The ThrottleStop Guide

Afterall, i* m CPUs can't be undervolted? I have an i5 2430m that is running hot 90C+ without turboboost (under load).

 

if you have a modded BIOS for the version of your laptop yes you can. But you not just have a modded bios..,

Have you properly cleaned your laptop and have you repasted it?

 

Yes I did. In idle runs at 47celsius, and in full load over 90. That's huge. I saw that in windows 7 temperatures are lower than windows 8 lower than 8.1.
So in windows 8 in full load cpu ran to 83C. Now with 8.1 over 90C.
I didn't found a modded bios after a deep search on forums.

 

Your laptop heatsink and fan are either poorly designed or they are filthy dirty. When was the last time you took your laptop apart and cleaned it inside out?

90C is a big number but it is within the Intel Specification for your CPU. Your CPU is rated to run 100% reliably at its full rated speed at up to 100C.

Here's my desktop CPU happily running overclocked, with both the CPU and Intel GPU fully loaded at over 100C. Intel builds good stuff.

http://img27.imageshack.us/img27/6216/torturetest.png

You can undervolt the newer 4th Gen mobile CPUs but I don't think you will find a way to undervolt your 2nd Gen mobile CPU.

 

Repasting with good thermal paste can reduce temps with 5-10C°. I dont think the new version of windows makes any difference in temps, that might rather been gathered dust or changed room temperatures..

I agree with unclewebb 90C isnt extreme and is perfectly workable..

 

I used AC- MX4. Laptop is more than clean.
edit: I think 2 months ago I cleaned it...every single dust particle..

 

Did your laptop originally come with a thermal pad between the heatsink and CPU or GPU? When switching from thermal pads to thermal paste, you have to make absolutely sure that the heatsink continues to apply the same amount of pressure. In some laptop designs, it might be impossible to maintain the same pressure. A small decrease in pressure can completely change how well your heatsink works.

What software are you using to fully load and test your CPU? Has the room temperature and fan speed been equal during testing? When fully loaded, there should not be any significant difference in your core temperature no matter what operating system you are using.

What laptop model do you have? Does it use a single fan? Have you overclocked your GPU?

The reason I am asking all these questions is because the temperature changes that you are reporting are unusual. It sounds like something else is going on.

Are you using ThrottleStop to disable Turbo Boost?

 

So: Laptop has arrived without thermalpads. Just with thermal paste. It has 1 fan. this temps are obtained when gaming, nfs rivals. Room temp are no more than 20-23C. Fan is running at low rpm or 0 idling (48C) , but loaded is runing fast. There is a huge difference of rpm. I OC the gpu but no difference when is noOC. Gpu runs under 80 at 770 mhz loaded (kombustor) (672 stock).
TB is disabled from power options (99%).
There is 5C difference between cores fully loaded. With prime95 loaded tems are 68 C0, 73C1.

 

Just as a heads up the heat pipe could be done with as well. Assuming all else working properly such as the fans etc. then if the liquid in the heat pipe(s) has leached out then they will not work properly. If this is the case only a new heat pipe assembly will work for you. A new assembly may not be the cure as the chip itself could be a defective one but your issues sound IMHO like a bad heat pipe.

 

What liquid from heatpipe are you talking about?? There is no liquid in no heatpipes.

 

Heat pipe - Wikipedia, the free encyclopedia

 

Didn't knew that. Sry. Thanks for info.

 

Hi i just download throttlestop and maybe someone can help me witch settings i should use? I have i7 2630qm and i dont want to damage my cpu by not knowing what im doing. So if someone also using i7 2630qm or knows everything about throttlestop maybe you can help me with settings. I want to get better performance in the games bicose im getting fps drop bicose cpu throttle i dont want to make cpu work on 100% but i want to work it smooth.

 

This shows tablet mode running TDP level 1. You could try either 0 or 2 instead and to be sure it sticks lock it at the same time. Unfortunately my 4700MQ supports 0 levels of TDP so can not do any testing for this.

 

So i did set multiplier 22 and mark it was working well but after 3-5mins in game pc shut down the logs showing the temp was 96C i7263qm. So i can do something about it? I was thinking to open my laptop and clear the dust but its imposible http://www.youtube.com/watch?v=maoRU9NhyJo also i have cooling pad but its not helping at all....

 

Edit : My apologies Dufus. I was hoping that you were working on a tool to solve the TDP issue in the low power 3rd and 4th Gen UM CPUs. It's one issue that ThrottleStop is still not able to solve.

MNSTR0 - Maybe blow some compressed air through your laptop to knock out the dust bunnies.

 

Yeasterday i tried run at 20multiplier and play bf3 after about 2mins in game my laptop shut down but the log is showing average temp was 86C max was 93C once, but the power is strange from 5 to 29 so maybe my pc shut down not bicose of temps but power adapter? Is it possible to change power settings in throttlestop? And witch settings should i use? This is the log http://www.filedropper.com/2013-12-09 . And should i mark BD PROCHOT? Im marked EIST C1E and Set Multiplier.

 

Hi Unclewebb,

both disabling real temp and running throttlestop, and running throttlestop delayed after real temp solved my problem of trying to run the two together, thanks for the advice!

As per dufus findings, I concur - my P states were also greyed out, I was hoping when I hit the TDP trip I would be able to force the 22/24 non turbo speed but, I guess my laptops (m/board?) doesn't support it. After reading:

AnandTech | AMD's Radeon HD 5870: Bringing About the Next Generation Of GPUs

I'm starting to come to the conclusion that perhaps, unless i'm willing to always cap 4 active cores at about a 22x/23x multiplier, I should just accept that the system is ok, and I'm putting too much load on it? The only time I witnessed this TDP trip in game, actually turned out the cpu was hitting TJmax.

After a drawn out battle with the manufacturer of my laptop re. their TIM application, they sent me some MX4. I'd been using MX2 but thought the 4 would be better. Long and short - I couldn't get it to spread under compression of the heatsink - it's an 'X' shaped plate with a sprung screw in each corner. after about 3 repastes, it was evident the paste just wasn't getting spread. I was getting about 20c higher idle temps and TJmax was being reached on load. As soon as I went back to the MX2, everything has been fine again. Perhaps as unclewebb has mentioned, the heatsink assembly is designed with the use of a thermal pad in mind, hence when cleaned up and using a paste, has very little clamping force left available.
My advice would be to look at the outright, on paper specs as secondary, and focus on finding the easiest to apply paste from group reviews

 

Hello to everybody. I got a lenovo y510p with nvidia sli gt750. How to enable turbo boost to the processor?

 

Do somebody has the laptop with i7-4702mq and sli gt750 settings for best performance with the throttle stop? I am little confuse what settings are best for maximum performance. Thanks in advance.

 

I am trying to write a script which will enable me to get similar results in Linux.

However, I can not find the msr register for the "Core Current Limit" and especially "Processor Current Limit" on Haswell...

 

This should help.

 

Thanks for taking a look at it. I suspected as much from what unclewebb wrote about TDP levels. I've tried locking it to level 0, 1, and 2, but the result is the same - it seems to always revert to level 1 when in tablet mode.

I've attached some test results for each of the three TDP levels. I hope it's useful. If a tool could be made that locked these U/UM CPUs into TDP level 2, that would be awesome. Here's hoping!

View attachment MSRList_TDP0-2.zip

 

Quibbler2412 - Thanks for the data. MSR 0x64C contains the MAX_NON_TURBO_RATIO in the lower 8 bits. When the CPU switches to Level 1, this register gets changed to 0x07 which limits the multiplier to the minimum value which is 8.

Code:

0x64C	00000000:00000007  00000000:00000007  00000000:00000007  00000000:00000007

In Level 0 or Level 2, this register contains 0xFF in the lower 8 bits which is the maximum value so this register no longer limits the multiplier.

Code:

0x64C	00000000:000000FF  00000000:000000FF  00000000:000000FF  00000000:000000FF

This is worth looking into further. This register also contains a lock bit in bit[31] which is not set in your example. I think the best thing to try would be to lock this register at 0xFF.

One of the worst parts about the CPU being in Level 1 is the drop down to the 8 multiplier and being stuck at about 798 MHz. Getting the multiplier back up to the default level would not totally cure things but if this worked, it would be a big step in the right direction compared to 798 MHz. I will see if I can do some testing this weekend and if the testing looks promising, I will add this feature to the next version of ThrottleStop so this can be tested further.

 

Quibbler2412, have you tried going to tablet mode and changing the TDP level to 0 or 2 AND setting the TDP lock?

If TS has the option to restore values of TDP level values through profiles you could do that as well. (Kevin?)


Kevin (Unclewebb), 64c is the minimum turbo activation ratio so by setting a value of 7 means turbo will always be invoked. Of course if hitting TDP limits expect the usual clock down. For instance using a value of say 0x10 (16) would mean requesting multi's from 8 to 16 should result in those multi's but 17 and above will request max usable turbo. Using 0xFF means all requested multi's should be 1 to 1 up to the maximum allowed turbo ratio. Hope that makes sense.


Also FYI seems XTU has a bug when changing Haswell voltages from other software in that XTU does not update the values shown although they are clearly changed.

Core offset voltage set to -100mV via XTU. Ive crudely added some of the offset values to the C-State monitoring program for reference as this also shows core voltages. Multi was set to 8x to keep things simple.

Here you can see VID showing vcore between 0.575V and 0.580V

Now using other software to change the voltages, in this case core offset voltage from -101mV to -1mV.

XTU still shows the old values it set, even if it is closed and re-opened. However you can clearly see VID now showing Vcore between 0.676 and 0.680V as expected with a core offset change from -0.101V to -0.001V.

Just something to keep in mind when adding CPU voltage adjustments for Haswell to TS.

 

:thumbsup: Dufus for your hard work learning about Haswell's offset voltage feature. This is definitely something I would like to add to ThrottleStop. Maybe next year when my new Haswell toy arrives I can learn some more about offset voltages and add this feature to ThrottleStop.

If you look in the log files that Quibbler2412 posted, the TDP Level lock bit is being set but it is being over ridden and ignored by the CPU.

 

As unclewebb says, I have locked the TDP level in all the tests I've done for you guys. But as you said, it is getting ignored completely. It's really awesome that you guys are willing to look into this though. If you need more testing done, please don't hesitate to ask.

 

Sorry Quibbler, just to be clear here you switch to tablet mode then run ThrottleStop to lock either level 0 or 2 then it reverts back to 1?

This is strange as it should only be changeable when locked after a reset, i.e. reboot or maybe going into sleep mode.

 

No, I lock it, and it stays locked (at least according to ThrottleStop) until I power down the computer. Going into or out of tablet mode changes nothing. However, the lock does not seem to have any effect on the actual TDP level. As unclewebb said (if I read that correctly), the lock bit apparently is not set, even if I lock the TDP level in ThrottleStop.

 

I discovered ThrottleStop yesterday and I like it very much. Unfortunately it shows that my PC does not enter the package C3, C6 oder C7 state. But the Core C States look fine. I'm using Windows 8.1 Pro x64.



My first question is what is the difference between the package C state and the CPU C State?

I activated all energy saving options on my mainboard (ASUS H87-Pro) following this guide: Maximus VI Power Saving & Power Tuning Guide | Republic of Gamers
Does somebody has an idea why my package C state is messing up?

 

On an Intel CPU, you can have the individual cores entering the various C States and when all of the cores are in the same or deeper C State, then the entire CPU package can enter that C State to save some more power. For example, if 3 of your cores were in C3 and one core was in C6 then the CPU package could enter C3.

What Windows power profile are you using? I like the Windows High Performance profile because it doesn't interfere with what the CPU wants to do. If your C States are working, the Balanced profile doesn't seem to save you anything when it comes to measured power consumption. The CPU multiplier becomes unimportant when a core is already in a deep C State.

Do you have the Package C States enabled in the bios? There are a couple of other settings within the CPU that can be used to discourage the CPU package from entering the deeper sleep states. Perhaps someday I will add a few more items to the ThrottleStop C State window so that users can fine tune their C States if they need to.

When overclocking my 3570K to 4500 MHz, I found that the package C States seemed to cause random BSOD issues. There is not a huge difference in power consumption whether the package C States are on or off so I would leave them off, especially on a desktop system. Some motherboard vendors will ignore your bios requests and adjust your CPU to what they think you really want. Asus might have left them off deliberately in the bios for this reason or maybe this is a Win 8.1 feature that needs some fine tuning. Not sure.

Edit: One other thing. The power consumption number reported by ThrottleStop and all software generated power consumption data is not 100% accurate. Intel included this information in their CPU but this data is intended to be used to control the Intel Turbo Boost feature. It is an approximation to actual power consumption so it is probably less than 100% accurate for C State testing. For proper testing you would need to measure the power going into the CPU socket.

Edit #2: You could try adjusting the C6 and C7 Latency in the bios from Short to Long. Maybe that will make a difference to CPU Package percentages.

 

Thanks for your fast answer.

Ok got it. But It's very strange that my CPU does can't even reach the C3 state in idle.

I'm currently using the balanced option.

Yes I enabled it. I did all options which are mentioned in the ASUS ROG guide.

Ok I try it.
EDIT: This option does not change anything. Also I installed Windows 7 for a short test. There is no difference.

 

jaegerschnitzel - The screen shot you posted shows your C0% at 17.1% which means your CPU is working on something. You might have something running on your system in the background, updates, etc. that is preventing your CPU from using the deeper package C States. Try getting your CPU to idle down to the low single digits in C0%. If you are still not seeing anything in the deeper package C States and you want me to investigate this further, run my MSR Finder tool and click on the Snapshot button to take a snapshot of your registers.

MSR Finder
MSRFinder

If downloading an unknown tool from a file sharing site like MediaFire scares you then you can also take a snapshot of the CPU registers using the CPU-Z program. Go to the About tab of that program, click on the Save Report (.TXT) button to create a file of your CPU's registers. I assume that it will cover the important registers that control this.

Copy and Paste the data to Pastebin.com - #1 paste tool since 2002! and then post a link here. My 3rd Gen Asus board has no problem using the C3 and C6 package C States. C7 core is not supported on the 3rd Gen desktop CPUs so there is nothing to report for that or C7 package. The mobile 3rd Gen can do C7 when you switch to battery power. I should be able to compare some registers to see if I can find out what is holding you back. Since this is happening in Windows 7 and 8.1, it is probably a motherboard issue. There are not a lot of freely available tools that report the C States so most users are in the dark about this stuff. Problems are rarely fixed by manufacturers when there are not enough complaints.



I find the C State data very useful when trying to track down a poorly written driver or some other crap running on one's system. The data is far more accurate than what the Task Manager provides users.

 

Here's a spin of from early Haswell testing, not by me, I just provided the SW. C7 needs ACPI support on DT IVB to work. No real difference in core C6 and C7 except C7 required as a prerequisite for package C7. C7 pkg should flush L3 cache so probably quite a bit of latency with that one. Not seen it used though, even on Haswell.

 

Thank you very much for your time and help. I'll provide you with the requested information later when I'm at home.

I know that the timing of my screenshot was bad ;-) But the problem is the same on full idle with max. 1% C0.
My working PC enters the C6 packet state even if I have over 20% C0 load, so it's definitely a problem with my computer at home.

EDIT: Your requested text files.
http://www55.zippyshare.com/v/48436157/file.html

 

Hi Unclewebb, first post so go easy on me.

I have just come across this thread and downloaded Throttlestop.

I am having problems with crackiling and poping during audio playback and recording via expresscard adaptor on a laptop. I have run a latency checker and disabled a few confilcting devices which has made the problem more bareable but still not right. The latency checker suggests it could be the cpu throttling causing the crackling and poping in the music which sounds reasonable however my BIOS will not let me stop the throtling down. I have also switched to Performance power mode and adjusted the minimum and maximum processor states to 100%.

My laptop is a HP Probook 4530's, with an Intel Core i5 2430m processor. According to CPU-Z and it is running anywhere between 8 x 100 and 30T x 100... I guess that 24 x 100 would be the max, non turbo value. I opened Throttlestop which suggests im running at anywhere between 13 x 100 and 18 x 100, if i click more data it suggests figures anywhere between 10 x 100 and 22 x 100.

What I would like to try and do is prevent the cpu from throttling down to keep it at a constant 2400 (i guess 10 x 24) to see if that indeed gives me the results im looking for audio wise. I guess I can always have another profile set up that reverts it to the original settings. On the performance setting, I tried setting the "set multiplier" to "24" and ticking the box but it doesnt seem to have any effect? Well when i say no effect in CPU-Z it seems to read a maximum of 24 x 100 now but still drops to 8 x 100.

Is there something I am missing?

Thank you very much for taking the time to read my query and help me! It is greatly appreciated. If you require any screenshots or logs I will happily upload them!

Regards, Matt

 

Dufus - My desktop motherboard does not let you enable C7 but I thought it was probably still hiding in the 3rd Gen desktop CPUs. Maybe you can find a trick to enable it.

jaegerschnitzel - Thanks for the data. I will have a closer look at your numbers to see what I can figure out but it might be a while before I add any new features to ThrottleStop to control this.

Welcome to NBR Matt.

All laptops are shipped with the low power C States enabled which significantly reduces power consumption and heat when lightly loaded. The C States are the main cause of the multiplier randomly hunting up and down in ThrottleStop as the CPU rapidly bounces into and out of various C States.

If you think this might be the main cause of your crackling problems then you can turn them all off if you like. If you are using Windows 7 or Windows Vista, you could try adding a new power profile to Windows that disables the C States as well as a few other things that might be causing problems. Here's some excellent background info including a download link.

Studio One - Optimizing Windows Vista/7 Power Settings for Studio One : PreSonus Audio Electronics

You will need to have C1E, C3 and C6 disabled so you can run your Core i5-2430M at a fixed frequency of 2400 MHz regardless of load. The two drawbacks are extra heat and disabling C3 and C6 also disables the Intel Turbo Boost feature.

To disable C States and to lock the CPU at a fixed frequency, I prefer to go into the registry and adjust one value.

Code:

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Intelppm

At the Windows search box type in regedit, navigate to the above registry key and change the Start value from 3 to 4. After you do that, exit regedit and reboot for the change to take effect. The ThrottleStop C States window will confirm if you have successfully disabled your C States. Adjust your CPU to a fixed frequency and see if that helps.

What latency checking program are you using and what did it show? Sadly, many laptops have been shipped with poorly written drivers which can cause horrible latency and sound crackling issues. The above fix might not be enough if your laptop falls into that category.

Here's some info from the DPC Latency Checker program that I used.

http://i.imgur.com/mpKv3Bf.png

With all of the C States enabled and with the multiplier randomly hunting up and down, I am not seeing any significant latency issues. Modern Intel CPUs can transition very rapidly out of the low power C States so they don't seem as evil as some forums suggest they are.

Here's an interesting test I did which compares ThrottleStop to CPU-Z and Intel XTU.

Event Viewer vs. Intel's ETU for reporting thermal throttling...

No need to run CPU-Z. You can trust what ThrottleStop is telling you.

 

It requires an ACPI table change but personally I would not worry about it as core 7 is the same as core 6. Never seen C7 package state used on both IVB and Haswell. Maybe it is only used with U/Y processors, idk.

 

Hi. I apologise for the dumb question (it may seem so to those more adequate in this area). I've read the initial post and the few first pages, but I still don't understand it completely.

Would it be possible to increase TDP on a ULV such as i5-4200U or i7-4500U? Currently they have a TDP value of 15W, would it be possible to increase it for a performance boost? And if it's possible, would it be worth the hassle, assuming the laptop can handle the increased heat properly?

 

Unclewebb,

Thank you soo much for taking the time to post this! It has been a huge help already!

I am using DPC latency checker which reads well in the green with absolutely no spikes at all like your example, and latency monitor which spikes every now and again quite high. Latency monitor suggested the CPU throttling but i think the program is a bit hit and mis because sometimes i run it without changing any settings it gives varying results, e.g. it will remain fully in the green, then try it again and there are spikes...

I tryed the method listed on the presonus website to add another power option but that did not work.

I then tryed your method of changing the registry value from 3 to 4 and then rebooting. I then fired up Throttle stop and on profile 1. The "set multiplier" box is ticked and set to 24. The other boxes that are ticked are "EIST" and "BD PROCHOT". It instantly stabilised the CPU at 24 x 100 which is exactly what I wanted!

This setting seems to enable my audio to work better with less pops and crackles which is quite a result... The best bit is I have been able to re-enable all the devices that I originally had to dissable and the audio still works as it did with them disabled where before it wouldnt work, so I would say that the throttling was partly causing my problems. I say partly because its still not 100% fixed my issues, (the device is an RME multiface, RME are known in music for outstanding drivers, stability and quality) although i now seem to be able to run at 64 samples buffer size (which is pretty swift) others with the same RME device are running at 32 samples... I understand im on a laptop and that does not compare to a high powered desktop but what i was getting before using throttlestop was crackling and poping at 128 samples and even some at 256 samples...

So what i then decided to do was creat a 2nd profile and set the multiplier to 12 x 100 for when im not working with audio or doing anything processor intensive. My temperatures measured on the Throttlestop page run around 60 on the high power setting and 45 on the low power setting... My question to you is as follows, is it ok to leave my laptop running like this i.e. selecting high power mode (24 x 100) when I want to do my audio work and then selecting low power (12 x 100) when say im browsing the net or reading emails etc... Will having the processor set at 12 x 100 cause any issues?

Again I would like to thank you for your help and your work on Throttlestop, I understand Im not really using it to anything like its full capabilities but for what I was looking for your guidance and throttlestop are one of the major things that has helped!

I look forward to your reply. Regards, Matt

 

Here another screenshot in idle and with a BIOS update. Looks a bit better, at least my CPU enters now the package C3 state. But the CPU stays only a few seconds in that state and then switches back to package C2 state.
Hope you can help me with analysing my data.

 

alfa_155 - Great to hear that you made some progress. You can run your processor however you want. Both 45C and 60C are well within the Intel specifications so there is no need to worry. Many desktop users turn off all of the C States just like you have done. Running your CPU fast or slow is also OK. Intel designs their CPUs to run safely at either speed for a long, long time so no worries. When you have the C States disabled with the registry mod, does the ThrottleStop C1E option do anything? I can't remember testing that. If it works, you can use that to reduce power consumption when needed. You can also set up a couple of keyboard shortcuts using ThrottleStop so you can switch profiles on the fly without having to open up the ThrottleStop user interface. Just keep it minimized in the system tray.

jaegerschnitzel - I am not surprised that a bios update changed this. I plan to have a close look at this when I have the time.

 

Will support ThrottleStop processors Intel Bay Trail?

My z3740 includes throttling too fast just 58С

 

Trying to figure out my processor problem, pretty sure it is thermal throttling. I would really appreciate any help.
Laptop: 14' dell inspiron 7000 i7 4500u

At start up the CPU is running fine, turbo boost is active, depending on load the frequency fluctuates up and down, above and below the 1.8ghz normal speed. After the temp hits a certain point and I hear the fans speed up for a min it drops to .77ghz and stays there. No mater what kind of load you try it just stays at .77ghz. I have not seen temps above about 60 besides once when i ran prime95 right after startup before the drop in speed. I saw temps hit in the 80's within a minute then computer shut down.

I have downloaded throttleStop and the Cmod and chip readings are both at 100%. Its running at 8 x 100 mhz.

If i uncheck BD prochot then everything starts working as it does before it gets stuck at .77

Can anyone explain to me what might be going on? Why does it get stuck even after cooling down? is not being able to run prime95 for a few mins before overheating say something is wrong with the cooling? I had to have the screen replaced and the tech did not seem like he knew what he was doing so maybe he damaged the cooling?

Thanks

 

I doubt that your laptop cooling has been damaged. Pretty much every major laptop manufacturer has turned out a variety of models over the last few years with one throttling problem or another. After reading Randall Cotton's excellent paper about his throttling Dell Latitude E6500, I decided to try and help users out by writing ThrottleStop one afternoon. His story is a long read but it details just how crazy some of these throttling schemes are.

throttlegate

Four years later, ThrottleStop is still the only solution that can properly diagnose throttling as well as correct it.

Some of the more popular monitoring tools still do not properly detect throttling so the majority of users are in the dark about throttling. Here's an interesting comparison.

Event Viewer vs. Intel's ETU for reporting thermal throttling...

Most users assume that throttling is triggered by high temperatures but it is frequently triggered by power consumption. BD PROCHOT stands for bi-directional processor hot. It is basically a signal path to your processor so other things on your motherboard can trick your CPU into thinking that it is too hot. The CPU responds by dropping the CPU multiplier to the lowest possible value so your CPU is left running at a little under 800 MHz. A bad sensor in your power adapter can be what is triggering this but it could be anything. When buying a new laptop, you are forced to do some thorough testing to see what sort of throttling problems are present. The throttling problem that you have discovered might have been there since day.

If it was my laptop and turning off BD PROCHOT allowed the CPU to run at its Intel rated speed, I would turn BD PROCHOT off. Unfortunately I can not recommend that you should do this because I have no idea why this throttling was added to your laptop. If this was intended to protect an under designed component on your motherboard then using ThrottleStop to bypass this might cause permanent damage to your motherboard or power adapter.

As always, use ThrottleStop at your own risk. I am happy to hear that it helped you find this problem and has the ability to correct it.

sanat - Bay Trail CPUs are low power CPUs. Intel designs these CPUs to throttle so they can operate in a low power environment. Post a CPU-Z screen shot so I can have a look but there might not be much that ThrottleStop can do.

Edit: The Z3740 has a 2 Watt Scenario Design Power (SDP) and sells in bulk for $32 per CPU. With a CPU-Z screen shot I might be able to get ThrottleStop to work with these but the difference in performance if any will likely be next to nothing. It is what it is.

 

Another culprit probably is crap bios/ec OEM love to use for sake of battery life or w.e reason ( like forcing geniue ac adapter). Usually check the owner lounge and see if other have similar issue, proceed for repair/return if it is an odd issue.

 

unclewebb, thank you
High performance is not particularly interesting to me.
I want to not include processor throttling at 58C.
And in games, such as DotA 2 after 15-20 minutes did not begin to slow down.
Cruising on the temperature specified by the manufacturer to the 90C.

Here are the data to the CPU-z http://sanatar.ru/pics/T100.html

 

Sorry, i am new at this.
So the only way to lower my temp is to lower the set multiplier number?

 

ThrottleStop 6.10 beta 3

New Features:
- added reporting of Overclocking Bins for Sandy Bridge and newer Core i CPUs.
- fixed ThrottleStop opening up from the taskbar when already running.
- reduced the maximum multiplier to 63.
- blocked access to the Turbo Ratio Limit adjusters when they are locked by the bios.

This version is a time limited beta. Just a few minor changes at the moment.
There are still quite a few things on the things to do list before the next release.

 

You could also reduce the VID value and adjust the set multiplier, I do find a stable combination by trying different combinations.