-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
Power Reporting Deviation?
- Thread starter freeagent
- Start date
- Joined
- Jun 21, 2002
I assume you're running a Ryzen? https://www.hwinfo.com/forum/thread...er-reporting-deviation-metric-in-hwinfo.6456/
The short answer, your motherboard is lying to the CPU about how much power it is using so that it will boost higher. Mind posting your specs? 67.9% is pretty dang low. I believe 90% is considered to be good, 80% around average. Also given that deviation, what is your PPT limit look like? It could be a case where the CPU just doesn't boost that high for other reasons. It's kind of like PBO but instead of raising the limits, the board is automatically tricking the CPU into raising them. There are still other features / safeguards / limits in place in the CPU.
Edit, important note
The short answer, your motherboard is lying to the CPU about how much power it is using so that it will boost higher. Mind posting your specs? 67.9% is pretty dang low. I believe 90% is considered to be good, 80% around average. Also given that deviation, what is your PPT limit look like? It could be a case where the CPU just doesn't boost that high for other reasons. It's kind of like PBO but instead of raising the limits, the board is automatically tricking the CPU into raising them. There are still other features / safeguards / limits in place in the CPU.
Edit, important note
Last edited:
- Thread Starter
- #3
Hey thanks for the reply!
I have a 3600XT and a Strix B550-F, and running 16gb Trident Z. That makes sense, the CPU is pretty boosty on its own, PBO doesn't really do anything I think. The CPU will boost to 4600 without any input from me and everything factory fresh in the bios. Sorry what is a PPT limit? I haven't snooped around the CPU section in the bios too much as I am stuck with a stock cooler atm.
- Joined
- Jun 21, 2002
I believe this link explains it, if not I'll find another one for you (I can't read it right now). https://www.gamersnexus.net/guides/3491-explaining-precision-boost-overdrive-benchmarks-auto-oc.
There are three separate parameters to the power delivery which are monitored and impact boosting behavior, primarily in multi-threaded workloads (as a single thread workload doesn't saturate the power draw of the CPU). PPT, TDC and EDC are the limits and the second two are shown in your screenshot. PBO just means increasing or effectively removing those limits (so instead of boosting based on three parameters: temperature, clock speed, power draw, the CPU instead boosts only based on clock speed and temperature). However this is not absolute as there is also the silicon fitness monitor which restricts voltage regardless of PBO (although a voltage offset can overcome this, but it is not a great idea to do so for long periods).
There are three separate parameters to the power delivery which are monitored and impact boosting behavior, primarily in multi-threaded workloads (as a single thread workload doesn't saturate the power draw of the CPU). PPT, TDC and EDC are the limits and the second two are shown in your screenshot. PBO just means increasing or effectively removing those limits (so instead of boosting based on three parameters: temperature, clock speed, power draw, the CPU instead boosts only based on clock speed and temperature). However this is not absolute as there is also the silicon fitness monitor which restricts voltage regardless of PBO (although a voltage offset can overcome this, but it is not a great idea to do so for long periods).
- Thread Starter
- #5
Thanks for that man. I'm a little chuffed that my CPU can overclock itself better then I can. Ill take a day off from it and come back. That pic was with core enhancement and PBO on, those values changed quite a bit depending on the load.
I completely understand. I ran this month's SuperPi 32m bench on the Overclocktagon at stock, I even forgot to shut down folding@home. Tried to overclock manually w/o F@H and lost 2 minutes on the bench! Need to spend a lot more time in the BIOS figuring out how to boost single & multi thread performance. I can see various saved BIOS profiles to accomplish that.
- Thread Starter
- #7
Yeah man, I think one or two of my cores is a dud. I might have even hurt the CPU a little, not sure yet.. but seems stable as a table @ stock settings, so maybe it will take awhile to show itself if I did do something. First time I tried to adjust vcore I input 1.35v but when I got to windows it showed 1.6v lol. I only say lol because @ stock without PBO and all that the board will shove 1.5v+ to the CPU, but with PBO only about 1.47ish. I can run it all core at 4500 at 1.3v, but will crash running R23. I tried adjusting CCX core voltage as well as the other one. I did also saw a brief spike of 103c once early on in my endeavor.. whoops.
Edit:
I am running 4x8GB but I removed the black and whites and just ran the royals for my cpu trials.
Edit:
I am running 4x8GB but I removed the black and whites and just ran the royals for my cpu trials.
- Joined
- Apr 19, 2012
Freeagent, you can leave the entire system stock and go into Ryzen master. Hit the gaming tab. In PBO section, simply raise the EDC, PPT and TDC limits and the CPU will clock higher based on the temps using the SenseMi algorithm.
It works off the hightemp alert the Ryzen chips are set at 70c. So on a stock system, at this temp, the Cpu fan should be 100% duty cycle. The cpu will not increase boost clocks at this temp. You can manually only adjust this temp to 75c and that includes all motherboards I'm aware of. On your particular asus board, when you F6 to set the fans, you will notice that probably. If you can keep the cpu under 60c or 65c sustained you should see better boosting.
My 2700X responded very well on my Geothermal loop. All the cores stayed Pinned at 4300mhz all the time on all factory defaults. Just a little something for you to ponder on. But anything beyond that, I did kinda try with sub zero cooling to see if the Cpu would try and boost more, but did not. I couldn't get the chip to budge past 4.5ghz even at 1.58v at -30c with only 1 or 2 cores running under a TEC just for a cpu-z validation. The processors are known to wall, and somewhere around -45c (I think it is) they will jump the wall and go faster.
It works off the hightemp alert the Ryzen chips are set at 70c. So on a stock system, at this temp, the Cpu fan should be 100% duty cycle. The cpu will not increase boost clocks at this temp. You can manually only adjust this temp to 75c and that includes all motherboards I'm aware of. On your particular asus board, when you F6 to set the fans, you will notice that probably. If you can keep the cpu under 60c or 65c sustained you should see better boosting.
My 2700X responded very well on my Geothermal loop. All the cores stayed Pinned at 4300mhz all the time on all factory defaults. Just a little something for you to ponder on. But anything beyond that, I did kinda try with sub zero cooling to see if the Cpu would try and boost more, but did not. I couldn't get the chip to budge past 4.5ghz even at 1.58v at -30c with only 1 or 2 cores running under a TEC just for a cpu-z validation. The processors are known to wall, and somewhere around -45c (I think it is) they will jump the wall and go faster.
- Thread Starter
- #9
Thank you kindly sir, Im going to try that right now. Well, as soon as I get the kids to bed. I haven't really played with software at all, only had the master installed once when I first got it.
Awesome man, really appreciate it
Awesome man, really appreciate it
- Thread Starter
- #10
Hey so I tried what you said, and it did as you said except.. it dropped clocks on the first 3 cores and raised the clocks on the last 3 cores. So I went back in the bios to play a little. I'm still a noob right now and just getting a feel for things. Its a bit to take in. So I am back to all core clocking for now. I have 4400 stable at 1.26v. Its really not that bad. I am at 4500 now, but need 1.36v to finish FP-32 ray trace tests in aida64. I chose that to start with before I move on to other test because for whatever reason that test will shut my rig down. It didn't this time lol. Is 1.36v safe for 24/7? I read some stuff about degradation, and don't want to be that guy..
- Joined
- Jun 21, 2002
I think the answer to degradation is it depends, or probably safe. Some Zen2 CPUs have degraded from less. Personally I would just go by PBO. I skimmed this post though, seemed to have some good info
. Some are suggesting for PBO the limits to be set to PPT 300W TDC 230W and EDC 230W instead of max. Why I don't know but if you feel like enduring this video there is an explanation somewhere,
.
