Question Computer Sluggish after CPU Upgrade

[works_arc]

howdy - I've upgraded the CPU in my build from a AMD RYZEN 7 3700X (not performing well, poor bench numbers) to an AMD RYZEN 9 5950X and the result are suboptimal (performing even worse, poor bench numbers)The computer is very slow to respond at login, and every operation after that is just as painful as equally slow. I ran a Userbenchmark test and the results were worrisome. I am hear hoping for guidance. My hope/suspicion is that I have a motherboard problem but I am ill-equipped to diagnose that issue.

I took the following step prior to physically replacing the CPU:

1. Updated Drivers for all current hardware and OS
2. Verified that the AORUS ELITE 570X was compatible.
3. Updated the AORUS BIOS to F37c, this was done via BIOS.

Here are the specs:

• CPU _ AMD RYZEN 9 5950X
• Motherboard_570X AORUS ELITE WIFI
• GPU_GeForce GTX 1660 6BG
• RAM_32 GB G Skills Ripjaws V Series DDR4
• Power Supply_EVGA 850 GOLD
• Storage_ 1TB Samsung 970 EVO M.2 SSD
• OS_Windows 10 64 Bit Pro

 

[kerberos_20]

goo catch!
does that do the same as un-ticking the check box in System Config /boot / advanced options?

yes, but sometimes msconfig gets bugged and doesnt lets you change number of cpus there

 

[works_arc]

so may answers and nobody noticed that his CPU runs 1 core - 1 thread?
no wonder its sluggish

run this command:
bcdedit /set {current} numproc 32
then reboot

Tried this - not recognized as internal or external command. Hmm.

 

[works_arc]

good catch!
does that do the same as un-ticking the check box in System Config /boot / advanced options?

@works_arc : definitely do what kerberos suggests forthwith!

And good benchmarks you can compare CPU performance and GPU performance are Cinebench R20 and 3DMark, the Timespy bench in the suite. They are somewhat synthetic but they're fairly repeatable so you can compare results to other CPU's and GPU's to see how yours compares. That's perfect to see if your system is performing correctly.

System Config /boot / advanced options shows number of processors checked and 1 in the dropdown FYI

 

[works_arc]

good catch!
does that do the same as un-ticking the check box in System Config /boot / advanced options?

@works_arc : definitely do what kerberos suggests forthwith!

And good benchmarks you can compare CPU performance and GPU performance are Cinebench R20 and 3DMark, the Timespy bench in the suite. They are somewhat synthetic but they're fairly repeatable so you can compare results to other CPU's and GPU's to see how yours compares. That's perfect to see if your system is performing correctly.

System Config /boot / advanced options shows number of processors has been checked and the number 1 is in the dropdown FYI

 

[drea.drechsler]

System Config /boot / advanced options shows number of processors has been checked and the number 1 is in the dropdown FYI

That's what needs to change...there are 32 virtual processors in your system. It's usually better to just clear the checkbox so it uses the default which is all available processors (threads).

 

[works_arc]

That's what needs to change...there are 32 virtual processors in your system. It's usually better to just clear the checkbox so it uses the default which is all available processors (threads).

Noted - I'll uncheck that box and reboot.

 

[kerberos_20]

System Config /boot / advanced options shows number of processors checked and 1 in the dropdown FYI

is it changeable there?

bcdedit /set numproc 32 should work, not sure why current identifier didnt pass

 

[works_arc]

is it changeable there?

bcdedit /set numproc 32 should work, not sure why current identifier didnt pass

It's changeable in the config>>advance settings, rebooting now and running the inferior UBM test.

 

[drea.drechsler]

is it changeable there?

bcdedit /set numproc 32 should work, not sure why current identifier didnt pass

@works_arc did you open the command prompt as administrator?

 

[kerberos_20]

@works_arc did you open the command prompt as administrator?

he probably used powershell, command for powershell needs quotes
bcdedit /set "{current}" numproc 32

 

[works_arc]

@works_arc did you open the command prompt as administrator?
[/QUOTE

he probably used powershell, command for powershell needs quotes
bcdedit /set "{current}" numproc 32

The quotes did it - command was performed successfully - rebooting now.

 

[works_arc]

@works_arc did you open the command prompt as administrator?

System normal

https://www.userbenchmark.com/UserRun/53195938

 

[works_arc]

Just to get started.

You have too much background stuff running.

Reboot then wait a few mins then run UBM with the browser closed.
Post a link.

Enable xmp for the ram your running it at low speed.

Thanks for the tip.
https://www.userbenchmark.com/UserRun/53195938

 

[drea.drechsler]

he probably used powershell, command for powershell needs quotes
bcdedit /set "{current}" numproc 32

hah....that's why i go out of my way to use command prompt. I'm still not familiar with the quirks of powershell and would have made the same mistake!

 

[works_arc]

Well, no idea what the cause of that was but everything seems to be in order now - thanks, everyone, for your help.

 

[drea.drechsler]

Well, no idea what the cause of that..

I wish i knew too. It may not be all that common but it's happened to me...and to many others as I recall.

 

[drivinfast247]

avoid userbenchmarks. it is hopelessly biased against AMD products and presents results in a bad light, if not flat out lieing about them. It's useless.

Here's a few citations, but run your own google. It goes on and on and on.

View: https://www.reddit.com/r/buildapc/comments/g2x49q/userbenchmark_should_be_banned/

UserBenchmark Can't Be Trusted, Here's Why - Gizmosphere

For awhile, Userbenchmark was the go-to site for many tech enthusiasts, but it's clear that the site is biased, leading to inaccurate results.

www.gizmosphere.org

View: https://www.youtube.com/watch?v=7sygt_lIeOk

Doesn't matter. LOOK AT THE RESULTS!!! Don't use it as a comparison use it as a diagnostic tool.

 

[drivinfast247]

Thanks for the tip.
https://www.userbenchmark.com/UserRun/53195938

WTH do you have running in the background? How do you have 45% bg usage after a restart?

 

[Karadjgne]

Moved from a 3k series Ryzen to a 5k series Ryzen, I'd suggest a windows reset at a minimum as many of the windows settings will be based on the old cpu. Still function within acceptable limits, but often not to the new cpu ability.

 

[Bob.B]

Doesn't matter. LOOK AT THE RESULTS!!! Don't use it as a comparison use it as a diagnostic tool.

Your right it does not matter UBM haters just love to carp.

 

[Karadjgne]

Doesn't matter. LOOK AT THE RESULTS!!! Don't use it as a comparison use it as a diagnostic tool.

Even that has to be taken with a grain of salt, or 3. You can run that test 3x, back to back and get different results. I find it's only real usefulness is more of a determination tool than a diagnostic tool. OP says something is wrong, system sluggish, UBM pointed at everything being slow, but also caught the single core and lack of xmp.

Which should have had almost no affect on storage performance, yet did, because it's all linked to the cpu, which it shouldn't be. All tests should be seperate, but aren't, that makes diagnostics somewhat hit or miss.

I'm not a hater, I just find UBM of limited usefulness.

 

[works_arc]

Moved from a 3k series Ryzen to a 5k series Ryzen, I'd suggest a windows reset at a minimum as many of the windows settings will be based on the old cpu. Still function within acceptable limits, but often not to the new cpu ability.

Yep, the windows reset fixed this up good and proper.