Overclock.net banner
Cancel
Create thread New Forums

5600x PBO negative curve offset and idle/low load black screen crashes

1 reading
Jump to Latest
1 - 20 of 38 Posts
X

xxbluedragonxx

· Registered
Joined
·
12 Posts
Discussion starter · #1 · (Edited)
I’m hoping I can get some help on why I’m getting random black screen restarts with my 5600x. This is happening when idling or light loads like browsing.

I have BIOS all defaults except PBO negative offset (all cores) and PBO Limits Disabled.

it’s in an MSI B550 Tomahawk motherboard with latest BIOS. Windows 10 “balanced” power plan.

How can I reliably test non-load crashes? How can I prevent?

Do i just have to wait for random crash and then adjust individual core offset based on WHEA CPID logger output?

I have tried 3 different 5600x CPUs and replaced the motherboard twice but still this random idle/low load restart happens when using negative PBO offsets over 26-27…

I guess is there any reliable way to test for this kind of crash?


UPDATE: All Core offset was the problem. Using Per Core and tuning each core individually seems to have fixed it! Thanks to everyone who contributed!
 
#2 ·
xxbluedragonxx said:
but still this random idle/low load restart happens when using negative PBO offsets over 26-27…
Click to expand...
Don't run that low of a curve? The CPU's that you have tried do not seem to be able to do it. I have my 5600X with boost capped at 4800, -25 all core with my power limits.
 
  • Rep+
Reactions: xxbluedragonxx
Save
Reply Quote Rep+ Rep+
#3 ·
xxbluedragonxx said:
I’m hoping I can get some help on why I’m getting random black screen restarts with my 5600x. This is happening when idling or light loads like browsing.

I have BIOS all defaults except PBO negative offset (all cores) and PBO Limits Disabled.

it’s in an MSI B550 Tomahawk motherboard with latest BIOS. Windows 10 “balanced” power plan.

How can I reliably test non-load crashes? How can I prevent?

Do i just have to wait for random crash and then adjust individual core offset based on WHEA CPID logger output?

I have tried 3 different 5600x CPUs and replaced the motherboard twice but still this random idle/low load restart happens when using negative PBO offsets over 26-27…

I guess is there any reliable way to test for this kind of crash?
Click to expand...
I don't like all cores CO, cores are not the same so why apply CO profile with same value per core?

Not optimal from what I have experienced.

I have done per core CO on several AM4 CPUs and not had an issue, now doing my same method on AM5. See OP here for what I do for per core CO, I'm slowly adding more info as and when, so do recheck OP.
 
#4 ·
gupstergv2.0 said:
don't like all cores CO, cores are not the same so why apply CO profile with same value per core?
Click to expand...
Because the newer silicon sucks, all of mine are from launch. I have 3 Zen 3 parts.
 
  • Rep+
Reactions: xxbluedragonxx
Save
Reply Quote Rep+ Rep+
Discussion starter · #5 ·
gupstergv2.0 said:
I don't like all cores CO, cores are not the same so why apply CO profile with same value per core?

Not optimal from what I have experienced.

I have done per core CO on several AM4 CPUs and not had an issue, now doing my same method on AM5. See OP here for what I do for per core CO, I'm slowly adding more info as and when, so do recheck OP.
Click to expand...
that’s helpful post but it does load testing. I’m getting crashes just idling, seems to be random… pain to test for
 
#6 ·
Give moar power.

I am feeding mine 200/140/180 +150 -25 AC and it is core cycler approved.
 
  • Rep+
Reactions: xxbluedragonxx
Save
Reply Quote Rep+ Rep+
#7 ·
Owterspace said:
Because the newer silicon sucks, all of mine are from launch. I have 3 Zen 3 parts.
Click to expand...
Putting aside 5900X and 5700X, I had 5x 5800X3D with varying production week/year I didn't note that.

xxbluedragonxx said:
that’s helpful post but it does load testing. I’m getting crashes just idling, seems to be random… pain to test for
Click to expand...
No worries, sorry it doesn't aid you. That is strange, not experienced this. Especially considering you've had 3x 5600X and 2x motherboards.

Owterspace said:
Give moar power.

I am feeding mine 200/140/180 +150 -25 AC and it is core cycler approved.
Click to expand...
I don't think increasing PPT/EDC/TDC is going to help in this case.

I like CoreCycler, but it's only one core at a time, and high clock. Which isn't how CPU behaviour is in normal usage. So do have to rely on other tests as well as just using the PC with a CO profile.
 
Discussion starter · #8 · (Edited)
gupstergv2.0 said:
Putting aside 5900X and 5700X, I had 5x 5800X3D with varying production week/year I didn't note that.



No worries, sorry it doesn't aid you. That is strange, not experienced this. Especially considering you've had 3x 5600X and 2x motherboards.



I don't think increasing PPT/EDC/TDC is going to help in this case.

I like CoreCycler, but it's only one core at a time, and high clock. Which isn't how CPU behaviour is in normal usage. So do have to rely on other tests as well as just using the PC with a CO profile.
Click to expand...
the instability is at idle/low load so perhaps the negative offset is too much.

I have also swapped GPU between AMD and nVidia and also a new PSU but still get this so I’m pretty sure it’s something to do with the PBO all offsets being too low

I guess I will just have to try per core offsets and a lot of patience to see if/when it happens again

my goal is “stock” speeds/boost but undervolted as much as possible
 
#9 · (Edited by Moderator)
xxbluedragonxx said:
the instability is at idle/low load so perhaps the negative offset is too much.
Click to expand...
Usually the better cores, ie which boost the most at stock and have lower voltage, using my method of per core CO tend to use lower CO offset.

Look at these CPUs, rest of the HW same, same method to see/set CO.

Text Number Font Screenshot Design

The CO value we set tweaks how the SMU of CPU profiles the core for frequency and determines voltage to give for frequency.

Say I have a core that is reaching ~ 5400MHz on a 9600X, lets say voltage is ~1.3V. I set a -5 CO on just that core. Lets say the CPU does ~5425MHz after adjustment, the SMU may decide to still give ~1.3V, as core clock increased.

Then you readjust CO to say -8 on that core only, you may see ~5450MHz, the SMU may decide to still give it ~1.3V, as core clock has reached FMAX.

Then you readjust CO to say -10 on that core only, you will continue to see ~5450MHz, but now the SMU will start dropping voltage, to say ~1.29V. Now further reduction of CO value will keep yielding voltage drop, whilst sustaining MAX core clock.

xxbluedragonxx said:
I have also swapped GPU between AMD and nVidia and also a new PSU but still get this so I’m pretty sure it’s something to do with the PBO all offsets being too low

I guess I will just have to try per core offsets and a lot of patience to see if/when it happens again

my goal is “stock” speeds/boost but undervolted as much as possible
Click to expand...
For sure to me seems like how your setting global CO is an issue.

It is difficult for me in one post to put all my observations on per core CO tweaking, I am adding to OP in my thread today more information.
 
#11 ·
Actually hate core cycler, as windows does not use the cpu the way that program does.
 
  • Rep+
Reactions: xxbluedragonxx
Save
Reply Quote Rep+ Rep+
#12 ·
Owterspace said:
Actually hate core cycler, as windows does not use the cpu the way that program does.
Click to expand...
I can't say I hate CoreCycler.

I'm quite grateful @sp00n82 created something to use with multiple test loads, is also updating it and sharing for free. Imagine the pain setting test loads to per core and cycling cores. The script does it with ease, tweakable, open source, etc.

We don't rely on one benchmark to say x performance out of a system, as there's more to it then just that a benchmark. I apply the same mindset to stability testing, don't reply on one program/method of testing.
 
#14 ·
Yeah.. lots of ways to test. I’m good.
 
  • Rep+
Reactions: xxbluedragonxx
Save
Reply Quote Rep+ Rep+
Discussion starter · #16 ·
Very helpful thank you!

I have set PBO Limits to "Disabled" from what I understand for MSI this forces limit to the CPU specs rather than using more power.

I have also set Per Core Offset to -20 for each core. I guess now just begins the wait to see if things crash in idle and slowly increase the offset.
 
#17 ·
No worries :) , there is a lot in OP.

On my to do list is to add key points in each section, so rather then reading text, looking at screen shots, etc person just see the point clearly.

Use Ryzen Master to see what PPT/EDC/TDC is, as 5600X is 65W part at stock you should see PPT = 87W, TDC = 60A, EDC = 90A.

This is another thing on my to do list for OP.
 
#19 ·
SuperPi 32M on each core, raise CO until crashes stop. Test heavy with Linpack Xtreme, OCCT is ok to throw in the mix too, I use it mostly for fclk testing, or to get the machine to throw a WHEA.

But really, you should find out the max the CPU can pull from the socket with PBO, set your power limits to those, and tune your curve to that.

But I lost the paper I wrote mine on, so I just use a basic 5900X tune and that is enough for me. 200/140/180 lol.. But I use 260/170/200 for my 5900X. She will do the full 260.
 
  • Rep+
Reactions: xxbluedragonxx
Save
Reply Quote Rep+ Rep+
1 - 20 of 38 Posts
About this Discussion
37 Replies
6 Participants
Last post:
seamen34
Overclock.net
posts
28M
members
562K
Since
2004
The premier forum for overclocking experts and enthusiasts. Discuss hardware optimization, custom builds, benchmarking, cooling solutions, and pushing the boundaries of computing performance. From beginner guides to extreme overclocking, join our technical community to master system tuning.
Full Forum Listing
Explore Our Forums
Intel CPUs Hardware News NVIDIA PC Gaming Water Cooling