CPU Clock with Turbo Boost

[aruff]

Hello,

I try to display the CPU Clock of a Processor with Turbo Boost enabled. (LGA 2011-v3)
If I display the CPU Clock within the TaskManager (Windows 10) it shows me every second a new value inert the Clock variant. It may differ extreme with two decimal values. E.G. 2.73 GHz- 3.45 GHz and so on. Of course, this is a calculated value.

If I show the same  CPU Clock value within Aida64 it will only show 2 Values with only a little differentiation. 1200.1 GHz what is indeed the lowest level and 4500.1 what is the highest value.
In some Cases it will show 4499.8 or 4499.9 instead of 4500.1.

If I look with an MSI Tool over the CPU real Frequency it will, of course, differ in many circumstances over all cores in a great variant. All cores runs with different clock scores.

So, what is the difference between the calculation of the real clock to the displayed value in Aida64?
​How did you calculate the value?
The calculation of the TaskManager seems to be smoother and smarter then the aida calculation.

PS: The same behaviour on the CPU Multiplier, this value shows also only two values, min and max. an never a value between.

[Fiery]

AIDA64 lets CPUs have a bit of "break" before it measures the CPU core clock. Hence when the CPU is idle, it should measure the lowest clock rate, the sleeping state. Other software may not let the CPU go to sleep before measuring the core clock, and so the CPU may be stuck at a higher clock rate due to previous tasks ran by the same software.

Regards,

Fiery

[aruff]

Thanks for the answer.

​So, this mean, I will never get another Value then min or max?

​Again, obviously the Win 10 taskmanager give me more Information then Aida, is this your intense?

Could you please integrate a Checkbox for enable/disable the "break" Feature?

​On the other side, why does Aida only shows the max value instead of different states between min/max?
​Is this also a Feature?

Again, please explain, how do you calculate the CPU clock? 
All of the other tools shows different values, some tools for each core (also taskmanager).

And Aida shows only min/max?

If so, this value is not “CPU Clock†it is then obvious only “CPU Sleep Mode yes/noâ€

 

There is no way to show the actual CPU Clock of any available CPU instead of the calculated CPU Sleep Value?

 

Or did I miss something?

[Fiery]

So, this mean, I will never get another Value then min or max?

You get other value than min or max when your CPU is currently operating at such a state. Like when it is under heavy load, it won't be able to sustain running at the maximum Turbo Boost clock, and will step back a few hundred MHz's.

 

Again, obviously the Win 10 taskmanager give me more Information then Aida, is this your intense?

No. Why would we want to copy a built-in Windows feature? It is already done, we don't intend to re-invent the wheel again.

 

Could you please integrate a Checkbox for enable/disable the "break" Feature?

I'm not sure what "break" feature you mean. Can you please tell me more about it?

 

On the other side, why does Aida only shows the max value instead of different states between min/max?

It will show the in-between states when the CPU is running at those. When the CPU is under partial load, it doesn't matter if the CPU clock measurement thread of AIDA64 provides the pause or not, the CPU won't be able to go idle and put its multiplier to the minimum state. That's when AIDA64 will measure an in-between clock.

 

Again, please explain, how do you calculate the CPU clock?

It depends on the CPU type. On modern Intel processors AIDA64 (and also CPU-Z, HWiNFO, SIV, and a lot of other monitoring software) measures BCLK (reference clock), detects the current CPU multiplier, and multiplies them to get the CPU core clock.

 

All of the other tools shows different values, some tools for each core (also taskmanager).

Task Manager may just show an average clock speed measured over a period of time. We don't know what exactly it does, since we are not the developers of that tool. We are not Microsoft

 

So, this value is not “CPU Clock” it is then obvious only “CPU Sleep Mode yes/no”

No. It is the current CPU core clock speed.

 

There is no way to show the actual CPU Clock of any available CPU instead of the calculated CPU Sleep Value?

AIDA64 does show the actual CPU clock.

 

Or did I miss something?

IMHO yes

[aruff]

Ah, you think my Questions are funny.
What is it? Is my English too bad?
Again, I did not get other values then min or max.
-You- wrote that Aida will have a “Break†before it measures the CPU Clock.
So the “Break†lets Aida measure the “Sleep Sate†without prevent Aida for itself that this point will never be reached. I understand.

But I will not measure that min CPU Clock to achieve the “correct†Sleep State instead of CPU Clock.
This may be interesting but not as CPU Clock. Because I know what the min is (min frequency), it may be interesting how often the Sleep State will be reached, but not what the frequency is in sleep State.

Nice if you add this as a new Measure point.
Measured Sleep State Frequency and how often this Point is/was reached.

So, if you “disable†the “break†before measuring I will get some other values then “Sleep Stateâ€, correct?
This is what I mean with “please integrate a Checkbox for enable/disable the "break" Featureâ€
Let the user decide if he wants to have the actual CPU Clock or the “Sleep Stateâ€.
And yes, I understand that such a measurement will prevent the CPU to come at any time to sleep state because the measurement itself needs a min. activity. 
 

It will show the in-between states when the CPU is running at those

 

This is my problem, it does not!
I never did see any other value then min or max, again 1200 or 4500 GHz, nothing between.
And this during other tools report me complete different values. E.G taskmanager or the MSI Command Center. Pic. attached.

So, obviously there are many different states between two measure time points.
All other tools show my different values only Aida will show only min/max.
Can you now see what I mean?
The Question is, can Aida show me different values instead of min/max or not?
And the explanation can’t be “we do like CPUzâ€, the question is how Aida can or shall do.

Possible it is the wrong way to calculate only the multiplier with reference clock?
Obvious other can do it better.

[Fiery]

If we disable the pause before measuring CPU clock, the CPU would be running in a high frequency state (either max or a few steps below max) simply because it switched up because of the stuff we do in AIDA64 before measuring the CPU clock. So it would never show the idle state, even when the CPU is actually at idle. Trust me, it would be a bad idea and would yield to misleading results.

 

The Question is, can Aida show me different values instead of min/max or not?

Yes, it can. Try to open the AIDA64 CPUID Panel (from AIDA64 / main menu / Tools), and grab the window, and vigorously move it around the screen. Shake the window if you will. You will see that as the CPU core clock is updated in the background, the clock will fluctuate between the min and max, showing multiple states in between them.

[aruff]

Thx again for you answer.

No, also with the CPUID Panel I can’t see any other then min or max.
Indeed, the last two digits, included the decimal will show me some little differences like I wrote in my first post.

And this is why I ask!

Aida shows my different values (also in the CPUID Panel)
1200
1200.1
4500
4500.2
4499.9

Many other tools show me complete different values at the same time.
Many other tools can show me the frequency of each Core and then they, maybe, calculate a complete other CPU Clock then Aida do.
Of course.

Is the Multiplier every time only in min/max state?
(Is it all the time only switching between sleep/full? Aida think so?)
 

Can Aida measure the multiplier of each core instead (of course in addition) of measuring only all cores?

Can Aida calculate other than the simple way (basefrequency X multiplier)?
(e. G. average between different clocks of each Core)

Would you integrate this?

[Fiery]

There's an experimental CPU Clock Monitor panel in the right-click debug menu of AIDA64. To start it, just right-click on the bottom status bar of AIDA64 main window --> CPU Clock Monitor. You may need to enable status bar in AIDA64 / main menu / View first.

Can Aida calculate other than the simple way (basefrequency X multiplier)?

(e. G. average between different clocks of each Core)

There's no averaging. The CPU core clock frequency is measured for the first logical processor (first core), or for the selected CPU if you use the dropdown list on the CPUID Panel.

[aruff]

Ok I see the experimental Clock Monitor.
Same behaviour inside this panel, only 2 Values, min or max.
Either 1200 or 4500, nothing more. And it will cause around 5% workload over all.
Anyhow, different for each core.
 

 

Ok, it is experimental but it seems to be an important Aida incapable issue.
​(PS: I never read before CORE 1 to 6, it is always CORE 0 to 5, isn't it?)
 

At the same time the impure MSI Software shows me exactly what going on in my CPU with 0.3% workload!

So, the answers to my questions seems to be very simple.

1. Aida cannot measure newly CPU exactly

2. If you measure, there will be a high workload to the CPU without increasing accuracy

3. It calculates nothing; it will always use the simplest formula (only one core is measured with baseclock X Multiplier)
​4. It uses a "break" to determine the exactly "Sleep State" moment without asking for
​5. Aida cannot report how often the sleep state will be occure

6. Aida cannot measure any modern CPU frequency between "Sleep State" and max. frequency 
 

Please, this is not what I expected from such a Software.
From my point of view, this function is essential for monitoring overclocking.
I see some posts inside your Forum asking for this feature.

Please, have a look to such tools like MSI Command Centre and try to find out what these developers do other then you. (with much more accuracy by using very low workload).
And then,  please implement the same measurment to Aida.
Until that, the overclock feature inside Aida seems to be complete useless.

And, again, rename CPU Clock in this context to “CPU Sleep Stateâ€. What seems to be the right name for.
In my environment the CPU Clock inside Aida shows me only how often the CPU went to Sleep State from maximum CPU Clock, nothing more.

[Fiery]

I'm sorry, but you're not right about most your points, and it's mostly because you don't understand how modern processors work, and how does CPU clock frequency measurement work.

 

It uses a "break" to determine the exactly "Sleep State" moment without asking for

If you don't let the CPU go to sleep, then you're not measuring the clock frequency of the CPU at any given time point, but the clock frequency that your software makes the CPU to switch to, right before you measure the frequency. By inserting that pause you eliminate the effects of the workload that the preceding parts of the monitoring software. Of course if the system is running not just the monitoring software, but some other task as well (like a partial workload caused by a background virus scanning process for example), then the pause will not cause the CPU to completely go to sleep -- since it still has other work to do.

If you want to do another experiment, then start a task that loads just one CPU core, like a WinRAR benchmark with the Multithreading checkbox cleared, and then check what the CPU Clock Monitor shows. After that, start a task that is not that constant about its CPU load impact, like start a background virus scanning or such.

---

Based on your posts, it seems to me that you simply don't like AIDA64. In which case you can just use any of the alternatives that came up here in this topic. Noone here would force you or persuade you to pick AIDA64 over the alternative monitoring solutions. We're still 100% sure that we measure CPU core clock frequency the right way. Not because it's our way, but because it provides the accurate results.

[aruff]

Not yet funny as you expected in the beginning, isn’t it?

I have paid for this and I want to use it.

For me a “Information Software†does not have any emotional attribute how “like or not likeâ€

But, to use it I have the need to fully understand what it is doing.

Ok, I have done your requested experiment and this shows, what wonder, the one Core presented me 4500 GHz during the Job, the others either 1200 or 4500 GHz.

The MSI Tool can display me some different states, between these 2 measure points during the same job.  So, the developer from this tool does not understand how modern processors work?

Let us go away from the “break†discussion. I understand that fully.

And if Aida wish to display the Sleep State correctly, be welcome.

 

But, why Aida can never display any value between 1200 or 4500 GHz? (in my environment)

Could you please explain this?
 

The Intel Extreme Tuning Utility write:

When the processor is operating below these limits and the user's workload demands additional performance, the processor frequency will dynamically increase in step sizes equal to the current reference frequency on short and regular intervals until the upper limit is met or the maximum possible upside for the number of active cores is reached.

 

But, I never can see any of these “step sizes†in Aida. Why not?

[Fiery]

I have no idea why you cannot see any middle steps between the min and max ratios. We've just checked: on our LGA2011-v3 system (Gigabyte X99-UD4 + Core i7-5820K) clock ratio fluctuates between 12x (min) and 36x (max) with several in-between steps, like 30x, 33x, 34x, etc. What kind of system do you use? What happens if you return to default settings, so no OC?

BTW, MSI Command Center, at least on the shot you've linked above, doesn't seem to be all that accurate. 100.0 x 40.0 = 4021.7 MHz? I don't think so... And when a software is obviously inaccurate, I wouldn't use it as a reference in a discussion.

[Fiery]

One feature that may influence the way the CPU switches between various power states is Turbo Boost Max 3.0:

http://www.anandtech.com/show/10337/the-intel-broadwell-e-review-core-i7-6950x-6900k-6850k-and-6800k-tested-up-to-10-cores/2

I'm not sure however what CPU you've got, and whether you have TBM 3.0 enabled or not. Also, when a CPU is overclocked, it may behave completely different to a similar CPU running at stock settings. So it would be very important to find out what CPU you've got, and also whether it works different without OC and special settings, so in its factory default state. It would also be interesting to see the list of PStates the BIOS Setup configures for your CPU.

Also, if you disable SpeedStep (EIST), the CPU may fall back to using C1E (Enhanced Halt State), to save power. C1E only alternates between the min and max ratio, no middle steps. Same thing could happen if you (or a software) alters the Windows CPU power management settings by forcing the CPU to work at the highest ratio all the same, effectively disable SpeedStep. In which case C1E may still work, so the min and max ratios may be used, just no middle steps.

BTW, you can check and modify the Windows CPU power management settings, enable/disable C1E, enable/disable Turbo Boost using the CPU Tweakings panel of AIDA64. Which you can launch from the same right-click context menu that you use for CPU Clock Monitor.

[aruff]

How funny.

I Have a look to a feature what Aida called „Overclock“ and you suspect that has something to do with overclocking.

However, interesting answer. I will have a look to.

 

If I disable C1E and Turbo Boost it changes between 1200 GHz and 3300 GHz without any steps between. I must investigate to TBM 3.0

 

To consider, the MSI Tool or TaskManager does not reflect in any way to this different states. They can see the different steps regardless my switches.

 

Withou Turbo Boost, C1E disabled___________________________

CPU-Eigenschaften:

      CPU Typ                                           HexaCore Intel Core i7-5820K

      CPU Bezeichnung                                   Haswell-E

      CPU stepping                                      R2

      Engineering Sample                                Nein

      CPUID CPU Name                                    Intel® Core i7-5820K CPU @ 3.30GHz

      CPUID Revision                                    000306F2h

      CPU VID                                           1.3019 V

    CPU Geschwindigkeit:

      CPU Takt                                          3300.0 MHz  (Original: 3300 MHz)

      CPU Multiplikator                                 33x

      CPU FSB                                           100.0 MHz  (Original: 100 MHz)

      North Bridge Takt                                 3000.0 MHz

      Speicherbus                                       1333.3 MHz

      DRAM:FSB Verhältnis                               40:3

    CPU Cache:

      L1 Code Cache                                     32 KB per core

      L1 Datencache                                     32 KB per core

      L2 Cache                                          256 KB per core  (On-Die, ECC, Full-Speed)

      L3 Cache                                          15 MB  (On-Die, ECC, Full-Speed)

    Motherboard Eigenschaften:

      Motherboard ID                                    <DMI>

      Motherboard Name                                  MSI X99A Gaming Pro Carbon (MS-7A20)  (2 PCI-E x1, 4 PCI-E x16, 1 M.2, 8 DDR4 DIMM, Audio, Gigabit LAN)

    Chipsatz Eigenschaften:

      Motherboard Chipsatz                              Intel Wellsburg X99, Intel Haswell-E

      Speicher Timings                                  14-14-14-36  (CL-RCD-RP-RAS)

      Command Rate (CR)                                 2T

      DIMM-B1: Kingston HyperX KHX3000                  4 GB DDR4-2133 DDR4 SDRAM  (16-15-15-36 @ 1067 MHz)  (15-15-15-36 @ 1067 MHz)  (14-14-14-35 @ 1037 MHz)  (13-13-13-32 @ 962 MHz)  (12-12-12-30 @ 888 MHz)  (11-11-11-27 @ 814 MHz)  (9-9-9-22 @ 666 MHz)

      DIMM-B2: Kingston HyperX KHX3000C15D4/4GX         4 GB DDR4-2133 DDR4 SDRAM  (16-15-15-36 @ 1067 MHz)  (15-15-15-36 @ 1067 MHz)  (14-14-14-35 @ 1037 MHz)  (13-13-13-32 @ 962 MHz)  (12-12-12-30 @ 888 MHz)  (11-11-11-27 @ 814 MHz)  (9-9-9-22 @ 666 MHz)

      DIMM-C1: Kingston HyperX KHX3000C15D4/4GX         4 GB DDR4-2133 DDR4 SDRAM  (16-15-15-36 @ 1067 MHz)  (15-15-15-36 @ 1067 MHz)  (14-14-14-35 @ 1037 MHz)  (13-13-13-32 @ 962 MHz)  (12-12-12-30 @ 888 MHz)  (11-11-11-27 @ 814 MHz)  (9-9-9-22 @ 666 MHz)

      DIMM-C2: Kingston HyperX KHX3000C15D4/4GX         4 GB DDR4-2133 DDR4 SDRAM  (16-15-15-36 @ 1067 MHz)  (15-15-15-36 @ 1067 MHz)  (14-14-14-35 @ 1037 MHz)  (13-13-13-32 @ 962 MHz)  (12-12-12-30 @ 888 MHz)  (11-11-11-27 @ 814 MHz)  (9-9-9-22 @ 666 MHz)

      DIMM-D1: Kingston HyperX KHX3000C15D4/4GX         4 GB DDR4-2133 DDR4 SDRAM  (16-15-15-36 @ 1067 MHz)  (15-15-15-36 @ 1067 MHz)  (14-14-14-35 @ 1037 MHz)  (13-13-13-32 @ 962 MHz)  (12-12-12-30 @ 888 MHz)  (11-11-11-27 @ 814 MHz)  (9-9-9-22 @ 666 MHz)

      DIMM-D2: Kingston HyperX KHX3000C15D4/4GX         4 GB DDR4-2133 DDR4 SDRAM  (16-15-15-36 @ 1067 MHz)  (15-15-15-36 @ 1067 MHz)  (14-14-14-35 @ 1037 MHz)  (13-13-13-32 @ 962 MHz)  (12-12-12-30 @ 888 MHz)  (11-11-11-27 @ 814 MHz)  (9-9-9-22 @ 666 MHz)

 

    BIOS Eigenschaften:

      Datum System BIOS                                 07/15/2016

      Datum Video BIOS                                  Unbekannt

      DMI BIOS Version                                  1.20

    Grafikprozessor Eigenschaften:

      Grafikkarte                                       Gigabyte GeForce GTX 1080

      GPU Codename                                      GP104-400  (PCI Express 3.0 x16 10DE / 1B80, Rev A1)

      GPU Takt                                          1328 MHz

      Speichertakt                                      2749 MHz

With Turbo Boost_________________________________

CPU-Eigenschaften:

      CPU Typ                                           HexaCore Intel Core i7-5820K

      CPU Bezeichnung                                   Haswell-E

      CPU stepping                                      R2

      Engineering Sample                                Nein

      CPUID CPU Name                                    Intel® Core i7-5820K CPU @ 3.30GHz

      CPUID Revision                                    000306F2h

      CPU VID                                           1.3019 V

 

    CPU Geschwindigkeit:

      CPU Takt                                          4499.8 MHz  (Original: 3300 MHz, Overclock: 36%)

      CPU Multiplikator                                 45x

      CPU FSB                                           100.0 MHz  (Original: 100 MHz)

      North Bridge Takt                                 2999.9 MHz

      Speicherbus                                       1333.3 MHz

      DRAM:FSB Verhältnis                               40:3

[Fiery]

Thank you. TBM 3.0 is only supported by Core i7-6800 and 6900 Series, so we can skip that. On your particular setup the best explanation may be that EIST is disabled for some reason, and what you can see the CPU alternating between the min and max multiplier is due to C1E activity. Another explanation might be that due some BIOS pecularity the EIST states (like 14x, 16x, etc) are not properly configured in the ACPI BIOS, and Windows fails to read the in-between states. In which case even though EIST is active, Windows may not be able to fully utilize its clock speed steps granularity.

[aruff]

Thx.

​I look at that and give you feedback

[aruff]

Hello,

sorry for the delay.

I have done some investigation into EIST and C1E and all other things that you have mentioned.
But, it is still the same regardless I switch those things. The measurement in AIDA is still the same and obviously wrong.

How funny is this clause of you:
>BTW, MSI Command Center, at least on the shot you've linked above, doesn't seem to be all that accurate. 100.0 x 40.0 = 4021.7 MHz? I don't think so... And when a software is obviously inaccurate, I wouldn't use it as a reference in a discussion.

________

 

If you look at this Video you can see that AIDA is doing the same wrong calculation as you characterize in your completely wrong statement.

In the Video you can see that AIDA will went to some different Values even if the “multiplication†is the same. So, why is your Software so inaccurate and how can we use it as a reference?

Please look at this Video and please explain me the difference between Taskmanager and Aida?

 

http://www.allesruff.de/Aida/cpu_boost.mp4

[aruff]

And, before you stated, no, it is not driven by the beta version of Aida.
​The issue is the same even when I use the stable version

[Fiery]

In your video I cannot see any miscalculated frequency readings in AIDA64. 4499.8 MHz instead of 4500.0 MHz exactly is due to slight fluctuations in the BCLK (APIC Clock) reading. It's way less than what I've mentioned above. MSI Command Center miscalculated (or incorrectly measured) the core clock frequency by almost 22 MHz. And even if it does that, the BCLK should be adjusted in a similar fashion. "100.0 x 40.0 = 4021.7" is not something you could take as mathematically correct.

[aruff]

OK, for me as a "normal" Person this difference is irrelevant.
You say, your method viewing on the BCLK is much more accurate then the method from another developer.
Aha.

So, and what is about the different between AIDA and the Taskmanager?
And, you are right, the Taskmanager measure the same as the MSI Tool.

So, you wrote some post before that you X99 Board will measure correct and show some differences between sleep mode and full performance.

In my Video you can see that the Taskmanager will do, and the MSI Tool also.
Why did  Aida measure completely other on my board?
Is this a board failure? Think not, because two other tool measure some difference.
Only Aida will not.

Please explain

[Fiery]

I'm the lead developer behind AIDA64, hence I know how that software works and what exactly it does. I'm not the developer behind Windows Task Manager or MSI Command Center, hence I have no idea what they do and how they work. I can only guess, but that's just a guess. I've already told you what I think they do: they measure average CPU clock over a certain period of time (short period, but still a period of time), and not the actual CPU clock frequency at a certain moment. Your CPU will never work at 4022 MHz when you use 100 MHz BCLK, so that single reading alone means MSI Command Center doesn't measure the CPU clock frequency in the proper way. It seems you stick to debating whether "proper" is really the way it should work. AIDA64 has its roots going back to over 20 years. It does it in a way that we believe, according to our over 2-decade experience, is the right way. CPU-Z, HWiNFO and several other monitoring software use exactly the same method to measure CPU frequency. BTW, CPU-Z and HWiNFO are also not exactly newcomers in this business. If you believe MSI Command Center and Task Manager both do a better job than AIDA64, CPU-Z, HWiNFO and other 3rd party monitoring software, then we have to agree to disagree. If this single issue -- where in my opinion you're clearly wrong -- is a make or break issue about using our software, then I'm afraid the only possible solution is for you to request a refund, get your money back, stop using AIDA64, and keep using Task Manager and MSI Command Center instead. Because we cannot and would not switch to using an average CPU clock frequency measurement method. Our method is the best and proper way to do it, and we wouldn't settle with an inferior solution.

[aruff]

Ok, I understand.
I think we go round in circles

I have no idea why you cannot see any middle steps between the min and max ratios. We've just checked: on our LGA2011-v3 system (Gigabyte X99-UD4 + Core i7-5820K) clock ratio fluctuates between 12x (min) and 36x (max) with several in-between steps, like 30x, 33x, 34x, etc. What kind of system do you use? What happens if you return to default settings, so no OC?

I want exactly this!
“Aida can see the middle steps†but obviously not on my board
Taskmanger see the middle steps, MSI can see the middle steps and of course on my board regardless any switches, configuration or something else.

And your answer is that they are wrong but you not?
Did you not think that this must be a failure by Aida and that it is not a failure by the other developers?

What can I do that my configuration (and there is nothing special with) will also “show the middle steps†like your configuration mentioned above?

[Fiery]

I want exactly this!

“Aida can see the middle steps” but obviously not on my board

Taskmanger see the middle steps, MSI can see the middle steps and of course on my board regardless any switches, configuration or something else.

And your answer is that they are wrong but you not?

Did you not think that this must be a failure by Aida and that it is not a failure by the other developers?

What I'm saying is that since on our motherboard -- and also several other motherboards with the same chipset and same CPU -- AIDA64 can track the CPU changing multipliers in the right granularity, there's nothing wrong with the software. There must be a configuration issue or other pecularity on your system that makes your CPU use only a small number of multiplier steps.

 

What can I do that my configuration (and there is nothing special with) will also “show the middle steps” like your configuration mentioned above?

First, please tell me why would that be beneficial? Your CPU seems to be able to utilize its power management features properly, so at idle it can down-clock itself to save power (it can switch to LFM). It can also switch up to the right high multiplier (it can switch to HFM) when the system is under load. What's wrong with it then? Why would it be beneficial to use the middle multiplier steps? And even if you miss those, what can we as software developers do about it? If there was a simple solution, I'd of course reveal it for you, but I've never seen a LGA2011 CPU behave like yours. To me it seems like a firmware (BIOS) issue or system configuration issue. If one can even call that an issue When both LFM and HFM modes work, ie. the CPU can both provide the expected performance and operate with little power at idle, I'd say it's working properly.

[aruff]

OK, again.

Other tools can, with the same Bios and same configuration.
Why can Aida not?
What is the different?

 

It is not a question why it is for me beneficial.
It is the question why you develop such a feature, so it must be beneficial for you.
If you provide such a feature I expect that it works correctly.

Are there more parts inside the software what I expect beneficial and you not?
Or vice versa?

 

I’m with you, if such a feature does not work within my configuration, it may be a board issue or something like that.
But, other tools can do it the right way. So, there must be an issue on your side, does it?
I think that the question about “beneficial†shows the way you see it.
I buy Aida as “information†Software,

 

 

AIDA64 Extreme has a hardware detection engine unrivaled in its class. It provides detailed information about installed software and offers diagnostic functions and support for overclocking.

I need the support for the overclocking! And first of all, it is my matter for what I need it.
But, if it works correct I promise you that I tell you.

[aruff]

There must be a configuration issue or other pecularity on your system that makes your CPU use only a small number of multiplier steps.

So, my CPU does only a small numbers of steps, even if other Software Shows more steps?