Constant "High" CPU core MHz reading - Important?

[ddDerelict]

I just did a fresh OS install which included also re-installing motherboard drivers and optional apps. Running HWinfo after driver install but before apps install I get:


After installing Gigabyte z270 HD3 apps:


The yellow readings remain roughly at that level (they fluctuate mildly (~200 MHz) but stay in the yellows range). Temperature status (generally 7-10c above ambient) does not appear to be negatively impacted. Uninstalling individual apps does not change the status. Uninstalling the full suite returns the status to normal. While I've found the majority of the apps to be useless I do find the BIOS & firmware update capability useful.

Are there negative ramifications -- like shortened expected lifespan -- to having the CPU cores operating constantly at those levels? (If it's relevant the CPU is i7 6700.)

 

[Martin]

This might be indeed caused by some of those apps. Check CPU usage in sensors and uninstall them if you don't need.

 

[Dalai]

Does that software change the (active) power plan to High Performance?

Regards
Dalai

 

[ddDerelict]

@Martin - that's what I suspected. Thanx.

@Dalai - hadn't even considered that. Does a High Performance power plan typically increase CPU usage to the degree observed in my screenshots?

 

[Zach]

Does that software change the (active) power plan to High Performance?

Regards
Dalai

I don’t think it can.

What a software can do is to occupy more CPU usage in a way that won’t let “him” drop speed/voltage. Thus less entering sleep mode. Meaning that the effective clock is high, even though overall load is low.

Usually board (crap)software do such things. Especially RGB control software. I don’t like them at all...!!! Every time I had install such, I’ve bitter regret it and was from different vendors. (MSI, ASUS, Gigabyte)

What short of apps have you installed from GB?

 

[Dalai]

My assumption was this: The High Performance power plan tells the CPU to not reduce clock speed (at all) - which can be seen in the second screenshot. Now I know that higher CPU frequency can also be caused by high(er) CPU usage. OTOH, ddDerelict said that the temperatures are not much negatively impacted which suggests low CPU usage.

But it could be both, a changed power plan, or high CPU usage, or at least a high polling intervall from the software which keeps the CPU from entering low power states and dropping clocks.

Does a High Performance power plan typically increase CPU usage to the degree observed in my screenshots?

There's only a high clock speed in your screenshots, nothing about CPU usage. Check Windows Task Manager or the appropriate sensors in HWiNFO to see the CPU usage.

Regards
Dalai

 

[Zach]

But it could be both, a changed power plan, or high CPU usage, or at least a high polling intervall from the software which keeps the CPU from entering low power states and dropping clocks.

Something like that yes.
Some software all that have to do is to add a few or a couple of % to usage to prevent CPU entering low power stages. I’ve seen it with RGB software. This can’t raise heat (temperature) by a significant amount but does the “damage” to low power stages. You will see the difference mostly on the average clock and for sure on the average effective clock because the latter contains all power stages (high and low/very low).

 

[VictorVG]

Remove the math from the motherboard kit, leave only the drivers. Vendor math is good only when the problem cannot be solved by other means, and in other cases it is not worth setting: less garbage in memory means more reliable OS operation.

As for the BIOS flashing - have you seen a "brick" for a long time? This is what I mean, flashing the BIOS chip from under an active multitasking OS is fraught with a "brick" and feverish searches for a programmer. And this is at best, if you manage to connect it to the microcircuit through a clothespin or the ISP connector.

Otherwise, it will mean looking for a soldering station with a hair dryer, blowing off the chip, going to the programmer and sticking it back in. Otherwise, the board will not start. I have seen hundreds of such "bricks" and the reason in 98% of cases was the same - a write error in the microcircuit due to the fault of a multitasking OS.

Therefore, if you want to flash the BIOS, then do it by booting the machine from a flash drive under DOS and not with programs like amiflash / awardflash, but Intel fpt from its SDK. fpt will also overwrite the Boot Block BIOS area and amiflash / awardflash or the firmware utility built into the BIOS (usually the same amiflash / awardflash, only hidden in the Boot Block BIOS area) this area does not touch, which is fraught with errors in the OS. For example, some critical technologies may not work.

For example, for example, they dragged a server under a guarantee on which, due to a firmware error in this area, the OS refused to start, considering that this machine was not certified to work as a banking system server. I had to solder the microcircuit, flash it on the programmer and solder it into place. And the guys were worried about this situation - the bank could not work normally for several days. So before you sew something, you need to calculate each of your actions a hundred times.

True, I was lucky with the board - the Chinese just sent a box with the inscription "X79 P3" in which the board itself, an adapter ring for a cooler, a SATA cable and a tin for the case lay. Everything, even a piece of paper, albeit with a minimal description was not. But everything is written on the board itself - each jumper is clearly signed, but the front panel connectors - the audio is simply written FAUDIO, and the front connector is color-coded. And another one that looks like an ISP with a DIAG signature. All the rest "hand over it yourself". True, it was great luck that the chipset was not the desktop B61 / H61 / Q67 / B71 / Q77 or ideally X79, but the server C602J chip PCH Intel BD82C602J - apparently the factory had a lot of them in the warehouse and the boards were soldered on them, and therefore, if you do not count forever working smartphone because CPU temperature sensor measures the price of seeds on the market on a market day - cheaper strapping of the multi I/O chip, then the rest is lucky - everything works without errors, and the machine is equal in performance to DELL Precission T5600, only it weighs not 14 kg, but about 12 kg - Zalman cases have thinner iron. And in terms of performance, it even surpasses it. Enough for CAD.

In to HWiNFO64 I see:

 

[ddDerelict]

What short of apps have you installed from GB?

BIOS & Firmware Updaters, System Fan Controls, Fast Boot (basically restarts and directly enters UEFI/BIOS), Game Boost BUT when looking at installed programs a LOT more than that actually got installed, the majority of which I could not directly connect with what I selected for install.

There's only a high clock speed in your screenshots, nothing about CPU usage. Check Windows Task Manager or the appropriate sensors in HWiNFO to see the CPU usage.
Regards
Dalai

OK I'm going to reveal my ignorance here but I had understood that when the system increases the clock speed that is (directly) because the CPU is being used more and the higher clock is required to maintain a consistent stable performance level ... does that mean clock speeds and CPU usage are unrelated?

Otherwise, it will mean looking for a soldering station with a hair dryer, blowing off the chip, going to the programmer and sticking it back in. Otherwise, the board will not start. I have seen hundreds of such "bricks" and the reason in 98% of cases was the same - a write error in the microcircuit due to the fault of a multitasking OS.

Therefore, if you want to flash the BIOS, then do it by booting the machine from a flash drive under DOS and not with programs like amiflash / awardflash, but Intel fpt from its SDK. fpt will also overwrite the Boot Block BIOS area and amiflash / awardflash or the firmware utility built into the BIOS (usually the same amiflash / awardflash, only hidden in the Boot Block BIOS area) this area does not touch, which is fraught with errors in the OS. For example, some critical technologies may not work.

Now that is very interesting. This is my first custom build (first PC tho was in 1990) so my learning curve is still vertical. I have, of course, read accounts of BIOS updating and "firmware flashing" going seriously awry. I recenty did my first firmware update to address the Intel Management Engine snafu (which incidentally didn't work & a ticket's been raised with Gigabyte tech support, who have not been impressive so far). I was of the belief that a process developed by the mobo manufacturer would be the safest way to do something as inherently risky as mucking about with machine code BIOS/firmware. I confess I was surprised to be performing that task from within Win7 (but presumed it was UEFI related) and can certainly understand the logic of your statement regarding "write error due to multitasking OS". Is there a tutorial you could point me to that can teach me the proper (maximize accuracy & minimize risk) way of doing a BIOS or firmware update?

 

[Zach]

BIOS & Firmware Updaters, System Fan Controls, Fast Boot (basically restarts and directly enters UEFI/BIOS), Game Boost BUT when looking at installed programs a LOT more than that actually got installed, the majority of which I could not directly connect with what I selected for install.

I would start by unistalling the Gameboost and all others that I dont really want/use. Any kind of software that claims to boost things as games or whatever else, usually keeping CPU on the edge and may not let it "cool off" (as sleep cores/low power stage). Those things offer nothing in my point of view. Its like using "High/Ultimate Performance" power plan for no reason.
And OS BIOS updater app is also unecessary though I dould this one is the cause of issue. I also have a Gigabyte board (X570 for Ryzen) and I would assume you have installed the "APP Center" that contains all other Gigabyte software. I only use SIV (System Info Viewer) for the fan control (SmartFan 5)

OK I'm going to reveal my ignorance here but I had understood that when the system increases the clock speed that is (directly) because the CPU is being used more and the higher clock is required to maintain a consistent stable performance level ... does that mean clock speeds and CPU usage are unrelated?

It is related but depends on what you mean as usage. As I said earlier a couple of more % could keep CPU from entering low power stage. But its not always the case. It depends on how a software uses the CPU.
+1% usage from a software is not equal to +1% usage from a different software.

Now that is very interesting. This is my first custom build (first PC tho was in 1990) so my learning curve is still vertical. I have, of course, read accounts of BIOS updating and "firmware flashing" going seriously awry. I recenty did my first firmware update to address the Intel Management Engine snafu (which incidentally didn't work & a ticket's been raised with Gigabyte tech support, who have not been impressive so far). I was of the belief that a process developed by the mobo manufacturer would be the safest way to do something as inherently risky as mucking about with machine code BIOS/firmware. I confess I was surprised to be performing that task from within Win7 (but presumed it was UEFI related) and can certainly understand the logic of your statement regarding "write error due to multitasking OS". Is there a tutorial you could point me to that can teach me the proper (maximize accuracy & minimize risk) way of doing a BIOS or firmware update?

I only update board BIOS from inside the BIOS (QFlash utility within BIOS) and have the update file in a flash drive. Once you're in the Qflash (hit F8 when you're inside BIOS) its pretty straight forward.

Alternative way is QFlash+ (plus) function that a board has if there is a white usb port on the back I/O plate and a QFlash button somewhere on the board. You have the file on a flash drive plugged on the white port and with the PC off but plugged on PSU (PSU ON) you hit the QFlash button and the board updates the BIOS by itself within 2 to 5 minutes. A flashing LED somewhere let you know if the procedure is started. When finished it reboots itself.

You can search for tutorials for both ways on YouTube.

 

[Dalai]

OK I'm going to reveal my ignorance here but I had understood that when the system increases the clock speed that is (directly) because the CPU is being used more and the higher clock is required to maintain a consistent stable performance level ... does that mean clock speeds and CPU usage are unrelated?

Higher CPU usage can directly cause higher clocks. In general, the more load processes put on the CPU the higher and more consistent the CPU frequency - up to a certain point of course. When there's less load on the CPU it will reduce the clocks. That's the (normal) behavior when using the Balanced power plan.

But if the power plan doesn't allow the CPU to increase clocks (e.g. when using Power Saver power plan), it won't increase clocks regardless of the load processes put on the CPU. And as I already said, there can be other reasons why a CPU won't reduce clocks, e.g. when the High Performance power plan is used.

In summary: There is a causality between CPU load and CPU frequency under certain conditions (i.e. Balanced power plan).

Regards
Dalai

 

[ddDerelict]

Responses greatly appreciated and usefully educational... chapeau!

 

[flyingdutchman]

Recently built a pc with ASUS Prime Z490-P mobo, 64Gb ram, Intel Core i7. All worked fine until last week when the cpu package temperature slowly climbs up to over 70 degrees and all the fans kick in. However, when I run HWiNFO, the temperature immediately begins to drop and settles at around 43 degree. Close the HWiFO program and the temperature starts to go up again.
This perhaps seems to indicate there is something on the mobo that needs adjusting?.
Any thoughts anyone?

 

[Zach]

Recently built a pc with ASUS Prime Z490-P mobo, 64Gb ram, Intel Core i7. All worked fine until last week when the cpu package temperature slowly climbs up to over 70 degrees and all the fans kick in. However, when I run HWiNFO, the temperature immediately begins to drop and settles at around 43 degree. Close the HWiFO program and the temperature starts to go up again.
This perhaps seems to indicate there is something on the mobo that needs adjusting?.
Any thoughts anyone?

You can open TaskManager on “Processes” tab and see which one loads the CPU when HWiNFO is closed.

 

[Martin]

This looks more like some issue with fan control or sensor when launching HWiNFO kicks the control in. Perhaps a BIOS update might help.

 

[flyingdutchman]

Thanks for the suggestions. As this is a recent build, all drivers, bios etc are up to date. Sounds like there might be some setting in the bios which is set incorrect..