Explaining the AMD Ryzen "Power Reporting Deviation" -metric in HWiNFO

junneh

New Member
You know PRD alone does not tell anything. You must take account of what PPT reports during that full load 70% PRD.
Depends on the CPU you have and it’s stock PPT.

Anyway, I have Aorus Pro X570 (F32 Bios) with R5 3600 that has a stock PPT of 88W.
During full load my PRD is around 89~92% and PPT report around 87W.
So the true total power consumption of CPU is:
87 / 0.89~0.92 = 94.5~97.7W

We have the same cpu. In prime95 it maxes the ppt and still on ~70% PRD. Getting upto 83c on a fuma2 in a mesh case with 140s which is plain ridiculous. Good to hear the x570 GB does better with this. Will probably buy the pro.
 

Zach

Well-Known Member
We have the same cpu. In prime95 it maxes the ppt and still on ~70% PRD. Getting upto 83c on a fuma2 in a mesh case with 140s which is plain ridiculous. Good to hear the x570 GB does better with this. Will probably buy the pro.
The numbers I mentioned was with CB-R20 which is a more normal (real life) benchmark.

This is P95 (ambient temp 23~24C) with PBO enabled.
As you will see I'm using 280mm AIO water cooler
Real PPT is 102 / 0.915 = 111.5W and CPU temp 70+C.

HWiNFO_P95_003.png

I guess that 88 / 0.7 = 125.7W would increase temp beyond 80C. Its normal for 125+W, depending ambient and case temperature.
Keep in mind that this only counts at full load. I wouldn't be too much concerned even with your numbers unless I'm running full load CPU tests every day for hobby.
My CPU always staying at 30~60C region during my kind of usage (gaming is the most higher load) and 60C is briefly reached.
ZEN2 (3000series) has max operating temp (throttle point) at 95C.
 

junneh

New Member
Im not too worried about it but the board is sold now anyhow ;). Tbf it only really showed up in Prime. R20 was around 65-70c with different coolers but also with around 70% PRD (cant remember R20 PPT usage, it wasnt maxed tho). Gaming temps never exceeded ~68c

I was doubting mostly between an x570 tomahawk and the x570 aorus pro, but it seems much better on the x570 end with Gigabyte regarding PRD from your postings, so I will go for that one since I value alot of usbs and the audio amp it has over the tomahawk :). Thanks for the info Zach :)
 

Zach

Well-Known Member
Any time…
When a got the board (Aug 2019), X570 was the only option for 500series. I picked Aorus Pro as the most balanced board for capabilities and features/ports. Also has a nice VRM heatsink and can easily support a 16core CPU with OC.
Early BIOSs was real pain (first 6-8 months from release) but now it’s very stable. I stayed at F32 ver. as the best for 3000 CPUs. Next year I will switch (most likely) to a 5900X and keep the system a few years more.
 

pma

New Member
Ryzen CPUs for AM4 platform rely on external, motherboard sourced telemetry to determine their power consumption. The voltage, current and power telemetry is provided to the processor by the motherboard VRM controller through the AMD SVI2 interface. This information is consumed by the processors power management co-processor, that is responsible for adjusting the operating parameters of the CPU and ensuring, that neither the CPU SKU, platform or infrastructure specific limits are being violated.

The weakness of this method is, that the telemetry essentially uses an undefined scale for the current (and hence power) measurements. This means that the motherboard VRM controller will send an integer between 0 - 255 to the CPU, and based the reference value known by the co-processor firmwares, this integer is converted to a figure, that represents a physical current drawn by the CPU. Based on the accurately known current flow and the voltage, it is possible to calculate to CPU power draw in Watts (V * I).

The reference value mentioned earlier is generally different for each of the motherboard make and model, unless there are boards which have an identical power circuitry. Because of that, it is on the motherboard manufacturers responsibility to find the correct value for their motherboard design through the means of calibration, and then to declare it properly in AGESA, during the bios compile time. In case the motherboard design specific, correct value differs greatly from the declared value, there will be a bias in the power consumption seen by the CPU. In case the declared value is greater than the actual value, the power consumption seen by the CPU is greater than it actually is. Likewise, if the declared value would be an understatement... the CPU would think it consumes less power than it actually does.

Since at least two of the largest motherboard manufacturers, still insist on using this exploit to gain an advantage over their competitors despite being constantly asked and told not to, we thought it would be only fair to allow the consumers to see if their boards are doing something they're not supposed to do. The issue with using this exploit is, that it messes up the power management of the CPU and potentially also decreases its lifespan because it is running the CPU outside the spec, in some cases by a vast margin. Also, it can cause issues when this exploit goes undetected by a hardware reviewer, since both the performance and the sofware based power consumption figures will be affected by it.

For example, if we take a Ryzen 7 3700X CPU that has 65W TDP and 88W default power limit (PPT), and use it on a board which has declared only 60% of its actual telemetry reference current, we'll end up with effective power limit of ~ 147W (88 / 0.6) despite running at stock settings (i.e. without enabling manual overclocking or AMD PBO). While the 3700X SKU used in this example typically cannot even reach this kind of a power draw before running into the other limiters and limitations, the fact remains that the CPU is running far outside the spec without the user even acknowledging it. This exploit can also cause additional cost and work to the consumer, who starts wondering about the abnormally high CPU temperatures and starts troubleshooting the issue initially by remounting the cooling and usually, eventually by purchasing a better CPU cooler(s).

HWiNFO will display "Power Reporting Deviation" metric under the CPUs enhanced sensors. The displayed figure is a percentage, with 100.0% being the completely unbiased baseline. When the motherboard manufacturer has both properly calibrated and declared the reference value, the reported figure should be pretty close to 100% under a stable, near-full-load scenario. A ballpark for a threshold, where the readings become suspicious is around ±5%. So, if you see an average value that is significantly lower than ~ 95% there is most likely intentional biasing going on. Obviously, the figure can be greater than 100%, but for the obvious reasons it rarely is ;)

As stated before, this metric is only valid during a relatively stable near-full-load condition. That is due to the typical measurement accuracy of the VRM controller telemetry, and also due to the highly advanced and fast power management on Ryzen CPUs, that not only result in extremely low idle, but also in extremely rapidly changing power consumption. A suggested workload to get a stable and reproducable deviation metric is Cinebench R20 NT, with the HWiNFO sample rate set to less or equal to 1000ms.

As of now, outside of certain MSI motherboards, the biasing isn't end-user controllable. In case there is clear evidence of biasing taking place on certain motherboards or their bios versions, please contact the manufacturer and ask them to remove the telemetry biasing from the bios. The biasing can be implemented in different ways, it can be tied to a specific setting(s) (known as an "auto-rule") in the bios or be fixed in a certain bios version or in all available bios versions.

Here is an practical example recorded on MSI X570 Godlike motherboard, using the most recent 1.93 beta-bios version.
For this bios version MSI has declared 280A reference current, when the correct value that produces near 100% result (i.e. no deviation) and also a matching power draw compared to other boards (same CPU and workload) is 300A. This means that the board allows 7.14% (300/280) higher power draw for the CPU than AMD specifications state. Compared to the worst violators (up to 50%) this is minor infraction, so MSI deserves a benefit of a doubt whenever this is intentional or a honest error.

With the proper 300A setting, the average HWiNFO "CPU Power Reporting Deviation" during Cinebench R20 NT is 99.2%.
With this setting, the average CPU core frequency is 4027.4MHz, power consumption seen by the CPU 140.964W (of 142W limit) and peak CPU temperature of 73°C.

qjankbG.png


With 225A setting (75% of the actual), the average HWiNFO "Power Reporting Deviation" during Cinebench R20 NT is 75.3%.
With this setting, the average CPU core frequency is 4103.5MHz, power consumption seen by the CPU 125.241W (of 142W limit) and peak CPU temperature of 80°C.

qRvEVCD.png


With 150A setting (50% of the actual), the average HWiNFO "Power Reporting Deviation" during Cinebench R20 NT is 50.2%. With this setting, the average CPU core frequency is 4106.6MHz, power consumption seen by the CPU 91.553W (of 142W limit) and peak CPU temperature of 79°C. This setting is already limited by maximum voltage allowed by the silicon fitness (FIT), so there were pretty much no addition performance gains, or ill-effects for that matter to be had.

t0Yifq1.png


I'd like to stress that despite this exploit is essentially made possible by something AMD has included in the specification, the use of this exploit is not something AMD condones with, let alone promotes.
Instead they have rather actively put pressure on the motherboard manufacturers, who have been caught using this exploit.

In short: Some motherboard manufacturers intentionally declare an incorrect (too small) motherboard specific reference value in AGESA. Since AM4 Ryzen CPUs rely on telemetry sourced from the motherboard VRM to determine their power consumption, declaring an incorrect reference value will affect the power consumption seen by the CPU. For instance, if the motherboard manufacturer would declare 50% of the correct value, the CPU would think it consumes half the power than it actually does. In this case, the CPU would allow itself to consume twice the power of its set power limits, even when at stock. It allows the CPU to clock higher due to the effectively lifted power limits however, it also makes the CPU to run hotter and potentially negatively affects its life-span, same ways as overclocking does. The difference compared to overclocking or using AMD PBO, is that this is done completely clandestine and that in the past, there has been no way for most of the end-users to detect it, or react to it.
Thanks for the write up, was wondering what that stat was all about.
If I understood it correctly, this MB is inflating the power by 17% (with a PRD of 117% at full load in Cinebench R20).
1626468470014.png
Not optimal though I'd rather stay on this side of the fence and tweak it later if possible than going about burning the life out of this CPU due to power underreporting.
 

Zach

Well-Known Member
Thanks for the write up, was wondering what that stat was all about.
If I understood it correctly, this MB is inflating the power by 17% (with a PRD of 117% at full load in Cinebench R20).
View attachment 6691
Not optimal though I'd rather stay on this side of the fence and tweak it later if possible than going about burning the life out of this CPU due to power underreporting.
Yes your board is overreporting CPU power and causing it to underperform.
Stock PPT of 5800X is 142W and the real power under CB-R20 is ~121W.
142 / 1.17 = 121W

1. Even if your board didn't do that, you still wouldn't see any benefits because your CPU already reaches the max operating temp (90°C) as it is.
2. This PRD "issue" cannot threat the lifespan of the CPU when a board does the opposite of yours and underreports PPT. The CPU is still under the regulation of its internal FIT controller (silicon FITness manager) when power management (Speed/Voltage) is Auto, regardless of PBO(Enabled/Disabled). All the underreport does is pushing the stock power limit but not the silicon stress/operating limit. Its there in Stilt's article if you missed it.
3. The unique thing about 5800X that distinguishes it from all other ZEN2/3 SKUs is that it is a single CCD CPU with all cores active and 142W power. The result is that 5800X has the higher heat density (W per die surface) among all CPUs from 3000/5000 series. That puts it in the hardest place for cooling. The need for high capacity cooling device and also high heat transfer thermal material (TIM/paste) is almost mandatory for starters. Room/Case ambient temp should also be optimal but this is secondary in significance. First is heat transfer and dissipation ability of the entire CPU cooling mechanism.
 
Last edited:

bloody74

New Member
Laut Gigabyte Sind die Gemeldeten Wertee Bei Der Power Reporting Nicht "aussagekräftig". der support sagt, sie wissen nicht was dieses 3thparty tool ausliest (hwinfo) und empfehlen, SIV zu benutzen.
Mein Power Report sieht so aus
neueste bios version F13, alle treiber aktuell
verbaut ist ein ryzen 3600
laut ereignisanzeige habe ich aber keine kritischen fehler, sondern nur die normalen dinge. und auch sonst läuft mein rechner sehr gut
vorher hatte ich ein ASUS-Board (tuf 550b gaming plus) und dort regelmäßig id41 Power kernel, as kritische fehler (teils mit unangekündigtem neustart oder auch einfach im hintergrund, egal ob last oder Idle)
 

Attachments

  • board.PNG
    board.PNG
    137.9 KB · Views: 10
Last edited:

Dalai

Well-Known Member
@bloody74:
You're in luck I'm German myself and can understand your post. However, I recommend you write in English next time since this is (primarily) an English language forum. As I'm just a user like yourself, this is no moderation or anything from my side. Just keep in mind that the probability for answers or a disussion is a lot higher when everybody is communicating in the same language.

To your problem: From the screenshot it's clear that your CPU is NOT at full load (CPU Package Power was under 40W at max). PRD is completely irrelevant unless the CPU is under full load like in CineBench (multi-core test) or similar. The first post of this topic mentions this, BTW.

Regards
Dalai
 

bloody74

New Member
@bloody74:
Du hast Glück, ich bin selbst Deutscher und kann deinen Beitrag verstehen. Ich empfehle Ihnen jedoch, das nächste Mal auf Englisch zu schreiben, da dies (in erster Linie) ein englischsprachiges Forum ist. Da ich nur ein User wie du bin, ist dies keine Moderation oder sonstiges von meiner Seite. Bedenken Sie nur, dass die Wahrscheinlichkeit für Antworten oder eine Diskussion viel höher ist, wenn alle in der gleichen Sprache kommunizieren.

Zu deinem Problem: Aus dem Screenshot ist klar, dass deine CPU NICHT voll ausgelastet ist (CPU Package Power war unter 40W max). PRD ist völlig irrelevant, es sei denn, die CPU wird unter Volllast wie im CineBench (Multi-Core-Test) o. Der erste Beitrag dieses Themas erwähnt dies übrigens.

Grüße
Dalai
es ist egal wie die CPU läuft, Volllast oder Idle. die werte von hwinfo sind seltsam
jetzt ein Bildschirm ca. 10 Minuten cinebench23 Multicore
this fucking google translater....
it doesent matter if cpu runing full load. now a screen aboput a 10 minute duration from cinebench23 multicore, with 3975mhz on 12 cores(6/6) at 1,250 cpu voltage about 60 watt PPT
second screenshot is from stresstest aida64 with cpu, fpu, memory
a strange thing is, that the dram voltage is at 1.368v and in the bios i have set 1.35v (XMP profile), the same on soc voltage, gigabyte set automaticly to high with 1.080-1.092v. setting in the bios is standart /auto
 

Attachments

  • cinebench23-10minuten.PNG
    cinebench23-10minuten.PNG
    139.5 KB · Views: 12
  • stresstest aida64.PNG
    stresstest aida64.PNG
    745.9 KB · Views: 10
Last edited:

Dalai

Well-Known Member
OK, let me rephrase. What PRD value do you get while the CPU is under full load (i.e. CineBench is running/looping)? As soon as the load is gone, the PRD value gets irrelevant. In other words, it's no use to let CB run for several minutes and reading the PRD value afterwards. It's only relevant while the CPU is under load.

Regarding the voltages: The BIOS sets some voltage. That doesn't mean that this exact voltage will be applied or can be measured exactly afterwards. Keep in mind that ICs have some variation and sensors have some variance in their precision. The 1.368 volts might be correct, but maybe not. But it's close enough to the 1.35 volts in my opinion. If you think it's too much, maybe you can set some negative offset in the BIOS. Watch out for system instability when doing so, though.

Regards
Dalai
 

bloody74

New Member
OK, let me rephrase. What PRD value do you get while the CPU is under full load (i.e. CineBench is running/looping)? As soon as the load is gone, the PRD value gets irrelevant. In other words, it's no use to let CB run for several minutes and reading the PRD value afterwards. It's only relevant while the CPU is under load.

Regarding the voltages: The BIOS sets some voltage. That doesn't mean that this exact voltage will be applied or can be measured exactly afterwards. Keep in mind that ICs have some variation and sensors have some variance in their precision. The 1.368 volts might be correct, but maybe not. But it's close enough to the 1.35 volts in my opinion. If you think it's too much, maybe you can set some negative offset in the BIOS. Watch out for system instability when doing so, though.

Regards
Dalai
when cinebenchr23 is running, the power report deviation is about 58%-62%(in the middle is a drop at 46% for ms) at the current.
ppt=68,1%; TDC=48,5% and EDC=100%
the voltages are strange, because the same cpu and ram takes on a asus tuf gaming b550 the correct voltages (default settings, XMP profile and soc voltage) from bios.
when i set manuell 1.35v for the ram, in the gigabyte bios, the board automatically set it higher.
no change when i set auto or manuell the 1.35v for xmp.
the board set 1.368v.
i don't think that the voltage is dangerous, but is strange, that a manuell set, in the bios, doesn't take a effect.
at the default settings the ram have 1.212v.
i think that gigabytes f13 bios is set the voltages a little bit higher, as they need
hope my English is good enough to understand^^.
gigabyte means about the power report deviation, that not a problem by the board, more a problem about the third party tool(HWinfo)
i stand in contact with the support about the ram/soc voltages and the PRD values....
 
Last edited:

Dalai

Well-Known Member
Around 60% PRD is weird (and too low), I agree. Do you run your system and CPU at stock settings? XMP for memory is OK, but every other voltage and frequency should be at stock/default/auto. If it is at stock, it's possible that this board indeed underreports what it supplies to the CPU. This won't lead to any damage (maybe in the really long term), but it's still out of the ordinary. Perhaps compare the CineBench scores with some results online and see if there's a lot of deviation.

Regards
Dalai
 

Zach

Well-Known Member
I have a Gigabyte board (AorusPro X570) and R5 3600.
Most Gigabyte boards (if not all) supply a little more on DRAM voltage from BIOS setting. It’s normal and this cannot affect a matter like PowerReportingDeviation. And it’s not BIOS version specific.

60W PPT and around 60% PRD gives a true power of around 100W. It’s not bad for a CPU like the R5 3600. CPU temp must be very reasonable for it to let push it self around the 100W mark.
@bloody74 have tried with BIOS F12 to see if PRD is more close to 100%?
 

PJVol

Member
With agesa 1.2.0.3b all abnormal deviations was fixed on most boards. At least, with the latest asrock bios I don't need to manually adjust "vdd cpu full scale current" anymore, since auto work fine now.
 

Gastoffu

New Member
Hello good day. I am from Argentina. I have doubts on this subject.

If the percentage is less than 100%, does it mean that it does not really consume what it shows? Or what is already showing is the addition to what is supposedly lacking to consume?

Another doubt:

I read that I should not take this issue into account if I overclock.

I use "manual" overclock of 4.125Mhz and 1.175v through Ryzen Master.

In Cinebench the maximum temperature I usually reach is 72 with a PPT of 57%.

The big question I have is that the "power reporting deviation" usually varies a lot, it is not fixed. It can be less than 100% or more. What does that mean?
 

Zach

Well-Known Member
Hello good day. I am from Argentina. I have doubts on this subject.

If the percentage is less than 100%, does it mean that it does not really consume what it shows? Or what is already showing is the addition to what is supposedly lacking to consume?

Another doubt:

I read that I should not take this issue into account if I overclock.

I use "manual" overclock of 4.125Mhz and 1.175v through Ryzen Master.

In Cinebench the maximum temperature I usually reach is 72 with a PPT of 57%.

The big question I have is that the "power reporting deviation" usually varies a lot, it is not fixed. It can be less than 100% or more. What does that mean?
Did you read the very first post (1) of this thread? Its all there...
And some general info you can get from sensors window when you hover the mouse pointer over the PRD sensor (if you have "Show ToolTips" enabled from sensor settings).
 

ixoza

New Member
I guess this is common thing rn, I've read the whole topic. I guess the best solution there will be just buy the new motherboard from another manufacturer?
 

Attachments

  • b550m aorus pro-p.jpg
    b550m aorus pro-p.jpg
    1.6 MB · Views: 15

AeroZ

Active Member
I guess this is common thing rn, I've read the whole topic. I guess the best solution there will be just buy the new motherboard from another manufacturer?
Why buy anything at all? Despite the PRD value everything runs in spec and works fine.
 
Top