HWiNFO "Power Reporting Deviation" metric & AM4 motherboards

[Bob Gnarley]

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.

Article in full: https://www.hwinfo.com/forum/thread...er-reporting-deviation-metric-in-hwinfo.6456/

Props to "The Stilt".

 

[BrunoBæ]

Doesnt sound good.
Im suddenly glad I still run an neanderthal Xeon x5670 on a 1366 MB

 

[Bob Gnarley]

It's been thought for a while that certain mb manufacturers, I think mainly Asus & Gigabyte, have been cheating with the power limits for Ryzen 3000. HwInfo recently added a metric that measures and proves this to be the case for some.

As well as the implications above it also nullifies a lot of Ryzen 3000 benchmarks as many of those tests will have been done with, effectively, with an overclocked cpu.

 

[robt100]

Asus does it for both manafacturers as far as I've seen/read. There was an article about it the other day on how they do similar with intel, rmoving the boost clock time limits etc. Probably in an effort for people to recommend the board as running CPUs better without the consumers knowledge of as to why.

 

[Case_]

 

[robt100]

105% reading here on all core load....so MSI is under-shooting rather than fake increases

 

[Bob Gnarley]

Interesting vid.

Back at work but will give CB20 a go at the weekend and post here.

I've read of I think 4 dead R3000 cpu's, all manual o/c's. It would explain why manual o/cing has been risky, if people have been unknowingly setting voltage plus deviation.

I very much doubt any manufacturer would push stock settings that kill cpu's so it should be a non issue for most. Bad for the enthusiast that likes to tinker. You really need a level base line to begin with otherwise you might as well pull random numbers out of the air.

 

[Case_]

The power deviation thing should not be relevant to manual OC. The power limits are ignored during manual OC.

 

[Bob Gnarley]

Power limits aren't ignored, though they can be manually set under PPT. Are you certain that the board doesn't interfere with that? Can't be bothered arguing mate.

Edit: just started an ignore list

 

[Terry Rock]

Not sure why anybody would want to manually overclock a Ryzen CPU anyway.
The darn things does such a good job on their own out-of-the-box.
Heck!... I don't even use precision boost or XMP.
I've just set the ram timings to 3600 MHz, tweaked the timing and let the thing do it's work.

 

[Case_]

Why manually OC - lower temperatures, slightly better multithreaded performance.

 

[Bob Gnarley]

Same here regarding ram and pbo Terry. I found that with pbo off, mine runs cooler and scores higher. I guess more sustained boosting. Didn't see the point in hours of tinkering and benching for maybe 100-200mhz when memory tuning potentially offers better results.

Looking at overclock.net's forum, generally, gains are either small or the cpu is being pushed hard. There are some that have found reasonable gains at low volts, but I consider those the lottery winners. Best results seem to be with 3900/3950 which stands to reason as they are said to have the best silicon in the stack.

 

[Durge Driven]

Im suddenly glad I still run an neanderthal Xeon x5670 on a 1366 MB

Glad I under clock these days had both my 3200G and 3900X for everyday you barely need 2GHz
run so cool too Kraken in silent mode while gaming 60C
Gaming limited to 4.2GHz @ 1.25v ( still miffed never got 1.2v ) <<<

 

[Case_]

Yeah, my 3600 also runs "so cool" that it runs at around 60C during gaming. Thing is, it can run below 50C with manual OC under same conditions and with slightly higher clock, all core.

The voltage PBO puts into the CPU for no real reason under normal loads is a lot higher than the CPU would need. Like 0.1-0.2 V higher.

 

[BrunoBæ]

Glad I under clock these days had both my 3200G and 3900X for everyday you barely need 2GHz
run so cool too Kraken in silent mode while gaming 60C
Gaming limited to 4.2GHz @ 1.25v ( still miffed never got 1.2v ) <<<

Hehe good for you.
But would you like to hear my temps when Im able to both oclock + undervolt my slightly ancient Xeon?
Oclock + undervolt at the same time

 

[Durge Driven]

Hehe good for you.
But would you like to hear my temps when Im able to both oclock + undervolt my slightly ancient Xeon?
Oclock + undervolt at the same time

I know some may think I saying no need to O/C , never said that and respect guys take the time to find the peak for others to follow

I have tracking for my rig so hopefully next week will tell the story
Up to now I have to use Ai control to test things
New 4K TV still blows me away so smooth and looks better then DSR

With 24 Community F3 cars on Nords never dropped below 90 and hit 120-130's
Mind you the F3 are among least hungry cars

This is settings and screen, high shadows and rest Sorry rough resamples
Running windowed mode working great as ever

*** New use for appcompatview too, drag in all apps and game launchers etc etc and set DPi compat to windows enhanced all saved every fresh OS including mod manager and mas2 In 4K mas2 bar was so small I could not see it

 

[Terry Rock]

If anybody else owns an MSI B450 Gaming Pro Carbon AC, there is a new beta BIOS which introduces AGESA 1.0.0.6.
I just installed it and it seems to be working fine.

 

[Bob Gnarley]

New Agesa has been a while coming. Happy days!

That 3900x is a beast, not jealous...honest!

Right, haven't touched the wheel all week. Must get some laps!

 

[robt100]

Not ANOTHER AGESA update ... Ive just updated the BIOS and had to start all over again with settings

EDIT:
Beta BIOS.....Gen 3 support perhaps? It was said that would be BETA BIOS only..

 

[BrunoBæ]

New 4K TV still blows me away so smooth and looks better then DSR

I have a HD(1080p) plasma 55" and have often wondered if I should check it out as a sim monitor.
But have allways hesitated because of the higher input lag in TVs compared to monitors.
Maybe I should try it out

CatsAreTheWorstDogs: Your pics looking extremely good