Question Turbo Boost - - on or off ?

[karasahin]

Hello,

To this day, I've been mostly using my CPU with turbo boost disabled because of generating extra heat when it's on (it's a mobile cpu).

When turbo boost is disabled, 2.5 GHz and 0.958V is the maximum for both. However, when turbo boost is active, it can often reach up to 3.1 GHz and 1.076V maximum.

Is more voltage THE cause for more heat? If so, more voltage means more wear on the CPU, right? Have I been doing right to keep turbo boost off to achieve less heat and more longevity of the CPU?

Lastly, if the main goal is to find a sweet spot for a laptop to perform not-so-horribly and last as long as possible (I can't really afford a good one rn), should I keep turbo boost off (cleaning and repasting done already)? Keeping turbo boost off also helps with the internal fan working less.

Thanks.

 

[TerryLaze]

Is more voltage THE cause for more heat?

Volts amps and watts are a triangle of correlation (ohms law) , volts is just the one that you can reduce without reducing performance, it just outright crashes if it doesn't get enough volts in some app.
Watts are the main cause of high heat but setting the max watts is setting a cut off point for performance.
Every CPU has a rated max temperature and as long as you stay below that the life span will be the same, it will break from other things before breaking from the temps.

 

[karasahin]

Volts amps and watts are a triangle of correlation (ohms law) , volts is just the one that you can reduce without reducing performance, it just outright crashes if it doesn't get enough volts in some app.
Watts are the main cause of high heat but setting the max watts is setting a cut off point for performance.
Every CPU has a rated max temperature and as long as you stay below that the life span will be the same, it will break from other things before breaking from the temps.

Logically, more GHz means more watt consumption as well, right? So it's more or less the same thing. I'm unable to lower voltage due to the CPU being Ivy Bridge.

You mean the max temp before throttling? It is 105C but yeah, I don't believe just because it runs at 104C and not throttling doesn't mean there would be no difference than running at 55C instead in terms of longevity. It doesn't make sense at all.

"it will break from other things before breaking from the temps"
Could you explain this part more?

 

[TerryLaze]

Logically, more GHz means more watt consumption as well, right? So it's more or less the same thing. I'm unable to lower voltage due to the CPU being Ivy Bridge.

When doing the same work more Ghz will use more watts.
If all three are locked down (volts amps watts) then yes, reducing Ghz will also reduce all three.

You mean the max temp before throttling? It is 105C but yeah, I don't believe just because it runs at 104C and not throttling doesn't mean there would be no difference than running at 55C instead in terms of longevity. It doesn't make sense at all.

The safety temp of a CPU is a good 10 degrees above the throttle temp, if you ever only reach the throttle temp then you are well below the safety limits.

"it will break from other things before breaking from the temps"
Could you explain this part more?

Basically it means that it will not break, it might break because the pins get bent if you remove/reinstall it or due to other mechanical factors.

 

[Misgar]

To this day, I've been mostly using my CPU with turbo boost disabled because of generating extra heat when it's on (it's a mobile cpu).

That's a bit like driving everywhere at 30mph (50kph) in a high gear (low revs) to make the engine last longer. Sure it's better for the planet and your wallet, but deliberately restricting performance means computing tasks can take longer.

Turbo boosting only oocurs when needed. If you're writing a document in Word, turbo will not be utilized much. If you're playing an intensive game, turbo boost should provide a better experience.

Is more voltage THE cause for more heat? If so, more voltage means more wear on the CPU, right?

To achieve high clock speeds, the BIOS raises the voltage on the CPU in a number of discrete steps. This voltage increase is needed to maintain stability. The faster the CPU speed, the more volts you need.

As a result of increasing the voltage, the current drawn by the CPU increases too and consequently, so does the power dissipation. Volts x Amps = Watts.

In digital electronics, a large proportion of the total power consumed by an IC occurs when individual transistors are switching between 0 and 1 and back again. The higher the clock speed, the greater the number of 0/1 transitions per second and the higher the power consumed.

That's why faster CPU clock speeds generate more heat. It's simple physics.

Lastly, if the main goal is to find a sweet spot for a laptop to perform not-so-horribly and last as long as possible

I have a fully working laptop that's 10 years old. A friend has a laptop that's close to 15 years old. I have even older laptops that might still work after more than 20 years (apart from their batteries).

How long do you want to keep your laptop? 10 years, 20 years or 30 years?

Limiting turbo boost will probably have no measurable effect on laptop longevity, unless you run Handbrake encodes 24/7/365.

 

[SkyNetRising]

Is more voltage THE cause for more heat?

Yes.

If so, more voltage means more wear on the CPU, right?

No. Unless you're doing heavy overclocking.

Have I been doing right to keep turbo boost off to achieve less heat and more longevity of the CPU?

No. Doesn't impact longevity.
Even with turbo boost enabled you're still running within spec.

should I keep turbo boost off?
Keeping turbo boost off also helps with the internal fan working less.

Probably not.
Unless your laptop battery has worn out and you have to conserve power at all cost.

 

[karasahin]

When doing the same work more Ghz will use more watts.
If all three are locked down (volts amps watts) then yes, reducing Ghz will also reduce all three.

The safety temp of a CPU is a good 10 degrees above the throttle temp, if you ever only reach the throttle temp then you are well below the safety limits.

Basically it means that it will not break, it might break because the pins get bent if you remove/reinstall it or due to other mechanical factors.

Thanks a lot.

So according to the pic above, the safety temp of my CPU is about 110 C? I try to never let it to reach above 100 C, but 98 C was reached during a stress test and it didn't throttle according to HWInfo.

Also, doesn't more heat due to more GHz cause the internal fan to work more? So, in this case by turning turbo boost off helps the fan to last longer no?

 

[karasahin]

That's a bit like driving everywhere at 30mph (50kph) in a high gear (low revs) to make the engine last longer. Sure it's better for the planet and your wallet, but deliberately restricting performance means computing tasks can take longer.

Turbo boosting only oocurs when needed. If you're writing a document in Word, turbo will not be utilized much. If you're playing an intensive game, turbo boost should provide a better experience.


To achieve high clock speeds, the BIOS raises the voltage on the CPU in a number of discrete steps. This voltage increase is needed to maintain stability. The faster the CPU speed, the more volts you need.

As a result of increasing the voltage, the current drawn by the CPU increases too and consequently, so does the power dissipation. Volts x Amps = Watts.

In digital electronics, a large proportion of the total power consumed by an IC occurs when individual transistors are switching between 0 and 1 and back again. The higher the clock speed, the greater the number of 0/1 transitions per second and the higher the power consumed.

That's why faster CPU clock speeds generate more heat. It's simple physics.


I have a fully working laptop that's 10 years old. A friend has a laptop that's close to 15 years old. I have even older laptops that might still work after more than 20 years (apart from their batteries).

How long do you want to keep your laptop? 10 years, 20 years or 30 years?

Limiting turbo boost will probably have no measurable effect on laptop longevity, unless you run Handbrake encodes 24/7/365.

Good points there, indeed by doing so I'm not fully utilizing the performance. But it never feels like the CPU being bottleneck with turbo boost off. In my experience it was the HDD and, now with the SSD, a few GHz increase doesn't seem to be noticeable except for maybe few FPS in games and benchmark tests.

My laptop is also about 10 years old. Obviously it wasn't working 7/24 for past 10 years but I still want it to work for about next few years.

 

[karasahin]

Yes.

No. Unless you're doing heavy overclocking.

No. Doesn't impact longevity.
Even with turbo boost enabled you're still running within spec.

Probably not.
Unless your laptop battery has worn out and you have to conserve power at all cost.

I see. Thanks for your reply.

 

[rgd1101]

I have a dell precision 5550 for work. turbo is off because the fan were constantly on. didn't really see any performance different

 

[karasahin]

I have a dell precision 5550 for work. turbo is off because the fan were constantly on. didn't really see any performance different

Does it have the i7-10850H? It is indeed a powerful CPU but there is about 2.4 GHz clock difference between the base and turbo clock, it seems a lot to me. To find a sweet spot, I believe it is possible to underclock turbo boost as well via Throttlestop. I forgot how exactly, but it was possible.

And yes, me neither and the fan working less and quieter is a good thing too.

 

[rgd1101]

i7-10750h 6 cores/12 thread.
turbo only last less than 30seconds.
unless you running old software that are running on 1 thread. don't really see big different

 

[karasahin]

i7-10750h 6 cores/12 thread.
turbo only last less than 30seconds.
unless you running old software that are running on 1 thread. don't really see big different

Possibly turbo duration is limited due to watt settings. I haven't noticed that in my case but then again the base and turbo clock difference for me is just 0.6 GHz for one core and 0.4 GHz for both cores.

Despite the performance loss (mostly in games and benchmarks), a laptop running quiet and cooler with turbo boost disabled is a luxury that you don't want to give up after getting used to 😆

 

[Misgar]

Despite the performance loss (mostly in games and benchmarks), a laptop running quiet and cooler with turbo boost disabled is a luxury that you don't want to give up after getting used to

When all said an done, it's your laptop so you can run "cool and quiet" for a peaceful experience. You'll miss out on snappier performance, but so what? It doesn't mattera jot.

I'm aware of fan noise on hard working laptops and desktops, but after working in server halls full of screaming high pitched Delta fans, I no longer worry about home computer fan noise. Must be going deaf.

Possibly turbo duration is limited due to watt settings.

Laptop manufacturers have the option to push some mobile CPUs harder for extre computing power on performance laptops where battery life is not important. Other manufacturers "tune" their BIOS to give slower CPU speeds, to extend battry life significantly. I'm not sure to what extent this is true for 10 year old CPUs with a more basic architecture.
https://www.extremetech.com/computi...bile-cpus-but-power-limits-may-throttle-chips

Also, doesn't more heat due to more GHz cause the internal fan to work more? So, in this case by turning turbo boost off helps the fan to last longer no?

You're probably extending the life of the fan by keeping the revs down. After 10 years, even the best fan bearings may start to wear. Replacing old laptop fans can be tricky, especially if you can't find an exact replacement.

Keep up the good work and you might get another10 years out of your laptop. Take care with older versions of Windows if you go on line, due to lack of security updates.

If you haven't replaced the BIOS battery, it might be getting weak. Brand new I'd expect 3.25V. Below 2.0V I'd replace it soon. If you have to enter the time and date at each power up, the CMOS battery has probably died.

The other "consumable" is the main laptop battery. Even the best OEM (Original Equipment Manufacturer) battery may be near the end of its useful life after 10 years.

I use a small app called BatMon to monitor battery condition on old laptops, especialy if I cannot get the manufacturer's battery app to work.
https://www.passmark.com/products/batmon/

If you click on the Info tab in BatMon you can check the original (claimed) capacity, together with its current fully charged capacity. You might find the original capacity was 45,000mAh, but over the years, it's dropped to 33,000mAh.

As a rough rule of thumb, I replace old batteries when they no longer keep the laptop running for at least 45 minutes of general use. A lot depends on how often the laptop is used away from a working mains supply for the charger. If the laptop is always always plugged in, I'm less concerned about nearly exhausted batteries, apart from their tendency to swell.

Laptop designs have improved considerably over the last 20 years. I used to be happy with 2 to 3 hours on battery. Now I'd expect 6 to 8 hours, but the best laptops can achieve 15 to 20 hours in run tests.

 

[karasahin]

When all said an done, it's your laptop so you can run "cool and quiet" for a peaceful experience. You'll miss out on snappier performance, but so what? It doesn't mattera jot.

I'm aware of fan noise on hard working laptops and desktops, but after working in server halls full of screaming high pitched Delta fans, I no longer worry about home computer fan noise. Must be going deaf.


Laptop manufacturers have the option to push some mobile CPUs harder for extre computing power on performance laptops where battery life is not important. Other manufacturers "tune" their BIOS to give slower CPU speeds, to extend battry life significantly. I'm not sure to what extent this is true for 10 year old CPUs with a more basic architecture.
https://www.extremetech.com/computi...bile-cpus-but-power-limits-may-throttle-chips



You're probably extending the life of the fan by keeping the revs down. After 10 years, even the best fan bearings may start to wear. Replacing old laptop fans can be tricky, especially if you can't find an exact replacement.

Keep up the good work and you might get another10 years out of your laptop. Take care with older versions of Windows if you go on line, due to lack of security updates.

If you haven't replaced the BIOS battery, it might be getting weak. Brand new I'd expect 3.25V. Below 2.0V I'd replace it soon. If you have to enter the time and date at each power up, the CMOS battery has probably died.

The other "consumable" is the main laptop battery. Even the best OEM (Original Equipment Manufacturer) battery may be near the end of its useful life after 10 years.

I use a small app called BatMon to monitor battery condition on old laptops, especialy if I cannot get the manufacturer's battery app to work.
https://www.passmark.com/products/batmon/

If you click on the Info tab in BatMon you can check the original (claimed) capacity, together with its current fully charged capacity. You might find the original capacity was 45,000mAh, but over the years, it's dropped to 33,000mAh.

As a rough rule of thumb, I replace old batteries when they no longer keep the laptop running for at least 45 minutes of general use. A lot depends on how often the laptop is used away from a working mains supply for the charger. If the laptop is always always plugged in, I'm less concerned about nearly exhausted batteries, apart from their tendency to swell.

Laptop designs have improved considerably over the last 20 years. I used to be happy with 2 to 3 hours on battery. Now I'd expect 6 to 8 hours, but the best laptops can achieve 15 to 20 hours in run tests.

Thanks for your detailed reply!

The thing is that this turbo boost feels like an extra/premium feature. Maybe it's because of its wording. It's cool to have it, if your laptop can handle it as same as desktop ones.

I can limit/underclock turbo boost clocks as well, so that I can find an excellent sweet spot for performance/heat/noise but I haven't tried it, since it doesn't really feel like performance drop is noticeable for simple Chrome usage, multimedia stuff etc. But I'll keep that in mind.

Haha I guess it's all about getting use to. I probably wouldn't mind my fan's noise that it's making when turbo is on if I were in your situation as well.

It's a good thing that you mentioned about security updates. Since Windows 10 paid extended security updates will last only for few years, I will stop using my laptop after that.

The BIOS battery is probably fine but I haven't had it checked. Hopefully it will last for more few years. And I replaced the main laptop battery relatively recent so it has still about 88% of its original capacity.

 

[Misgar]

When turbo boost is disabled, 2.5 GHz and 0.958V is the maximum for both. However, when turbo boost is active, it can often reach up to 3.1 GHz and 1.076V maximum.

I've just run a quick test on a Dell Latitude 5480 with an i5-7200U processor.

Aida64 stress test.

CPU clock rate 3094MHz
CPU temperature varies between 72 and 77C
Vcore 1.012V

My CPU didn't hit 80C. If your CPU (I'm not sure what you've got) is reaching 100C, it sounds a bit high. Perhaps not if it's a top of the range i7?

My 7950X desktop CPU will reach 95C under heavy load, but perhaps your laptop's heatsink could benefit from re-pasting with new thermal compound? Modern Intel desktop CPUs are designed to throttle at 100C, but AMD tend to be 95 or 90(ish).

Now that it's cooled down after the stress test:

i5-7200U idle
CPU clock rate 898.2MHz
CPU temperature varies between 39 and 40C
Vcore 0.638V

It's a good thing that you mentioned about security updates. Since Windows 10 paid extended security updates will last only for few years, I will stop using my laptop after that.

You could install Linux. I've tried a few distros but I'm used to Adobe software, so I'm sticking with Windows.

 

[karasahin]

I've just run a quick test on a Dell Latitude 5480 with an i5-7200U processor.

Aida64 stress test.

CPU clock rate 3094MHz
CPU temperature varies between 72 and 77C
Vcore 1.012V

My CPU didn't hit 80C. If your CPU (I'm not sure what you've got) is reaching 100C, it sounds a bit high. Perhaps not if it's a top of the range i7?

My 7950X desktop CPU will reach 95C under heavy load, but perhaps your laptop's heatsink could benefit from re-pasting with new thermal compound? Modern Intel desktop CPUs are designed to throttle at 100C, but AMD tend to be 95 or 90(ish).

Now that it's cooled down after the stress test:

i5-7200U idle
CPU clock rate 898.2MHz
CPU temperature varies between 39 and 40C
Vcore 0.638V


You could install Linux. I've tried a few distros but I'm used to Adobe software, so I'm sticking with Windows.

Very good temps. I think newer CPU, newer laptop, better cooling design, ambient temp etc. are all related for your lower CPU temps.

I have the i5 3210M which is according to PassMark results, a bit less powerful than the i5 7200U and has higher power consumption and older tech.

It reaches 100C under combined stress test of Prime95 and Furmark both without additional cooling solutions and without throttling. It is indeed high and I don't know why it's not throttling but at least my cooling pad helps about 10C drop in temps.

I don't trust myself to disassemble the laptop to reapply thermal paste. It would likely break. Maybe I buy a high quality thermal paste and give it to a trusted tech service for them to do it but not sure. Had that done already last year in summer without giving them a thermal paste.

Yeah I'm not sure about Linux either. My custom modified GPU driver barely works stable on W10. I don't want a new adventure.