F5F Stay Refreshed Power Users Overclocking Undervolting and PBO

Undervolting and PBO

Undervolting and PBO

Pages (2): 1 2 Next
J
john0404
Member
64
09-19-2021, 05:45 PM
#1
It seems like the undervolting process for my CPU isn’t cooperating well, especially when trying to combine PBO with Ryzen Master. When I leave BIOS settings untouched and just run PBO in Ryzen Master, everything works smoothly and boosts performance nicely. However, if I switch to BIOS and enable both RM and PBO, the core speeds drop compared to running them separately. During BIOS updates, I set the 'max cpu boost overdrive' to 150MHz, which keeps the system stable but causes a reset back to zero when PBO is activated in RM. This inconsistency makes me wonder if there’s an issue with my configuration.
J
john0404
09-19-2021, 05:45 PM #1

It seems like the undervolting process for my CPU isn’t cooperating well, especially when trying to combine PBO with Ryzen Master. When I leave BIOS settings untouched and just run PBO in Ryzen Master, everything works smoothly and boosts performance nicely. However, if I switch to BIOS and enable both RM and PBO, the core speeds drop compared to running them separately. During BIOS updates, I set the 'max cpu boost overdrive' to 150MHz, which keeps the system stable but causes a reset back to zero when PBO is activated in RM. This inconsistency makes me wonder if there’s an issue with my configuration.

T
TommyTheLommy
Posting Freak
846
09-27-2021, 12:43 AM
#2
The system might still require using RM because it ensures a more reliable backup process, even if the BIOS setting produces identical outcomes.
T
TommyTheLommy
09-27-2021, 12:43 AM #2

The system might still require using RM because it ensures a more reliable backup process, even if the BIOS setting produces identical outcomes.

A
Agus6w
Member
60
09-27-2021, 04:51 AM
#3
Precision boost is turbo/MCE, and the precision boost override takes control of the PB.
All functions operate correctly as intended.
PBO aims to increase voltage, amps, and watts delivered to the motherboard or CPU, ensuring maximum clock speeds without compromising efficiency.
https://www.gamersnexus.net/guides/3491-...ks-auto-oc
A
Agus6w
09-27-2021, 04:51 AM #3

Precision boost is turbo/MCE, and the precision boost override takes control of the PB.
All functions operate correctly as intended.
PBO aims to increase voltage, amps, and watts delivered to the motherboard or CPU, ensuring maximum clock speeds without compromising efficiency.
https://www.gamersnexus.net/guides/3491-...ks-auto-oc

F
FranceskoKing
Junior Member
11
09-28-2021, 11:46 AM
#4
Use BIOS or Ryzenmaster to adjust PBO settings, never both at once. BIOS applies changes every time you boot. Ryzenmaster is meant for advanced overclockers. Undervolting should be gradual after choosing your preferred PBO values, and offset adjustments should not be fixed. Check both multi-thread and single-thread performance, as well as stability during the process. Results will differ from others. The 'silicon lottery' exists, and some manufacturers may falsely report power usage to boost performance. Not all motherboard BIOSes support this behavior.
F
FranceskoKing
09-28-2021, 11:46 AM #4

Use BIOS or Ryzenmaster to adjust PBO settings, never both at once. BIOS applies changes every time you boot. Ryzenmaster is meant for advanced overclockers. Undervolting should be gradual after choosing your preferred PBO values, and offset adjustments should not be fixed. Check both multi-thread and single-thread performance, as well as stability during the process. Results will differ from others. The 'silicon lottery' exists, and some manufacturers may falsely report power usage to boost performance. Not all motherboard BIOSes support this behavior.

B
babcraft
Member
71
10-03-2021, 03:30 AM
#5
Honestly, after experimenting, even though I can achieve slightly better multicore performance with PBO2 and undervolting, the stability is much higher for me using the latest CTR. Games tend to crash after some time with PBO2/undervolting, but I’m still maintaining faster single-core speeds while experiencing slower multicore speeds (around 50MHz) with CTR—no crashes have occurred. (Plus, cooler temperatures.) I might get more consistent PBO2 results if I spent a lot of time fine-tuning each core individually, but since I’m relatively new to overclocking/undervolting, it’s simpler to rely on CTR.
B
babcraft
10-03-2021, 03:30 AM #5

Honestly, after experimenting, even though I can achieve slightly better multicore performance with PBO2 and undervolting, the stability is much higher for me using the latest CTR. Games tend to crash after some time with PBO2/undervolting, but I’m still maintaining faster single-core speeds while experiencing slower multicore speeds (around 50MHz) with CTR—no crashes have occurred. (Plus, cooler temperatures.) I might get more consistent PBO2 results if I spent a lot of time fine-tuning each core individually, but since I’m relatively new to overclocking/undervolting, it’s simpler to rely on CTR.

J
Jessicaught
Junior Member
10
10-03-2021, 10:36 AM
#6
I've found that with my third generation CPU, it's not accurate to depend solely on core clock readings when evaluating performance under any configuration that keeps clocks in AUTO mode, such as PBO. I believe this issue will persist with the fourth generation as well. The performance changes are extremely rapid—constantly fluctuating up and down up to 100 times per second. Even analyzing average core clocks is ineffective because the long polling intervals I need to use distort the data significantly.

I now depend entirely on performance testing to judge how different settings compare. Running Cinebench, both multi-threaded and single-threaded, is the only reliable method.

For stability checks, I perform a 10-minute test and a 30-minute one. It's actually helpful to observe the performance decline over time, as I know the processor is actively managing heat. To ensure fair comparisons, I use identical tests for any fixed overclock settings, although performance metrics can't be artificially lowered. This results in continuous heating until the system fails—either by crashing or overheating due to insufficient cooling.
J
Jessicaught
10-03-2021, 10:36 AM #6

I've found that with my third generation CPU, it's not accurate to depend solely on core clock readings when evaluating performance under any configuration that keeps clocks in AUTO mode, such as PBO. I believe this issue will persist with the fourth generation as well. The performance changes are extremely rapid—constantly fluctuating up and down up to 100 times per second. Even analyzing average core clocks is ineffective because the long polling intervals I need to use distort the data significantly.

I now depend entirely on performance testing to judge how different settings compare. Running Cinebench, both multi-threaded and single-threaded, is the only reliable method.

For stability checks, I perform a 10-minute test and a 30-minute one. It's actually helpful to observe the performance decline over time, as I know the processor is actively managing heat. To ensure fair comparisons, I use identical tests for any fixed overclock settings, although performance metrics can't be artificially lowered. This results in continuous heating until the system fails—either by crashing or overheating due to insufficient cooling.

_
_FreeZe_YT_
Member
145
10-04-2021, 06:48 PM
#7
General guideline – employ tools like Ryzen Master to fine-tune configurations, remove the program and reinstall the adjustments through the BIOS.
_
_FreeZe_YT_
10-04-2021, 06:48 PM #7

General guideline – employ tools like Ryzen Master to fine-tune configurations, remove the program and reinstall the adjustments through the BIOS.

Z
ZERTOX_YT
Member
62
10-05-2021, 01:24 AM
#8
I don't really like RM much, but it works okay on my PC. I've tried ClockTuner2 (guru3d.com) and it seems to be exactly what I need. It lets me fine-tune VID and clocks for the best stable performance without hurting idle. At my current settings, around 18% loads I get a slight boost, so I should use higher VID per CCX and stronger boosts. Up to 54% loads give a tighter, more stable gain for better performance compared to temperature changes versus voltage changes. The overall result is a solid balance—no need to set the BIOS to a rigid gain, just a small drop in speed when temperatures rise by about 20°C.
Z
ZERTOX_YT
10-05-2021, 01:24 AM #8

I don't really like RM much, but it works okay on my PC. I've tried ClockTuner2 (guru3d.com) and it seems to be exactly what I need. It lets me fine-tune VID and clocks for the best stable performance without hurting idle. At my current settings, around 18% loads I get a slight boost, so I should use higher VID per CCX and stronger boosts. Up to 54% loads give a tighter, more stable gain for better performance compared to temperature changes versus voltage changes. The overall result is a solid balance—no need to set the BIOS to a rigid gain, just a small drop in speed when temperatures rise by about 20°C.

W
wahleno
Member
243
10-05-2021, 01:45 AM
#9
Yes, it suggests using CTR in place of RM or bios tweaking.
W
wahleno
10-05-2021, 01:45 AM #9

Yes, it suggests using CTR in place of RM or bios tweaking.

H
HippoMonk
Member
186
10-17-2021, 01:14 PM
#10
For someone still finding their footing with BIOS tweaks, CTR is perfect. On my 5600X I’m seeing improved all-core overclocking and single-threaded boosts reaching 4.8Ghz across three cores while using lower voltages and keeping temperatures down. I also follow Yuri on Patreon, so I can try the pre-release versions of CTR.
H
HippoMonk
10-17-2021, 01:14 PM #10

For someone still finding their footing with BIOS tweaks, CTR is perfect. On my 5600X I’m seeing improved all-core overclocking and single-threaded boosts reaching 4.8Ghz across three cores while using lower voltages and keeping temperatures down. I also follow Yuri on Patreon, so I can try the pre-release versions of CTR.

Pages (2): 1 2 Next