmsr -a 0x1fc 0x0xfffffffffffffffe
sudo rdmsr -ax 0x1fc
0
0

? Still see no change in CoreFreq afterwards (also checked unloading daemon + module and loading those again).

[cyring]

Maybe you meant this?

sudo wrmsr -a 0x1fc 0x0xfffffffffffffffe
sudo rdmsr -ax 0x1fc
0
0

? Still see no change in CoreFreq afterwards (also checked unloading daemon + module and loading those again).

Try 0x4005c

[olejon]

So I did:

sudo rdmsr -ax 0x1fc
4005c
4005c

Still can't toggle off Power Limitation. I guess you're clear that's what I could before but can't now right? Well seems to interest you more than me. When this bit is set on boot anyways, it would be e.g. set in Windows too right?

[cyring]

So I did:

sudo rdmsr -ax 0x1fc
4005c
4005c

Still can't toggle off Power Limitation. I guess you're clear that's what I could before but can't now right? Well seems to interest you more than me. When this bit is set on boot anyways, it would be e.g. set in Windows too right?

Windows ThrottleStop is alerting this bit

[olejon]

Yeah so no change with 4005c. Do you consider this a bug?

FYI: A few days ago an update to the package intel-microcode came. Of course I updated as any would do. Since I've also tried with the Master branch now and it's the same situation there (can't be toggled off), and since it's been months and months since I last actually toggled it off because why? Does nothing.

So... My guesses: Kernel and/or Microcode updates since months back, the fact that Hyperthreading = DISABLED in BIOS and maybe that confuses your code, if it can even see that it really has Hyperthreading capabilities when it's off (no software shows it), or some CoreFreq code change, but if so to the Master branch as well as it's as said the same situation there.

Maybe this is how Intel wants it? I'm not sure if the words "Power limitation" are chosen by you or are specified in one of:

1. Intel docs
2. Kernel docs

And what it actually means, yeah you still really haven't told me.

But, hey, I'm totally guessing here but considering this bit is ALWAYS set on boot, maybe Intel has made a change so that one can't toggle it off because having it on actually serves some purpose, e.g. telling kernel/system/programs that "Hey I am a Ultra Low TDP so don't expect too much mmmkay?" 😃

[olejon]

Windows ThrottleStop is alerting this bit

Okay, had to google that. I guess it reports it's on after boot on these CPU's right? And turning it off might improve performance by reducing throttling? As it's called ThrottleStop. However, my NUC does ZERO throttling. Always 1.7 GHz on all threads no matter temperature.

So why would it matter on this PC? Maybe on a MacBook with Linux, yes... If it does reduce throttling. My experience is that Linux does that anyways. But at a big expense of noisy fans and way worse battery life. Can still check my MacBook another day and see if I can toggle it off on that one.

[cyring]

This is an experimentation of presumed BD_PROCHOT to either remove throttling introduced by components surrounding and notifying Processor to decrease its operating state: could be target p-state, frequency id or other limitations. Sorry no datasheet.

Thus you have to monitor, stress, observe any thermal events while toggling on and off and on ... this bit.

Purpose was to explain why CoreFreq is showing HOT a few sec again, as you mentioned previously.

Hope it helps.

[olejon]

Aha. Well, it does in no way throttle. It can run a full CLI Blender Benchmark and not throttle. Both before when this could be toggled off, and now. Same result.

showing HOT a few sec again, as you mentioned previously

The other way around. HOT is always on (colored). Mostly the HOT never changes from colored to gray.

Some random times it might show HOT with a gray color after toggling it off but it becomes colored again within 1 second (and that's for sure because of the default UI refresh time).

Probably happens when the UI refreshes right right after I've toggled it off (and then stays gray until it refreshes again). If I had decreased the UI update time to a much lower value I guess it would've been gray for much less than 1 second. Barely visible, basically.

[cyring]

I remember that in this commit 0ac6c83 I made a change to allow clearing thermal event bits if and only if a thermal interrupt handler has not been installed.

Although the initial feature was operational with SNB and above, including your Haswell, I encountered a lock issue with Coffee Lake. And investigations revealed some conflict within Kernel handler.

The fix has been to check MSR_IA32_THERM_INTERRUPT before attempting to clear MSR_IA32_THERM_STATUS

nomce prevents kernel to boot without such handler

[cyring]

Edit: nomce is the kernel boot argument.
See the issue #225 for detail.

Edit: as a side note for my roadmap. UI should pop a message box up that clearing operation is not safe, and thus refused.

[cyring]

More easy than nomce is to unload or prevent from loading the Intel thermal kernel module, because of these Interrupt bit enablements

https://elixir.bootlin.com/linux/v5.12.5/source/drivers/thermal/intel/therm_throt.c#L679