cancel
Showing results for 
Search instead for 
Did you mean: 

Single core - 100% CPU - high interrupts on ACPI

iransofaraway
Level 7
Hi folks. I bought an Prime X299-Deluxe and a Core i9-7900X - I've had them for months now but keep having a random issue. Sometimes, after rebooting, one of my cores will be maxed at 100% constantly, and checking DPC interrupts shows ACPI.SYS is the culprit. No amount of rebooting, changing drivers, etc seems to correct it - rebooting sometimes fixes it - if I shut down completely and power off the system for like five minutes, then boot back up, the problem sometimes goes away. Sometimes I have to swap around PCI-E cards. It's just bonkers. I've built many computers in my life, and know what I'm doing, but this has me stumped.

I found this other user on Reddit having the same problem: https://www.reddit.com/r/Windows10/comments/7puiip/win_10_bug_acpisys_always_maxes_out_one_of_my_cpu... - and I have that same motherboard and SSD, but slightly different CPU. The Microsoft tech in that thread seemed to indicate that they looked at the logs and it's possibly an issue with ASUS?

Has anyone experienced anything like this on their X299 hardware? Any tips/ways to get it fixed? I'm already running latest BIOS firmware for my board, and latest drivers on everything.
329 Views
86 REPLIES 86

I have the same issue here, x299-a latest bios, 7980xe and 64gb corsair ram, 1050Ti.
I am using this machine for music production, and i am having issues every time i start my computer with latency.
Using latencymon.exe it shows that ACPI.sys gives alot of latency and uses core 2 at 100% all the time.
*

Silent_Scone
Super Moderator
I've been able to replicate this issue on my own system with the Apex. I'll let you when I find a solution.
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

Silent Scone wrote:
I've been able to replicate this issue on my own system with the Apex. I'll let you when I find a solution.


is it allowed to post links to reddit or other forums?

Anyways i read that somwone found a temporary workaround *by Disabling wake-on-magic-packet (in the network adapter's properties)
Have not had the time to try this though since my computer is in my studio.*

Silent_Scone
Super Moderator
Grab this version of the chipset driver: http://dlcdnet.asus.com/pub/ASUS/mb/LGA2066/ROG_RAMPAGE_VI_EXTREME/DRV_MEI_Intel_Consumer_TP_W10_64_...

If running AIDA64, or any other application that is polling devices, close and retest.
Also, uninstall Geforce drivers and Geforce Experience, run DDU driver removal utility and reinstall the latest Geforce driver.
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

Silent_Scone
Super Moderator
Yes, that's fine.

For me, following my post worked - but it's not necessarily the same root cause. I'm also running insider build 171115.

Wake on magic packet isn't enabled by default, either.
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

Silent Scone wrote:
Yes, that's fine.

For me, following my post worked - but it's not necessarily the same root cause. I'm also running insider build 171115.

Wake on magic packet isn't enabled by default, either.


ok ill try your solution.

*https://www.reddit.com/r/Windows10/comments/7puiip/win_10_bug_acpisys_always_maxes_out_one_of_my_cpu...

DooRules
Level 10
I have one core running full blast as well it seems. As soon as I closed out Aida64 it dropped right off. Started it back up and one core picked right up again. I use it to monitor some temps for my Aquaero controller. i guess i will try Hwinfo64 and see if the same thing persists.


For me it seems Aida64 was the culprit. Killed it and tried Hwinfo and on several reboots no core going full bore at startup. I will monitor to see if it stays this way.
7980XE on Asus ApexVI
Zotac 2080Ti
Intel 900P
2 x Samsung 950 Pro
EVGA 1600W
Alienware 34"
Samsung 28" 4K

DooRules wrote:
I have one core running full blast as well it seems. As soon as I closed out Aida64 it dropped right off. Started it back up and one core picked right up again. I use it to monitor some temps for my Aquaero controller. i guess i will try Hwinfo64 and see if the same thing persists.


For me it seems Aida64 was the culprit. Killed it and tried Hwinfo and on several reboots no core going full bore at startup. I will monitor to see if it stays this way.



Update to the latest beta, there is far less latency for me but it's still there.
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

Silent_Scone
Super Moderator
The problem seems to be with AIDA and the OS putting significant overhead on the system. This is easily replicated by disabling and re-enabling Low-Level Sensor Operation under the stability panel in AIDA.

I've contacted AIDA regarding the issue.
13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

Silent Scone wrote:
The problem seems to be with AIDA and the OS putting significant overhead on the system. This is easily replicated by disabling and re-enabling Low-Level Sensor Operation under the stability panel in AIDA.

I've contacted AIDA regarding the issue.



All I did was open task manager and go to cpu performance. Change look to logical processors and you could see one core hammering away. Close Aida and it dropped right off. Open it back up and the same core picked back up again. Little different method but same result.
Good find Silent scone. Thanks for the heads up. 😄

No such issue with Hwinfo.
7980XE on Asus ApexVI
Zotac 2080Ti
Intel 900P
2 x Samsung 950 Pro
EVGA 1600W
Alienware 34"
Samsung 28" 4K