• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Cinebench r20 performance loss

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

xxNAPxx

Member
Joined
Jan 8, 2019
Hello everyone, I have an issue with Cinebench after the last windows Update 2004. with my I9-9900k clocked at 5.05 ghz I was able to score up to 5315points with minimum of 5277points,after the update and messing with some services and HPET I lost up to 100 points less now scoring max 5242point with minimum of 5165,why this? Do you have any idea what affect the score so much in windows? I tried everything, reenabling services, reactivate HPET, uninstall and reinstall drivers, reinstall cinebench, nothing worked, this is driving me crazy, any help please?
 
I'm guessing it may have something to do with more security code added to Windows 10 with the 2004 Feature update, similar to what was done last year in the wake of Meltdown and Specter and the other chip-based vulnerability revelations that followed.
 
So do you think it is nothing releted on what I did but just something on windows side? I didn't find any comparision just to prove it is not me but microsoft
 
Well, you indicated you reversed "what you did" and it didn't improve the score.
 
Pretty much yeah, I did some research and it looks like it depends on windows, it looks like the last may update actually decrease the cpu performance on Cinebench
 
a background process may have been running or something...cinebench can 'pop' higher scores now and again.. so that may be something.
 
I've run cinebench with task mamager open and the only background process that takes some cpu percentage is "system interrupts " within the 0.3/0.7% range, no trace of high cpu usage processes. Ialso changed the priority of the Cinebench benchmark up to realtime and this didn't get the score back to up 5300
 
It's also ~2% difference. The only way to test would be on fresh install, different versions, and see.
 
Well I should try on a fresh install of thos version to check if I broke something, because at the same conditions with the previous version the score was ok
 
Do you have an installer for the previous version? I believe now you can only get 2004 from the Microsoft website where the ISO is downloadable.
 
Do you have an installer for the previous version? I believe now you can only get 2004 from the Microsoft website where the ISO is downloadable.

No i Don't have it I have a ISO of the very first version but to do this I have to format and honestly I prefer not to. Probably I understimated the impact this update could have on the system, I trusted Microsoft. I can't say it's all windows fault but whatever optimization I did it is somethig I had on ver 1909 as well like HPET disable and services here is the services I have disabled

9AQ2imt.png

TVSNAgS.png
 
So... one test shows are barely more than margin of error difference...

What about anything else?
 
I'm saying it is one test and it hardly shows a difference in performance... what about other tests? I assume you render on your pc which is why youre focused on cinebench?
 
This is why benching folks run custom versions of windows, some of us have many versions of windows, with everything taken out.
If you are rendering on this rig, why use windows, any linux will be faster.
100 cb? it might be defender running in the background, that small..... who knows.
 
I'm saying it is one test and it hardly shows a difference in performance... what about other tests? I assume you render on your pc which is why youre focused on cinebench?
Probably I am a bit exaggerated but I don't like when I see a performance hit of any kind, it's quite annoying also because I spend time to trim out the system to keep it minimal, clean, and mantained to get the best performance at all times and when I see this drops it drives me nervous
This is why benching folks run custom versions of windows, some of us have many versions of windows, with everything taken out.
If you are rendering on this rig, why use windows, any linux will be faster.
100 cb? it might be defender running in the background, that small..... who knows.
What are these custom versions of windows you are talking about? I'm quite curious, what do you take out? I've taken out many things to be honest but I'm always likely to know new things.
I do not render, just gaming, I just like cinebench to test the cpu for stability and performance and check the temperature, I used to run other tests but I don't remember the results to compare in case I want to repeat them in this scenario
 
One test.. negligible/minimal change. Don't worry about it. :)

Cinebench is too short of s test for stability or temps.

If you're a gamer, test games. Since you don't have a baseline, id move on and enjoy your rig. :)
 
We remove all the stuff that make an os usable in daily life.
language packs
internet connections
fonts
word pad
note pad
windows defender
we make them ourselves and for anything other than running the bench, they are useless.
If you are not benching, I have to agree with earth dog, don't worry about it, just enjoy your rig, we spend way to much time bitching about what we ain't got and not enough time enjoying what we do got.
 
Honestlty I don't see the point of keeping active stuff I don't use, like smart card services or virtual machine stuff, Microsoft gives people out of the box stuff that make the OS pretty rather than performing and user friendly in terms of functions and services that most people may use, removing stuff is actually optimize the OS for personal use with the stuff that you need and not everyone needs.
Cinebench is a way to check the performance of a system if it gets less points it means there is something they or you did to do so, now I cannot fix what they did but I can try to fix what I did and asking here if there is something you know that I don't to get my performance back. I was pretty happy and enjoying my rig before this happened and before Microsoft asked me to update windows with their great updates
 
If gaming is your primary usage scenario, CB is not the app you should be using to judge performance. It only tests the CPU. Most games rely more heavily on the GPU.
 
Back