The ThrottleStop Guide

Just copy the ini file of expired TS to new obsidian version.

 

First of all, thank you for making this great tool. I like it far more than the Intel XTU.

I can fully understand your position, but seriously, who do they stop with this ? If AMD or any other big player wants to reverse engineer their CPUs, they just will to this. There is no reason to hold this information back.

Anyway, this lead me to

Windows annoyed me again and I want to switch to linux finally but I like my undervolting settings.
So for everyone interested in undervolting in linux, I collected some ideas:
http://forum.notebookreview.com/threads/undervolting-e-g-skylake-in-linux.807953/

 

Which version of Windows 10?

 

Ah, yes, it's Windows 10 (1703 to be precise)

 

Its quite known issue. One way is to fix it is install Anniversary Edition RS1.

 

Ehm ... ok, seems to be a fix for many things ^^

I didn't even write the details yet, I just wrote Windows annoyed me again. (In detail there are a lot of small reasons, including that every app and EVEN DRIVERS want to install their permanently running updaters&telemetry.)
I think the discussion about windows drifts off topic here a bit ...
I just needed a way to contact @359RdrMb2Jee8h7K (because I couldn't message him) and link interested users to my post XD

But thanks to your fast replies

 

Hahaa. I thought someone hacked Windows 10 and telemetry can be killed completely using OOSP 10, Winaero Tweaker etc..

 

I'm sitting here with a laptop of mine that has an i7 4600u, trying to get the best battery life of it. I looked at the c-states and what puzzled me is that it doesn't seem like it utilizes C-states below C2? Atleast according to 'Package C State Percent'. I've tried looking at even on a clean install of Windows with no programs running, without having AC connected and no settings changed in ThrottleStop.



Maybe this is how it should be? The reason I'm wondering is because my battery times on this, allegedly, 0% wear battery is no-where near the times other people report with this laptop.

Can anythnig be done to perhaps more aggresively make the CPU use the powersaving c-states, if it isn't already?

 

@Kocane - Not sure why but some manufacturers have disabled the deeper package C States. Your individual CPU cores are using C7 like they should be but it looks like the entire CPU package is being limited to C2. I have a desktop Asus board that does the same thing. I am using Windows 7 on that computer so I will not blame Asus until I get around to installing Windows 10.

A low power U CPU should be using at least package C8. Intel designs these power saving features but they cannot force individual manufacturers to use them. The average Joe does not have the smarts to troubleshoot this problem. If you contact tech support they will probably try to blow off your question and tell you that they do not support third party monitoring software. For the record, ThrottleStop works fine at reporting the C States.

 

Not sure how many would benefit since my issue was with an Ivy Bridge 3940XM I just got (and overclocking newer CPU gens works differently) but this workaround I found may be of help to people trying to avoid using the XTU GUI for the things TS can't do - by using the command line XTU client I was able to get things working automatically at startup - AND ON RESUME - instead of having to manually use the XTU GUI.



I finally found a way around an annoying cTDP/turbo power limit issue.

With the 3720QM/3740QM on both my P170EM and in this P370EM previously, I would just set the short and long power limits way high in XTU and they set and stayed there, persisting reboots and everything (until a CMOS clear)

As soon as I put the 3940XM in, on boot, it kept on setting cTDP profile 2 for a 81.25W short and 65W long turbo limit. There are zero BIOS overclocking options enabled in the BIOS (Premamod 1.02.09/1.02.08). XTU would set the multis correct but the power limits would not be applied on startup - to change the turbo power limits I had to do it manually through the XTU GUI, which would try and tell me the power limits in the profile were currently loaded, lucky I had HWInfo to tell me otherwise (and clocks would drop to as low as 3.7ghz on an extended bench with the 65W long power limit). Unacceptable.

Changing the power/turbo settings on the TPL page in Throttlestop wouldn't change anything at all. I have it set to run at startup via Task Scheduler so it sets the turbo multipliers (x43) as normal.

I found the command line way to use XTU - XTUCli.exe - and using a short Powershell script, it runs automatically on startup and sets the Turbo Long power limit to 95W and the short to 90W:

Here's the script: (adapted from https://www.reddit.com/r/Surface/comments/3vslko/change_cpu_voltage_offset_with_intel_xtu_on/)

Code:

$status = get-service -name "XTU3SERVICE" | Select-Object {$_.status} | format-wide
if ($status -ne "Running") { start-service -name "XTU3SERVICE"}
& 'C:\Program Files (x86)\Intel\Intel(R) Extreme Tuning Utility\Client\XTUCli.exe' -t -id 48 -v 90.0
& 'C:\Program Files (x86)\Intel\Intel(R) Extreme Tuning Utility\Client\XTUCli.exe' -t -id 47 -v 95.0
sleep 5
stop-process -id $PID -force

Put it in a file, put it somewhere on your OS drive (N.B.I had trouble getting it to run when I had spaces in the file path)

Set it to run via Task Scheduler (very similar to the standard Throttlestop task scheduler setup)
General tab:
Run only when user is logged on
Run with Highest privileges
Triggers tab:
1) At log on of any user; Delay 20 seconds
2) On Event triggers for resuming from sleep/hibernate. This is a WIP right now, I haven't found a single reliable trigger here yet; my laptop seems to be erratic as to which "event IDs" are triggered on resume. (To see which event is in the logs, run Event Viewer and look under Windows Logs->System and look for the timestamp to find the combination of Log/Source/EventID to use in the Task). I'm trying the kernel-power/boot one now. But the important thing is that when an event occurs that has an event trigger set up for it, the script runs as desired.

Actions:
Program/script: powershell.exe
Add arguments: -ExecutionPolicy Bypass -Windowstyle Hidden c:\pathtoyourscriptfile\filename.ps1 [the bypass executionpolicy switch may help solve administrator issues; windowstyle hidden means the blue box flashes up for only a fraction of a second when it runs]
Conditions: all unchecked
Settings: all unchecked; tick 'allow run on demand' for testing the script

The startup delay is to give the system time to start the XTU services, and also give a window to stop the script if somehow it starts borking up (unlikely with power limits but that's a possibility if you set an undervolt or a typo which BSODs as soon as it's done). A blue box flashes up for a fraction of a second and it's done.

You could automatically set your whole XTU profile to load using these commands if you wanted, but of course TS is the preferred program for everything that it is possible to do (much lower background resource usage, hotkey support, etc).

Open cmd.exe as admin and navigate to your XTU install directory (default C:\Program Files (x86)\Intel\Intel(R) Extreme Tuning Utility) and under " Client" subdirectory you'll find XTUCli.exe
Run " XTUCli -i ALL" to output a list of all the parameters you can alter on your system. The ID of the parameter is what you're after - in my case 47 for turbo power short and 48 for turbo power long. I could set them to crazy high numbers but 90W is enough for now as the highest I've seen on this CPU was 88W at x45/1.4V and it couldn't handle that heat for very long at all and 1.4V is getting up there for Ivy Bridge.

 

No wonder my battery life even on idling is inferior. Like said, people in reviews have reported up to 12 hours of idling on this laptop. I haven't done a real test but I find windows estimates accurate enough and it reports ca. 5-6 hours on lowest brightness, WLAN disable etc. I've even done a clean install of Windows 10 on another partition, to rule out running programs/services but it's the same thing. It's an Elitebook 840 G1. I doubt HP would've done anything to limit the CPU power saving. Giving my work laptop, an Elitebook 850 g1 (same CPU, and generally just same hardware but 15" over 14" and even same version of Windows 10) it reports that atleast some other c-states are in use, even when having a ton of programs running, and being docked:



Do you have any ideas? There's little to know help to get anywhere regarding such an issue. It's so critical for batterylife but doesn't seem like anyone much care for :/

 

I just took out my two SSDs and inserted a HDD, installed a clean Windows 7. Resettet my bios (not that the darned thing have any options) and I still cant use anything above C2. It's killing my battery life. Jesus what a ridiculous thing to stop functioning.

 

Tell us your CPU model and battery capacity in Wh or mAh.

 

i7-4600u, battery ca. 4500mah. But I know from previous experience how much the c-states mean for powersaving. I can compare my own laptop with the several ones we have at work and see the difference.

 

Any HP utilities installed or running currently? Are you on latest BIOS?
EDIT: Optionally check your microcode revision in HWINFO > CPU > Microcode Revision
Try applying the latest microcode update through here Skylake / Kaby Lake Hyper-threading bug

 

Thanks for replying, mate.

I have some interesting development though. I took my work laptop (850 g1, the 15" version of my laptop) and put my temporary Win7 HDD in it (no drivers installed at all, I figure(d) that the CPU shouldn't care and use the power saving states regardless) but on that laptop, which normally uses C6, it didn't enter anything but C2 either. Then I put my Win10 SSD for the work laptop into my personal 840 g1, and it did indeed use the c-states... i've tried to compare drivers between the two installs but I really cannot find anything different?! The only thing different - that I konw of - is that the install is done with the 'legacy' mode enabled rather than uefi with crm, in the BIOS.

Any ideas?

 

Do you have Intel Rapid Storage Driver + GUI installed? If so, disable SATA link power saver and will allow deeper low power states.

 

Yeah I got it installed. Just tried disable SATA Link power saver and rebooted. No change

 

Hello.My laptop shutdown every 30-40 minutes(no bsod) with error 41 task 63.After this error i got this

Sent from my SM-G920F using Tapatalk

 

Maybe its overheating.

 

I Changed the termal paste a month ago

Sent from my SM-G920F using Tapatalk

 

Sorry for double post

 

Can you show the current wattage on battery in throttleStop or HWINFO > Smart Battery

 

Is your power plan set to Balanced/High Performance?

 

High performance

Sent from my SM-G920F using Tapatalk

 

Can you post a screenshot of current temps in TS?

 

Right now i can't because i am not at home but tems in idle are 40-55 and in game 70-90

Sent from my SM-G920F using Tapatalk

 

It seems a lot higher both in idle and full load. What's your ambient temps like?

 

Around 17 -28

Sent from my SM-G920F using Tapatalk

 

Then your temps are quite high.

 

I think I finally cracked this one.. .Its been bothering me for a few years now. The problem is with many versions of intel HD graphic driver... I installed version 20.19.15.4444 and got almost 50% more battery life..
Link: http://drivers.softpedia.com/downloadTag/Intel Graphics 20.19.15.4444 Beta

Please try it and report back..

 

Just gave that version a go before I went to work this morning. It didn't work in my case

Thanks for suggestion though,


I even installed Windows 7 clean yesterday, installed every single driver and tool from HPs Softpaq (driver repository) and it didn't change a thing... why on earth would it work with the Windows install from my work laptop? What can it possibly be in software that makes the c-states not work? Is it Throttlestop not reporting correctly? Just seems very unlikely as well..

 

Your battery capacity is 100% with no wear and tear. Okay, post a similar screenshot using battery.

 

Did you go and manually install driver - auto install didnt work for me as this is an older driver... Go into device manager->display adapters->Intel HD Graphics Family->Update driver->Browse my computer for driver software->select folder where you extracted the drivers linked above wait for it to install and restart computer.

 

Yeah I located the INF and rebooted after the succesful downgrade. No dice.'

On my work PC where the c-states work fine I inserted a different disk and did a clean install of Windows 10. It downloaded all the drivers etc, but it was exactly the same - c-states weren't being used. What the ****

 

Question for an older machine. We had an older version of ThrottleStop on an Alienware m15x. It seems that Throttlestop has expiration timers in place. We upgraded to the latest edition, and found that it doesn't seem to work. So two questions come to mind. 1st: Is there a way around the timed kill switch? 2: Do you have a long term version for machines on the 1st gen i7?

 

@WaldoJim42 - What doesn't work? There are thousands of ThrottleStop users but I have not heard about any new bugs recently. The system tray / notification area bugs can usually be solved by reading the included ReadMe file.

I have never owned any first gen hardware so I do not have any way to do any proper ThrottleStop testing. Some bugs would not surprise me.

Most bugs go unreported for days, weeks or months. By the time I hear about a bug, it takes me a long time and a lot of back tracking before I can get around to fixing them. Everyone is too busy. Me too. I hate buggy software so now I just let ThrottleStop expire on a regular basis.

If you have an older first gen Core i, try running ThrottleStop 6.00. Most new TS features are for the newer CPUs so TS 6.00 should work OK for a first gen Core i. If you have the time, explain your problem as best as you can.

 

Unfortunately, it will be another week now before I can test any further. Didn't realize my brother was taking the machine for the week on a trip.

The machine is using an i7 920XM and an NVidia 970m GPU. Both upgrades from the original build (720qm and amd m5870) - I was using the custom turbo profiles to OC the chip a bit for gaming. In particular, I was trying to use 22x 4 core active, 22x 3 core, and 24x 2 core, and 26x 1 core, with a 20x minimum while in games. In the previous version, the machine held those clocks wonderfully, and games ran perfectly. When the older version expired, and I switched to the latest version, it would no longer hold any of the multipliers. It tends to throttle down to 17x or even 15x in some games on the newest version. Though it doesn't give a reason for the throttle that I can find. 15x is the base multiplier for that chip.

I know the platform itself is somewhat power limited. That was the reason I went with the 970m instead of the 980m for the upgrade, as well as trying to hold clocks that are enough to keep smooth gameplay, rather than going for anything extreme.

 

This sounds like a problem with settings. We have an m15x, and TS has consistently worked well on it. When the machine comes back, post some screen shots so we can see what is going on. There are also lots of TS users on the m15x forum here, so you couldd check settings there. Good luck, Joe

 

About that C state thing - @Kocane

Read my post here, fiddle with the power settings, and I keep forgetting which one. Power Saver for me allows C8 package states, all other power plans lock it down to C2

http://forum.notebookreview.com/threads/the-throttlestop-guide.531329/page-624#post-10557107

I can't be bothered to do it every single time I reinstall windows, so essentially what I do is copy the Windows installation that came with the HDD by default, to my SSD. Yes, I have to deal with bloatware and upgrading to Windows insider etc but for me it is definitely worth it - I use my laptop at school, I'm paranoid about battery health, and I get about 10 hours with C8, where the CPU idles at about 0.4W. With only C3, the package uses 1.4W, which significantly drains battery life.

For you gaming laptop/desktop users out there, every single millwatt makes a difference to battery life to us 'normal laptop' users.

 

Thanks for throwing some input on this issue, I'm still trying to fix it.

Like I've mentioned, the issue exists on my work laptop as well with any other OS that I install, aside from the one that I've had installed for a while now, where C7 works. I've no idea what makes it special. I tried HP OEM Win7 and Win10 - no difference. At the moment I'm cloning the "working-c-states"-Windows to another HDD where I'm gonna try and 'reset' the Windows in order to make it clean. Then I'll see if the power states still work.

We'll see. Still open to any other suggestions as well.

 

Yes, that's what I've done. It seems to be a Windows issue, since it happens to me as well. It must be a driver, but I've literally copied the entire driverStore folder in C:\WIndows\System32 to the other non-working C8 state but that does nothing.

Perhaps it is a registry key that determines this?

What did you do to your OS that has working C7 states? How did you install it?


Next time I have to reinstall Windows, I'm going to try copying the entire Windows folder from the working C7 to the non-working installation of Windows .It'll bother me with bcd corruption and everything but that's easy to fix for me. I'll have to try it though.

It certainly is sad, nevertheless, that there is not a single thread out there that complains about not having working C8 states. It's sad that Microsoft isn't fixing a problem and forcing its users to endure battery life that is lower than accepetable.

Then again, perhaps we must have done something to muck up our installs.

 

I'm completely lost. I cloned that "resetted Windows install" from my working-c-states-laptop and it works fine in my personal laptop - when it's on an HDD. I cloned that HDD to my M2 SATA and it doesn't work there. I can switch between the two drives and see it working when using the HDD, atleast according to my testing, but at the same time it also seems like it stops working whenever the M2 sata is inserted, even when its not booted to. But I'm really not sure, it's all very flaky. But that doesn't change the fact that I've had clean installs of Windows 7 and 10 on an HDD with all drivers installed on BOTH computers where it has not been working.

I did however get a quick glance at Windows battery estimate when I had the working C7-state and it was like 8-9 hours as compared to the usual 3-4 hours. Painful that this problem is even here and even more painful is that no-one else aside from the people in this thread gives a **** about it. My googling primarily takes me to "HOW TO DISABLE C-STATES" threads and talks.

 

I'm just lost more than you now.

I can switch between SSD and HDD with no issues, but you can't? That is really weird. Is it a matter of a controller or something? We know that the issue is related to either the Cache or one of the many system controllers (ACHI, USB, etc) that are integrated into the CPU, since the CPU CORES are able to go into C7 state, but the package is stuck in C2 state. OR was it C3.

Now the fact that you mention you tried both Windows 7 and 10 and cloned them onto different machines suggests it is a hardware issue, probably drivers or something. Oh well when I have time in the holidays I'll do more testing. Rn I'm so busy at school with HW and tests etc.

 

It doesn't seem to work as long as I got my M2 SSD inserted, lol...

Windows is on a HDD and the M2 SSD only got storage stuff on it:


Without the SSD on the same HDD. (Notice the temporary 2GHz but still usage of C7)


I cloned this Windows to my regular 512 GB SSD and the c-states worked fine as long as I kept the M2 SSD out of machine. Same goes for work laptop which also lost the ability to use the power saving states as long as the M2 SSD was inserted.

However my "original" Windows 10 install from before I found out about all this - which originated on my regular Samsung SSD and was later cloned to the M2 SSD - doesn't work either, no matter what media I clone it to... and my first tests gave me c-state trouble without the M2 SSD inserted. So I still find it hard to say whats the culprit but I guess its pretty alarming that it doesn't work whenever the M2 is inserted lol...

So I'm about as lost as lost can be. Already lost trying to explain my tests... I guess I could just use the install of Windows 10, where c-states work, on my 512GB Samsung SSD alone and then ditch the 256 GB M2 SSD, but damn what a waste. It might be some China SSD without any firmware upgardes or anything but how this thing should have an affect on how the Intel CPU does power saving?! What the heck...

@unclewebb, is there nothing that can be done about such an issue from the programmering side? There's no way to aggressively enforce usage of c-states?

 

I have not made any changes to ThrottleStop recently that would change how a first gen 920XM overclocks. If one of these CPUs is dropping the multiplier while loaded, check the TDP TDC values that you are using. Also check to make sure your CPU is not overheating.

@Kocane - I do not know of any magic way to toggle the deeper package C states on and off. It could be a poorly written driver or maybe some manufacturer wanted to improve their SSD benchmark numbers so they decided to disable some of the package C States. Better numbers = more hardware sales = more $$$$$. Easy to get away with this If review sites never report C state residency time. I think HWiNFO also shows C state data but i have not tested or compared it to TS.

 

@unclewebb, may I ask a question?
What is the difference between TDC and TDP?
Searching on google shows absolutely -nothing-.

Why are there options in the Bios for 'TDC' AND for "ICCMax" (CPU Current Limit)? Aren't they redundant?

 

Intel says:

TDC is continuous and IccMax is instantaneous. The difference would be what sort of time frame TDC is measured over. An 80 amp TDC limit might not trigger any throttling as a CPU varies between 70 amps and 90 amps but an 80 amp IccMax setting would immediately trigger throttling as soon as the CPU went beyond 80 amps.

The 1st Gen Core i CPUs used a TDC setting. The latest gen CPUs mostly use IccMax instead. This is the first I have heard of both being available in a bios. If you are using an unlocked bios, sometimes there will be options available that do not actually work but it is possible that they can both be used together.

Some very old Intel specs from over a decade ago listed TDC at 85 amps and IccMax at 100 amps. If not significantly overclocking, those continue to seem like reasonable numbers to me.

TDP is for the power limits which are measured in watts. TDP and TDC are related to each other.

Power (W) = Current (A) X Volts (V)

That's why when overclocking, you can raise your power limits but you might still throttle because your current limit is not high enough. Likewise, raising only your current limit will not do anything if your reason for throttling is because your power limits are set too low.

 

Thank you Unclewebb! Best explanation on the net +5

 

Myeah, maybe. If I throw it into my other HP laptop with a regular 2,5" adapter, the M2 SSD doesn't cause C-state issues. But no matter what it gives trouble in the M2 port of my other two HP laptops. I guess I'll just stop using it... what a waste of 256gb