Overclock.net banner
1 - 20 of 24 Posts

·
Registered
Joined
·
189 Posts
Discussion Starter · #1 ·
Hello fellow overclockers. So, my 9900K has been working fine for over a year and is overclocked to 5Ghz Core and 4.7 Ring and it always boosted to the max overclocked speed as indicated in HWInfo under the core effective clock reading. I noticed recently when doing a few benchmarks with Cinebench that it now only get up to 4950mhz. I've tried to narrow it down to what could suddenly be affecting this, but I'm at a loss. I've literally stopped every piece of software and retested and intermittently it does start to function at 5Ghz at times.
 

·
Registered
Joined
·
367 Posts
Maybe as processors get worn out they require more voltage to hit the same frequencies. The higher voltage required for 5GHz might be pushing your CPU further into a thermal situation. Maybe. I could be wrong - this might not happen automatically (might not be implemented yet - if not, I'm sure it'll be how silicon works in a decade or less)

If I were you I'd clock the chip to 5GHz all core manually, it'd fix your problem and your system will feel much snappier during normal daily tasks, because you don't have to wait for the system to boost to 5GHz, and all tasks will be done at 5GHz instead of some at 2-3. Idle power consumption only goes up 10-12 watts :)
 

·
Registered
Joined
·
189 Posts
Discussion Starter · #4 ·
Maybe as processors get worn out they require more voltage to hit the same frequencies. The higher voltage required for 5GHz might be pushing your CPU further into a thermal situation. Maybe. I could be wrong - this might not happen automatically (might not be implemented yet - if not, I'm sure it'll be how silicon works in a decade or less)

If I were you I'd clock the chip to 5GHz all core manually, it'd fix your problem and your system will feel much snappier during normal daily tasks, because you don't have to wait for the system to boost to 5GHz, and all tasks will be done at 5GHz instead of some at 2-3. Idle power consumption only goes up 10-12 watts :)
I was probably vague with the finer points of my overclock. I have manually clocked each of the 8 cores to 5Ghz. With regard to the clock maintaining the 5Ghz on each core, the analysis software that is HWinfo has much more measurement readouts than any of the other software out there. Similar software like HWMonitor will not give you a reading for "Effective Clock Rate" and you would never know that your cores aren't reaching your dialled in speeds. I can also confirm that I have edited the power plan for maximum cpu power and no energy saving.

I will try upping the VCore voltage to next level and see if that corrects it. Its not a big issue in the grand scheme of things, but just grates my OCD in having a perfectly operating system to my specifications. I'll be looking to upgrade my entire system later this year when Zen 4 hits the ground.
 

·
Registered
Joined
·
367 Posts
I was probably vague with the finer points of my overclock. I have manually clocked each of the 8 cores to 5Ghz. With regard to the clock maintaining the 5Ghz on each core, the analysis software that is HWinfo has much more measurement readouts than any of the other software out there. Similar software like HWMonitor will not give you a reading for "Effective Clock Rate" and you would never know that your cores aren't reaching your dialled in speeds. I can also confirm that I have edited the power plan for maximum cpu power and no energy saving.

I will try upping the VCore voltage to next level and see if that corrects it. Its not a big issue in the grand scheme of things, but just grates my OCD in having a perfectly operating system to my specifications. I'll be looking to upgrade my entire system later this year when Zen 4 hits the ground.
Font Number Screenshot Pattern Parallel



Are you saying HWiNFO is showing less than 5000 where mine are 5000? (and that you have it set to 5000 in the BIOS)

I like HWiNFO64, it's the greatest. And AIDA64, I have both. AIDA is missing some things (even sensor data), which is surprising since it's paid software. But it has many great things HWiNFO doesn't.
 

·
Registered
10850k, Apex XIII, RTX3080, G.Skill 3200MHz Cl14 GTZR
Joined
·
58 Posts
I know it's a virus from microsoft. I've been monitoring it since March.

Picture 1: MS without virus:
Product Font Screenshot Line Building

17071 Points

Picture 2: With MS virus:
Furniture Light Product Font Screenshot

15809 Points
I've reinstalled Windows a lot of times and error, error error ... Core Effective Clock is low. I did a lot of Benchmarking until March, I know my PC well. OK, new installation without internet, drivers are all installed and the PC works without error!!!
I tested it for several hours without internet, totally error free. OK, let's join the net ...

This is a must see!
8:50, still good, KB890830 finish 10:20 and next test is .... sh..t. at 10:50.

I made a video of it because it can't be disabled. I have disabled Win10 updates everywhere, but it will infect my machine after 8 minutes.
There are 2 solutions.
One is the complete exclusion of MS, such as ReviOS. I used it for 2 weeks, the PC works perfectly as it used to.
Another solution is to turn off Windows Defender, although it's not easy. Automatically reconnects in a few minutes, I tried in many ways, always activates itself.
The 12th generation doesn't have this bug, it's not planned to become obsolete yet, only the older 9th, 10th and 11th generations.
No one cares, unfortunately, I'm glad someone noticed. Cinebench R23 test is 16,000 vs 17,000 points, no small difference.


"Buy a new one, that's the motto. Does the old one work? No problem, we'll break it!"

Mod:
Try out this: Method 2 works

Mod2: just 10 min long works it, too.
 

·
Registered
Joined
·
189 Posts
Discussion Starter · #7 · (Edited)
I tried upping my VCCIO and VCCSA from 1.15V to 1.2V and initially it worked but soon failed to reach the golden value. Just tlike the recent member just posted, this is probably something Windows 11 has borked on my system with one of their crap recent updates!
 

·
Registered
10850k, Apex XIII, RTX3080, G.Skill 3200MHz Cl14 GTZR
Joined
·
58 Posts
You don't believe me, np. Tested W11 and 10,no matter. Pls disable the Realtime Protection, and test it. Unfortunately, it activates again and again every 10 minutes, making it impossible to turn it off. Please rename this thread in the topics to something right. (Effective Clock too low or something) Here's a test with a graph to better see the whole situation.
Started PC, and works ~ 10 min, and 5GHz Eff. Clock it goes bad due to windows defender.

Rectangle Computer Font Screenshot Software


Manually disabled:
Computer Personal computer Rectangle Font Screenshot


This ~ 50MHz would not even interest, but see the Cinebench test -1200-1500 points. **** Intel and Microsoft!
 

·
Registered
Joined
·
1,034 Posts
Effective Clock too low
I have been following this issue for a while. You are not the first user to trace the reduced effective clock problem back to Windows Defender.

Can you do some testing for me? The next time the clocks are down, can you try running ThrottleStop 9.3? This program has some code in it that seems to snap some Intel computers out of this low effective clock mode. I am still not sure if this is 100% a Windows Defender issue or if this is partially an Intel 9th and 10th Gen hardware level issue.


Click on Show older versions to find version 9.3. This version applies some fixes automatically while the newer versions do not.

Font Rectangle Parallel Screenshot Technology


I have never had this problem on my computer so I had a look at my Windows Defender settings.
I have Real-time protection permanently turned off. That is probably why I do not have this issue.

Rectangle Font Handwriting Circle Magenta
 

·
Registered
Joined
·
1,034 Posts
I turned Real-time protection back on so I could do some further testing.

Rectangle Font Circle Number Paper product


Using Windows 10, I ran Cinebench R23 for some consistent load. The reduced Effective Clock was easy to see. The cores and threads were averaging 4951.7 MHz.

Font Screenshot Rectangle Technology Software



Next I ran ThrottleStop 9.3
Before ThrottleStop does anything or makes any changes, it displays a warning message so a user can safely exit the program.

Font Rectangle Screenshot Number Parallel


Now here is the weird part. Without agreeing to anything and before any changes are supposedly made, the Effective Clock problem is suddenly gone. All cores and threads are happily running at 5000.0 MHz. Full performance has been restored.

Font Rectangle Parallel Software Electronic device


This issue has been around for at least a year. Many have noticed this problem and shrugged it off. A drop of only 50 MHz in the Effective Clock might seem like no big deal but it drops full load benchmark scores a lot more than that.

Windows Defender is definitely involved but I have not been able to understand how running a different program seems to solve this problem. Other users have reported that if they exit ThrottleStop, the problem does not return. I am not sure if this Defender bug will return after a sleep resume or hibernate cycle or it might only return after a full reboot.

Hopefully some more testing next week will help me get to the bottom of this Windows Defender bug.

Edit - After exiting ThrottleStop 9.3 and after doing a quick sleep resume cycle, the reduced effective clock bug returned immediately.

Did some Cinebench R23 testing.

R23 with Reduced Effective Clock
16043
16048
16013

R23 after ThrottleStop 9.3 fix with full 5000.0 MHz
16925
16931
16903

R23 after fix with full 4700.0 MHz
15928
15913
15933

The 50 MHz drop in effective clock causes a performance drop equivalent to reducing the CPU speed by almost 300 MHz.
 

·
Registered
Joined
·
25 Posts
This issue has been around for at least a year. Many have noticed this problem and shrugged it off. A drop of only 50 MHz in the Effective Clock might seem like no big deal but it drops full load benchmark scores a lot more than that.

Windows Defender is definitely involved but I have not been able to understand how running a different program seems to solve this problem. Other users have reported that if they exit ThrottleStop, the problem does not return. I am not sure if this Defender bug will return after a sleep resume or hibernate cycle or it might only return after a full reboot.

Hopefully some more testing next week will help me get to the bottom of this Windows Defender bug.

Edit - After exiting ThrottleStop 9.3 and after doing a quick sleep resume cycle, the reduced effective clock bug returned immediately.

Did some Cinebench R23 testing.

R23 with Reduced Effective Clock
16043
16048
16013

R23 after ThrottleStop 9.3 fix with full 5000.0 MHz
16925
16931
16903

R23 after fix with full 4700.0 MHz
15928
15913
15933

The 50 MHz drop in effective clock causes a performance drop equivalent to reducing the CPU speed by almost 300 MHz.
My brother and I spent a day this last week diagnosing this, and we found ThrottleStop fixed this issue as well. I came in to recommend it, but it looks like you've reached the same point I did.

I also made some stuff to auto-execute and shut down at startup and from resuming from sleep using Task Scheduler. I attached the files, just change .txt to .zip and check it out.

Also, how are you getting ~16k in R23 on a 9900k?? I'm on a 9900KF and so's my brother, and we only hit ~13400... We're also running 5.0, x47, my ram is OC'd to 4133Mhz and his is XMP at 3600Mhz.
 

Attachments

·
Just hangin' out
Joined
·
1,583 Posts
I only get 13,700 on my 9900K @ 5000MHz, 4700MHz ring, 4000MHz DDR4.

16,000 is amazing!

EDIT:

Just noticed, HWiNFO screenshot shows 20 cores, so not a 9900K.

Nevermind.
 

·
Registered
Joined
·
25 Posts
I only get 13,700 on my 9900K @ 5000MHz, 4700MHz ring, 4000MHz DDR4.

16,000 is amazing!

EDIT:

Just noticed, HWiNFO screenshot shows 20 cores, so not a 9900K.

Nevermind.
oh yeah, didn't realize it was someone else, lol.
 

·
Registered
Joined
·
1,034 Posts
The 10850K has two more cores to help them Cinebench scores.

I just wanted to show that the Windows Defender bug happens to both Intel 9th and 10th Gen CPUs. Not sure about 11th or 12th Gen yet.
 

·
Registered
10850k, Apex XIII, RTX3080, G.Skill 3200MHz Cl14 GTZR
Joined
·
58 Posts
"
I have been following this issue for a while. You are not the first user to trace the reduced effective clock problem back to Windows Defender.

Can you do some testing for me? The next time the clocks are down, can you try running ThrottleStop 9.3? This program has some code in it that seems to snap some Intel computers out of this low effective clock mode. I am still not sure if this is 100% a Windows Defender issue or if this is partially an Intel 9th and 10th Gen hardware level issue."


Sure, first, sorry, i'm hun, it's not my main language. I use sometimes google translate, just a little speak english.
ThrottleStop looks like works fine, but i'm not happy. I don't wanna third party software, just want my old PC back.
We tested this weird situation with some friends, and 9, 10 and 11! cpu-s throttling. 12th ADL not!
Here is my test for u, over 10 mins, and without problem.
Colorfulness Product Font Screenshot Rectangle


Easyest way use ReviOS, that has no connect to MS, sure, has no problem, works fine. Tested two weeks long. Original Windows works not, 100%.
Unfortunately, people do not care, "only 50MHz", but this is not true, power loss is too large.
In my opinion, intentional slowdown, "take the 12th generation ADL family!"
One curiosity: Friend from the Hungarian forum, using ESET. If you watch my video, update KB890830 is considered dangerous.
Azure Computer Font Rectangle Line


Soory, whole day was work, i'm tired. I just wanna back my old PC, i don't wanna buy new PC every half year, as Intel and ms thinking. Comet Lake is good, lol, if works.
Cheers ;)
 

·
Registered
Joined
·
189 Posts
Discussion Starter · #16 ·
This is the way of the big corporations unfortunately. My 9900K may be a few generations old now, but its still plenty fast for my needs, which are mainly FPS gaming. They literally force you to upgrade to a new generation, by making older hardware fail to perform at their optimum levels. I will be upgrading later this year to most probably the new Ryzen platform and my current PC will be handed down to my kids. I usually follow this cycle with my older systems and the one that leaves the house, goes to a person most in need of a working system, albeit with 50mhz missing off their overclock.
 

·
Registered
Joined
·
189 Posts
Discussion Starter · #17 ·
You guys that have spent vast amounts of time researching this issue, I take my hat off to you all! I just tried the disabling of "Real Time Protection" and it did indeed restore the CPU back to full speed when running Cinebench.

I wish, I could figure out my final problem. I get this logged in the events, but I don't actually observe any apps crashing at all. It doesn't help, that there's no trace of what is causing the supposed app crash.

Description
Faulting Application Path: bad_module_info

Problem signature
Problem Event Name: APPCRASH
Application Name: bad_module_info
Application Version: 0.0.0.0
Application Timestamp: 00000000
Fault Module Name: StackHash_0000
Fault Module Version: 0.0.0.0
Fault Module Timestamp: 00000000
Exception Code: 00000000
Exception Offset: PCH_84
OS Version: 10.0.22000.2.0.0.256.48
Locale ID: 2057
Additional Information 1: 0000
Additional Information 2: 00000000000000000000000000000000
Additional Information 3: 0000
Additional Information 4: 00000000000000000000000000000000

Extra information about the problem
Bucket ID: 48565509f28d3c5915fb08e8f96ea833 (1583869490661074995)
 

·
Registered
10850k, Apex XIII, RTX3080, G.Skill 3200MHz Cl14 GTZR
Joined
·
58 Posts
You guys that have spent vast amounts of time researching this issue, I take my hat off to you all! I just tried the disabling of "Real Time Protection" and it did indeed restore the CPU back to full speed when running Cinebench.

I wish, I could figure out my final problem. I get this logged in the events, but I don't actually observe any apps crashing at all. It doesn't help, that there's no trace of what is causing the supposed app crash.

Description
Faulting Application Path: bad_module_info

Problem signature
Problem Event Name: APPCRASH
Application Name: bad_module_info
Application Version: 0.0.0.0
Application Timestamp: 00000000
Fault Module Name: StackHash_0000
Fault Module Version: 0.0.0.0
Fault Module Timestamp: 00000000
Exception Code: 00000000
Exception Offset: PCH_84
OS Version: 10.0.22000.2.0.0.256.48
Locale ID: 2057
Additional Information 1: 0000
Additional Information 2: 00000000000000000000000000000000
Additional Information 3: 0000
Additional Information 4: 00000000000000000000000000000000

Extra information about the problem
Bucket ID: 48565509f28d3c5915fb08e8f96ea833 (1583869490661074995)
Maybe help: Link
 

·
Registered
Joined
·
189 Posts
Discussion Starter · #19 ·
Maybe help: Link
Thanks for the link, but I've already tried these fixes. Fortunately, this doesn't appear to be affecting the OS or any of my programs from working correctly. I get no game crashes or any other program issues. Its just annoying that something is failing on Windows start-up, but then it just stays silent without any indication of what or where it is.
 

·
Registered
Joined
·
501 Posts
I tried running CB R23 with real-time protection on and off. No difference of note (maybe 20 points?). Around 13100 points either way.

9900K + Win10 + 16 GB XMP 3200 MHz CL16. No CPU overclock, I just have PL1/PL2 raised so I don't throttle on CB. I don't see any difference in clocks either. Pegged at 4.7 GHz all cores when running the bench.
 
1 - 20 of 24 Posts
Top