1usmus Power Plan for AMD Ryzen - New Developments

[TheLostSwede]

Just in case anyone is looking to remove the previous one, or these, see link.
It would seem you have to use the command line option, as the other one is missing in the recent Windows 10 versions.

Delete Power Plan in Windows 10

How to Delete a Power Plan in Windows 10

www.tenforums.com

This does indeed seem to help with using the high performance cores first though, which is nice, as my fastest core is in the secondary CCX and didn't see much use so far. However, it still defaults to the second fastest core most of the time, as it's in the primary CCX. The boosts also seems to last longer, at least from what I can see in HWiNFO, as rather than just flashing up to a certain speed and then dropping straight away, it now stays there for 2-3 seconds before dropping back down.
However, AGESA 1.0.0.4B doesn't give me as high boosts as one of the 1.0.0.3 releases where I was hitting 4,600MHz, now the best I've seen is 4,550MHz.

As a side note, Global C-state Control, CPPC Preferred Cores, and AMD Cool'n'Quiet are not visible in Gigabyte's UEFI, but all of them are enabled by default. I guess they didn't want people messing with those settings, but it's also confusing in a way...
We might get to see them in the future though.

 

[Frick]

AMD didn't give me a sample for CPU review. Yuri is completely separate both geographically and logistically

been wondering about that.

 

[TheLostSwede]

So I've switched to Hwinfo for monitoring my speeds and i'm getting closer to the 4.6 than I thought. Still nothing over 4466 Mhz with my 3900x though. I've tried with PBO both on and off (doesn't make a difference). This seems pretty good but really I'm not sure what to expect tbh.

View attachment 137042

Care to throw up some screenshots of your UEFI settings? Just plug in a FAT32 formatted USB drive and hit F12 to grab a screenshot.

 

[VulkanBros]

This Power Plan tool - it only applies to RyZEN 3000 series CPU´s - correct? Or does it help any on RyZEN 1000 and 2000 series CPU´s as well?

 

[1usmus]

I have a similar spec'd machine and I'm curious what the PBO setting should be set to as well. I too cannot boost passed 4.2Ghz with the power profile, and my temps hover in the mid-40s under 10% load with my AIO Corsair H115 Platinum for cooling.

• Ryzen 3900x - haven't touched OC settings outside of the 1usmus instructions, or Ryzen Master settings.
• x570 Auros Extreme Mobo
• 64GB (3600mhz XMP profile) Corsair RAM
• Windows Build 1903
• F10c BIOS from Gigabyte website
• Corsair H115 Platinum AIO
• 1.09.27.1033 chipset drivers,
• 1usmus 1.1 power settings and bios settings to match. I know for sure I got the following BIOS settings: Global C-state Control, Low Current Idle, and CPPC Preferred Cores to Enabled - I also couldnt find PPC).

Do I understand correctly that your processor also does not receive maximum boost in single-threaded loads?

Guys, everyone who needs help, attach to your message a log made using Hwinfo 6.15 in CB15 (single thread).

P.s. System must be in default + there should not be any background activity during logging.

 

[TylerTr]

Made an account to just say thank you.. my 3700x recorded highest averages for core clocks and i got the best c20 scores with the power plan and the right things enabled in bios. Cores are sleeping when not in use, and average core voltages are super low when im just browsing the web and doing normal things. will test gaming tomorrow, but i imagine i'll see the same results there.

 

[jamexman]

On Ryzen Univeral power plan, this thing constantly keeps the CPU core voltage at around 1.45V. Even in 2129Mhz. This is bullshit.

Even the balanced plan of windows itself nicely puts the CPU into 0.890mv instead of 1.45v. Are you kidding me? If you make a power plan please test it properly and dont cram this up your system without monitoring it's abilitys. Ryzens are proned to suffer from degradation fast and having it in idle on 1.45V / 2129Mhz (CPU-Z, Ryzen 2700X) is just asking for problems.

Windows balanced is proper enough for me.

Did I just read you have a 2700X? Dude, this power plans are for Ryzen 2nd gen (3XXX).

 

[TheLostSwede]

Coming soon™ to a Gigabyte board near you...

Still not getting as good clocks as with AGESA 1.0.0.3 something.

This Power Plan tool - it only applies to RyZEN 3000 series CPU´s - correct? Or does it help any on RyZEN 1000 and 2000 series CPU´s as well?

No, it's only for Ryzen 3000.

 

[DeathtoGnomes]

Just in case anyone is looking to remove the previous one, or these, see link.
It would seem you have to use the command line option, as the other one is missing in the recent Windows 10 versions.

Delete Power Plan in Windows 10

How to Delete a Power Plan in Windows 10

www.tenforums.com

This does indeed seem to help with using the high performance cores first though, which is nice, as my fastest core is in the secondary CCX and didn't see much use so far. However, it still defaults to the second fastest core most of the time, as it's in the primary CCX. The boosts also seems to last longer, at least from what I can see in HWiNFO, as rather than just flashing up to a certain speed and then dropping straight away, it now stays there for 2-3 seconds before dropping back down.
However, AGESA 1.0.0.4B doesn't give me as high boosts as one of the 1.0.0.3 releases where I was hitting 4,600MHz, now the best I've seen is 4,550MHz.

As a side note, Global C-state Control, CPPC Preferred Cores, and AMD Cool'n'Quiet are not visible in Gigabyte's UEFI, but all of them are enabled by default. I guess they didn't want people messing with those settings, but it's also confusing in a way...
We might get to see them in the future though.

thanks

 

[bonezy]

Universal Power Plan

Hey Yuri and thanks for all your hard work! I also follow you on Twitter to keep up with your releases.

I have 2 quick questions.

1. Which power plan works best with 1909?
I am running Win10 Pro 1909, OS build 18363.476 - KB4524570 is also installed.

2. In your previous post, you suggested the following changes to the bios but I can't find all of them in the bios.

You have to set the following in your BIOS, under "CPU Features" or "AMD_CBS":
Global C-state Control = Enabled
Power Supply Idle Control = Low Current Idle
CPPC = Enabled
CPPC Preferred Cores = Enabled
AMD Cool'n'Quiet = Enabled
PPC Adjustment = PState 0

Are they needed for the new (1.1) power plan to work as intended and does PBO have to be disabled?
I am running the latest bios version (1105) for my board (ROG Crosshair VIII Impact)

Again, many thanks!

 

[RH92]

AMD didn't give me a sample for CPU review. Yuri is completely separate both geographically and logistically

Yeah i don't know what AMD PR team plans where for 3950X but something went really wrong on the priority list to say the least .

 

[Danielsweb]

Do I understand correctly that your processor also does not receive maximum boost in single-threaded loads?

Guys, everyone who needs help, attach to your message a log made using Hwinfo 6.15 in CB15 (single thread).

P.s. System must be in default + there should not be any background activity during logging.

Hi there. See attached log of HWInfo 2.15.3390, logged while running CB R15 using your universal power plan. Screen shots of "About my PC" and "Power Plan" settings here. See this reddit post for details on my hardware config and some other benchmarks I've run on Windows 1903 and 1909 using the .

I'm curious to see what you think. I was seeing 4625 Mhz peak on preferred cores at some point in my past (could have been 1003ABB or 1003ABBA) but now the highest peak I ever see is 4550 Mhz on preferred core.

I'm actually happy with this performance, I think it's great, so I'm not really concerned. This is more "for science" than anything else. I'm also willing to do any testing you'd like, if it helps, on fresh installs of Windows and BIOS reset/flashback.

Thank you for all you do for our community.

 

[Artex]

Do I understand correctly that your processor also does not receive maximum boost in single-threaded loads?

Guys, everyone who needs help, attach to your message a log made using Hwinfo 6.15 in CB15 (single thread).

P.s. System must be in default + there should not be any background activity during logging.

Correct - I will attach the log this evening. I've also disabled PBO and verified all other required settings visible to me in the Gigabyte BIOS.

 

[Jism]

Did I just read you have a 2700X? Dude, this power plans are for Ryzen 2nd gen (3XXX).

There's a Universal powerplan included for "Ryzen processors". So i assume thats from Gen 1 to 3. But it's worse for balanced, it keeps the voltage up to a constant 1.45 while even on 2100 mhz idle.

 

[speedgoat]

both power plans boost really high for me, the first one was giving me constantly 4.6 on 2 cores and all cores over 4.5 but i think i probably prefer the most recent one since on this i see the freq down to 2.200 MHz when idling but i ve only been using it for a couple of hours so not sure yet.

they are both on 1.0.0.3, with a C7H, the first one is with 1903 and the 2nd with 1909, ignore the 1.5Vs, they are not happening i am constantly running a small offset.
Really thank you for your contribution Yuri ! i have now the 3800X i thought would have came out of box when i first bought it.

 

[Wickedt]

Do I understand correctly that your processor also does not receive maximum boost in single-threaded loads?

Guys, everyone who needs help, attach to your message a log made using Hwinfo 6.15 in CB15 (single thread).

P.s. System must be in default + there should not be any background activity during logging.

 

[kara.new]

There's a Universal powerplan included for "Ryzen processors". So i assume thats from Gen 1 to 3. But it's worse for balanced, it keeps the voltage up to a constant 1.45 while even on 2100 mhz idle.

If you read the article, it even states it's for the 3000 series.

 

[Wickedt]

CPU Cooling: the boost frequency of the Zen 2 processors is very dependent on temperature. AMD calculated their rated boost clocks at 50°C.

Depending on the processor, maximum boost will go down with temperature:
- 3900/3950 - 75 MHz per 10°C
- 3800/3700 - 50 MHz per 10°C
- 3600/3500 - 35 MHz per 10°C

So when bench marking your computer, unless you can stay at 50 or below, you will never see full boost. I think this is the reason why we see 4525 Mhz so often when running benchmarks, because the temps hover around 50-57 C
Then once it cools down over time, we see 4600+ showing up on some cores, the temps are usually idle temps, and then some background process causes it to boost to its max of 4600+.
Is it possible with water cooling to keep temps at max 50 C?

 

[W1zzard]

Yeah i don't know what AMD PR team plans where for 3950X but something went really wrong on the priority list to say the least .

tell them on social media

 

[Jism]

CPU Cooling: the boost frequency of the Zen 2 processors is very dependent on temperature. AMD calculated their rated boost clocks at 50°C.

Depending on the processor, maximum boost will go down with temperature:
- 3900/3950 - 75 MHz per 10°C
- 3800/3700 - 50 MHz per 10°C
- 3600/3500 - 35 MHz per 10°C

So when bench marking your computer, unless you can stay at 50 or below, you will never see full boost. I think this is the reason why we see 4525 Mhz so often when running benchmarks, because the temps hover around 50-57 C
Then once it cools down over time, we see 4600+ showing up on some cores, the temps are usually idle temps, and then some background process causes it to boost to its max of 4600+.
Is it possible with water cooling to keep temps at max 50 C?

Yes. I have a 2700x and the boost algorithm pretty much works the same way as the 3x00 series. Its all dependent of temps and it has to stay within the 60 degree mark for it to fully utilitize. Now i have a 360mm rad with 6 fans sandwiched in between. On stock it's not enough and it takes a slight undervolt to keep it within 60 degrees mark.

What happens on IBT all core is that for a few minutes it's just stomping all core on 4.2Ghz, and slowly goes down as the water heats up. Eventually tuning in on 4.15Ghz which is really not bad for a 2700x.

So if you want to have best / max boost, keep the temps within. The undervolting on some 3x00 series seem to harm performance instead of bringing.

 

[jamexman]

Hi there. See attached log of HWInfo 2.15.3390, logged while running CB R15 using your universal power plan. Screen shots of "About my PC" and "Power Plan" settings here. See this reddit post for details on my hardware config and some other benchmarks I've run on Windows 1903 and 1909 using the .

I'm curious to see what you think. I was seeing 4625 Mhz peak on preferred cores at some point in my past (could have been 1003ABB or 1003ABBA) but now the highest peak I ever see is 4550 Mhz on preferred core.

I'm actually happy with this performance, I think it's great, so I'm not really concerned. This is more "for science" than anything else. I'm also willing to do any testing you'd like, if it helps, on fresh installs of Windows and BIOS reset/flashback.

Thank you for all you do for our community.

Exactly the same for me and my 3900x. Highest boost was with agesa 1.0.0.3 abba 4625mhz. Agesa 1.0.0.4b lowered them to max 4550mhz.

1usmus new plan at least brought 4615mhz back, but only on one core.

AMD fixed boost clocks with abba and the ruins then again with 4b...

 

[nangu]

Exactly the same for me and my 3900x. Highest boost was with agesa 1.0.0.3 abba 4625mhz. Agesa 1.0.0.4b lowered them to max 4550mhz.

1usmus new plan at least brought 4615mhz back, but only on one core.

AMD fixed boost clocks with abba and the ruins then again with 4b...

I observed the same behaviour, regardless the power plan utilized. On 1.0.0.4, sustained clocks on ST are 100/150 mhz lower than 1.0.0.3, and ~50 Mhz in MT. Benchmarks numbers correlate with this lowered clocks.

Windows 1909 changed things for the worst to my system. The scheduler is now splitting single threaded workloads between Thread0 and Thread1 on my best two cores, instead to sustain the workload on the Thread0 or Thread1 on the core. It's bouncing like mad between Core0 T0 T1, and Core2 T0 T1. With 1.0.0.3 ABBA and Windows 1903 ST Workloads always hitted Core2 T0 and/or Core0 T0, never went to the SMT portion of the core like now.

All of it it's regardless the power plan I use.

I'm really tired to the apparently Microsoft, AMD and Gigabyte incompetence on the software and/or firmware side of things. I'm done with PB and PBO, I dialed a stable CCX Overclock with Ryzen Master and I'm done :-(

 

[Weshya]

I observed the same behaviour, regardless the power plan utilized. On 1.0.0.4, sustained clocks on ST are 100/150 mhz lower than 1.0.0.3, and ~50 Mhz in MT. Benchmarks numbers correlate with this lowered clocks.

Windows 1909 changed things for the worst to my system. The scheduler is now splitting single threaded workloads between Thread0 and Thread1 on my best two cores, instead to sustain the workload on the Thread0 or Thread1 on the core. It's bouncing like mad between Core0 T0 T1, and Core2 T0 T1. With 1.0.0.3 ABBA and Windows 1903 ST Workloads always hitted Core2 T0 and/or Core0 T0, never went to the SMT portion of the core like now.

All of it it's regardless the power plan I use.

I'm really tired to the apparently Microsoft, AMD and Gigabyte incompetence on the software and/or firmware side of things. I'm done with PB and PBO, I dialed a stable CCX Overclock with Ryzen Master and I'm done :-(

Owner of a 3600 here. I also observed the same behavior after updating to Agesa 1.0.0.4b, lower sustain clocks overall and the inability of Window's scheduler to properly function. For me 1usmus power plans (both the old and the new universal after updating to Win'1909) & correct BIOS settings have permitted my CPU to regain its pre Agesa 1.0.0.4b boost clocks in a more sustained manner and more consistently on the fastest cores.

Bottom line, It's been 4 months since the Zen 2 processors are out, and it's becoming really tiring to have to check every day for the best tweak/setting/driver/bios in order to finally feel we're getting the most out of our processor

 

[BoMbY]

Still not using the best core on 1903 with that patch and newest plan.

 

[zlobby]

@1usmus - Yuri, what about the mobile variants of Ryzen?

I for example have the 2700U running Win 10 1909 w/ the AMD chipset drivers, which bring the 'Ryzen Balanced' power plan.

Needless to say, the battery life is far from stellar even under low load. With all fairness, laptop doesn't feel sluggish under any task.

Can something be done for the battery life of the mobile Ryzens while maintaining responsiveness?