pastrychef's build - Asus Maximus VIII Gene - i7-7700K - GTX 1080

[pastrychef]

OpenCore doesn't have X86PlatformPlugin so I used AppleIntelInfo.kext. Here are the results:

First image is OpenCore and Second one is Clover using iMac17,1

View attachment 433308 View attachment 433307

As you can see that OpenCore mostly rest at 10 after briefly touching 8. Clover is at approx. 20 so there's significate gap there. Does this help?

Both should be proving more than enough to drop it down to 800MHz at idle.

I still think there's some kind of process running that's preventing it from going down.

 

[remington]

Both should be proving more than enough to drop it down to 800MHz at idle.

I still think there's some kind of process running that's preventing it from going down.

I have no idea what's preventing it from going down some more as it's running on clean Catalina without any apps. I've seen many posts in other threads that people are complaining that 6700K is idling at higher frequency. It seems that nobody was able to figure out how to get it down to 800 and stay there while idle.

So the only things I think can cause this:

1) Clover/OpenCore patches or drivers not managing 6700K very well.
2) Apple designed macOS to manage 6700K at higher frequency.
3) It could be the chip design.
4) Motherboard design or BIOS.

I have not changed anything in BIOS in regards to CPU. My generated SSDT.aml doesn't really help and it only pushes the frequencies higher so I'm running Clover with PluginType enabled and nothing in OpenCore.

 

[joevt]

As you can see that OpenCore mostly rest at 10 after briefly touching 8. Clover is at approx. 20 so there's significate gap there. Does this help?

I don't think you can compare them that way. A new line is generated in the output only when a new multiplier is encountered (notice that no line is the same and that every new line is longer than the previous line and that each line contains all the multipliers of the previous line). It does not convey information about how often a multiplier is encountered.

They both have multipliers between 8 and 42. What is interesting is that Clover hasn't encountered multipliers 9 to 14. Maybe you need to run it longer. Try clicking stuff or dragging a window.

 

[remington]

I ran CPUID in Windows 10 on my same machine. It was able to comfortably idle at 803 MHz so we can rule out #3 and #4 above.

 

[remington]

I don't think you can compare them that way. A new line is generated in the output only when a new multiplier is encountered (notice that no line is the same and that every new line is longer than the previous line and that each line contains all the multipliers of the previous line). It does not convey information about how often a multiplier is encountered.

They both have multipliers between 8 and 42. What is interesting is that Clover hasn't encountered multipliers 9 to 14. Maybe you need to run it longer. Try clicking stuff or dragging a window.

I understand but that's not the point. The point is trying to get macOS to idle at 800 MHz... not higher. I ran the test for 2 to 3 days with various options, different SSDTs and I was not able to get it to stay at 800 MHz. Many of the different options in Clover/OpenCore I've tried just made it worse so it's best to leave the options at default. You can see my snapshots in my post #1,267 as it doesn't idle at 800 MHz.

So the culprit is either Clover/OpenCore or macOS itself.

 

[joevt]

I understand but that's not the point. The point is trying to get macOS to idle at 800 MHz... not higher. I ran the test for 2 to 3 days with various options, different SSDTs and I was not able to get it to stay at 800 MHz. Many of the different options in Clover/OpenCore I've tried just made it worse so it's best to leave the options at default. You can see my snapshots in my post #1,267 as it doesn't idle at 800 MHz.

So the culprit is either Clover/OpenCore or macOS itself.

An older version of Intel Power Gadget (3.0.3) would draw the graph differently. I think it would draw the minimum instead of the average so you would see it dip below 1GHz when idling. In the current version of Intel Power Gadget (3.6.1) you see the line being drawn for the average, but the minimum is still indicated as the blue shading in your screen shot. The OpenCore screen shot is strange because the minimum doesn't go below 1 GHz and the max is wrong - there are times the graph goes about 4 GHz but the max says 2.9 GHz.

 

[remington]

An older version of Intel Power Gadget (3.0.3) would draw the graph differently. I think it would draw the minimum instead of the average so you would see it dip below 1GHz when idling. In the current version of Intel Power Gadget (3.6.1) you see the line being drawn for the average, but the minimum is still indicated as the blue shading in your screen shot. The OpenCore screen shot is strange because the minimum doesn't go below 1 GHz and the max is wrong - there are times the graph goes about 4 GHz but the max says 2.9 GHz.

RehabMan said once before that Intel or iStat graphs are not reliable as they may not always read the frequency correctly. Do you know of a reliable CPU monitoring software?

 

[joevt]

RehabMan said once before that Intel or iStat graphs are not reliable as they may not always read the frequency correctly. Do you know of a reliable CPU monitoring software?

Nope. Are you using Intel Power Gadget 3.6.1?

 

[remington]

Nope. Are you using Intel Power Gadget 3.6.1?

Yes and iStat 6.40