Overclock.net banner

12341 - 12360 of 14568 Posts

·
Registered
Joined
·
253 Posts

·
Registered
Joined
·
414 Posts

Ryzen Master just shows you Windows preferred cores and not necessarily the best(see article from Anandtech). I found out about this while tuning curve - the second best core that RM was showing was actually my worst. I was only able to stabilize that particular core by adding more voltage via CO.
I want some more info about this. Trying to learn .. whare are settings in bios and so on. Pleas.

Skickat från min SM-G973F via Tapatalk
 

·
Registered
Joined
·
209 Posts

Ryzen Master just shows you Windows preferred cores and not necessarily the best(see article from Anandtech). I found out about this while tuning curve - the second best core that RM was showing was actually my worst. I was only able to stabilize that particular core by adding more voltage via CO.
Okay but can you explain what you did to see the best cores? If Ryzen Master aint showing it properly I mean.
 

·
Registered
Joined
·
414 Posts
Okay but can you explain what you did to see the best cores? If Ryzen Master aint showing it properly I mean.
The discrepancy that’s been discussed by the community in recent weeks, and that’s been prevalent since the launch of the Ryzen 3000 series in July, is that in the majority of situations and setups, the actual CPU cores that are being loaded in the operating system under single-threaded or lightly threaded workloads mostly never matched the best CPU cores as reported by Ryzen Master. This can be seen with any generic monitoring utility such as the task manager.

Skickat från min SM-G973F via Tapatalk
 

·
Registered
Joined
·
253 Posts
Okay but can you explain what you did to see the best cores? If Ryzen Master aint showing it properly I mean.
I set all my cores to -25, +75mhz, crashed, set core 0 to -10, crashed, set core 0 back to -25 and core 1 to -10, crashed, and so on until it didn't crash anymore when I got to core 6.
 

·
Registered
Joined
·
209 Posts
I set all my cores to -25, +75mhz, crashed, set core 0 to -10, crashed, set core 0 back to -25 and core 1 to -10, crashed, and so on until it didn't crash anymore when I got to core 6.
okay so you have 1 core @ -10 with 75 mhz and rest -25?

I feel I dont understand how the curve optimiser works yet, havent fiddle to much yet
 

·
Registered
Joined
·
253 Posts
okay so you have 1 core @ -10 with 75 mhz and rest -25?

I feel I dont understand how the curve optimiser works yet, havent fiddle to much yet
Correct but I got an error in R20 on my third run so I had to lower core 7 to -10 as well making two cores at -10 and the rest -25. However, in doing that, my achievable clocks in Warzone and BFV lowered. So I'm testing 7 cores at -25 with one core at -10, +50mhz right now.
 

·
Registered
Joined
·
127 Posts
Correct but I got an error in R20 on my third run so I had to lower core 7 to -10 as well making two cores at -10 and the rest -25. However, in doing that, my achievable clocks in Warzone and BFV lowered. So I'm testing 7 cores at -25 with one core at -10, +50mhz right now.
How much is your effective clock at warzone (because that is the important one)?
I have my 5900X at -15 two best cores and -20 others, just for comparison.
 

·
Registered
Joined
·
253 Posts
How much is your effective clock at warzone (because that is the important one)?
I have my 5900X at -15 two best cores and -20 others, just for comparison.
With my current setting, between 4750 and 4825. I see I'm mostly at 4775 and 4800 all cores. What about yours?
 

·
Registered
Joined
·
127 Posts
With my current setting, between 4750 and 4825. I see I'm mostly at 4775 and 4800 all cores. What about yours?
Thats impossible for effective clock, warzone never use 12 cores fully loaded.
That has to be the core clock, take a look at effective one in HWINFO, thats the important one

Edit: There is a better way of testing ur cpu, just download 3dmark and do a timespy run, then check ur cpu score, here is mine:
NVIDIA GeForce RTX 3080 video card benchmark result - AMD Ryzen 9 5900X,Gigabyte Technology Co., Ltd. X570 AORUS MASTER (3dmark.com)
 

·
Registered
Joined
·
209 Posts
Correct but I got an error in R20 on my third run so I had to lower core 7 to -10 as well making two cores at -10 and the rest -25. However, in doing that, my achievable clocks in Warzone and BFV lowered. So I'm testing 7 cores at -25 with one core at -10, +50mhz right now.
yeah but how do you know what the best cores is?
 

·
Registered
Joined
·
253 Posts
Thats impossible for effective clock, warzone never use 12 cores fully loaded.
That has to be the core clock, take a look at effective one in HWINFO, thats the important one
Ah sorry I see what you meant. My effective clocks are lower than my core clocks. Around 4.6 but i don't see the effective clock the way I think you're thinking when playing games. I see the effective clock as indication on how heavily the core is loaded. If it was 100% (equal to core clock) like on Cinebench, it won't be anywhere near 4.8.

But i know my CO tweaks are working because I confirm with the FPS I get.
 

·
Registered
Joined
·
214 Posts
Still tinkering with CO on my 5800x. HWinfo seems to show consistency as to which cores are my 2 best as well as being favored by the OS in single core tasks. I've found that CB20 or CB23 single core is a good test to see if you are remotely stable on your 2 best cores. If you can pass multiple single core runs without erroring then very likely you have a good setup. OCCT large worked for checking all-core load. I had failures in OCCT at -15 on 6 of my cores and -3 on my 2 best. I scaled back to -10 and -3 and I passed an hour. I'm also finding that maybe it isn't a great idea to crank up the core boost to +200. I'm testing again to see if I can run +150 or +175 and use more negative offset. This should net a more stable overclock and may improve benchmark scores for more consistent results.
 

·
Registered
Joined
·
127 Posts
Ah sorry I see what you meant. My effective clocks are lower than my core clocks. Around 4.6 but i don't see the effective clock the way I think you're thinking when playing games. I see the effective clock as indication on how heavily the core is loaded. If it was 100% (equal to core clock) like on Cinebench, it won't be anywhere near 4.8.

But i know my CO tweaks are working because I confirm with the FPS I get.
Thats completely correct, but is a bit of a mess to measure that that way.
And as i said, warzone doesnt load all cores, so a lot of them will be even at idle in effective clocks.
As said is better to do a fast 3DMARK run and take a look at your cpu score.
It takes you 4 min and is far more objective than the other way.
 

·
Registered
Joined
·
253 Posts
Thats completely correct, but is a bit of a mess to measure that that way.
And as i said, warzone doesnt load all cores, so a lot of them will be even at idle in effective clocks.
As said is better to do a fast 3DMARK run and take a look at your cpu score.
It takes you 4 min and is far more objective than the other way.
I agree thanks for the tip. What is the best test to find CPU score in 3D Mark? I only used PR to test my 3090 OC a couple of months ago.
 

·
Registered
Joined
·
209 Posts
None can help me figure out the best cores or if its a program that shows the best cores?
 

·
Registered
Joined
·
105 Posts
None can help me figure out the best cores or if its a program that shows the best cores?
Ryzen master show what core is best the cpu reports it it's self. Hwinfo also labels cores under (per#x/y) you can also watch which core clocks the highest.

If C2 and C9 is my best cores, what counts as core 2 in bios? starts from core 0 in bios and not 1.
core 1 to 8 is equal to 0 to 7. HWinfo also labels from 0
 

·
Registered
Joined
·
127 Posts
I agree thanks for the tip. What is the best test to find CPU score in 3D Mark? I only used PR to test my 3090 OC a couple of months ago.
Timespy should be fine
 
12341 - 12360 of 14568 Posts
Top