Help with PRD readings

[fatehasfans]

Hi there,

I'm a bit of a novice with computers but I've noticed my fans are ramping up and down so I downloaded HWiNFO to monitor temps and my CPU will be fine (all while idle) and sit at 35 degrees then for no reason (nothing loaded, literally booted up and left) will ramp up to 70+ degrees where this process will continue. Once it starts that's it like every 60 seconds it'll ramp up then die down but then idles at around 50-60.

I then look at some info and decided to take a screenshot of something in red. I've pasted it. Now the screenshot taken was down after my computer literally loaded which is when my temps will idle at 35ish, so before the spike thing begins.
Any help on this would be great because the HWiNFO tooltips said something about it not being below 90% or above 110% - I have no idea what any of it means but some help would be fantastic as I've just pulled my PC apart and dusted, repasted etc etc to help cooling and these spikes arent changing, the whole pc feels really off. Like small things, if I click once it clicks twice now and again or keyboard keystrokes dont register or things taking long time to load etc.

my specs are:

- Ryzen 5950x
- Fractal Lumen cooler
- Meg Ace Max mobo
- 16 gig 3200 ram
- Corsair 1600AX

Thanks.

 

[Martin]

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

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...

www.hwinfo.com

 

[Zach]

Hi there,

I'm a bit of a novice with computers but I've noticed my fans are ramping up and down so I downloaded HWiNFO to monitor temps and my CPU will be fine (all while idle) and sit at 35 degrees then for no reason (nothing loaded, literally booted up and left) will ramp up to 70+ degrees where this process will continue. Once it starts that's it like every 60 seconds it'll ramp up then die down but then idles at around 50-60.

I then look at some info and decided to take a screenshot of something in red. I've pasted it. Now the screenshot taken was down after my computer literally loaded which is when my temps will idle at 35ish, so before the spike thing begins.
Any help on this would be great because the HWiNFO tooltips said something about it not being below 90% or above 110% - I have no idea what any of it means but some help would be fantastic as I've just pulled my PC apart and dusted, repasted etc etc to help cooling and these spikes arent changing, the whole pc feels really off. Like small things, if I click once it clicks twice now and again or keyboard keystrokes dont register or things taking long time to load etc.

my specs are:

- Ryzen 5950x
- Fractal Lumen cooler
- Meg Ace Max mobo
- 16 gig 3200 ram
- Corsair 1600AX

Thanks.

View attachment 7978

Enable from Sensor settings the "Show Tooltips" so you can read a short/general info about every sensor when you hover pointer over it, like this:



What you see with temp and hear with fan, is totally unrelated to PRD.
The temp that controls the fan is the "CPU (Tctl/Tdie)" and if you observe it you will see that this reading is jumping all around, all the time.
Its typical Ryzen 3000/5000 behavior. How much it jumps up/down depends on what you running at time, whats on background (tasks) and most likely from Windows Power plan. If you use a high performance plan it makes the CPU boost react aggressively even to small loads and so the (Tctl/Tdie) temp increase rapidly because it reports always the highest CPU spot temp.
Its best to use a balanced power plan.

If you already use a balanced power plan you can compensate a little the fan rapid ramping(up/down) by inserting a delay (delta) time on the fan from the fan control app you are using. If you dont use any app then from BIOS. By setting a 3~5 seconds delay(delta) on the CPU fan it makes it ignore the temp ups and downs for the given time (3~5) seconds. If the load of the CPU is sustained, so the temp increase is sustained then after the 3~5secs the fan will increase rpm. Same for temp decrease.

 

[fatehasfans]

Enable from Sensor settings the "Show Tooltips" so you can read a short/general info about every sensor when you hover pointer over it, like this:

View attachment 7986

What you see with temp and hear with fan, is totally unrelated to PRD.
The temp that controls the fan is the "CPU (Tctl/Tdie)" and if you observe it you will see that this reading is jumping all around, all the time.
Its typical Ryzen 3000/5000 behavior. How much it jumps up/down depends on what you running at time, whats on background (tasks) and most likely from Windows Power plan. If you use a high performance plan it makes the CPU boost react aggressively even to small loads and so the (Tctl/Tdie) temp increase rapidly because it reports always the highest CPU spot temp.
Its best to use a balanced power plan.

If you already use a balanced power plan you can compensate a little the fan rapid ramping(up/down) by inserting a delay (delta) time on the fan from the fan control app you are using. If you dont use any app then from BIOS. By setting a 3~5 seconds delay(delta) on the CPU fan it makes it ignore the temp ups and downs for the given time (3~5) seconds. If the load of the CPU is sustained, so the temp increase is sustained then after the 3~5secs the fan will increase rpm. Same for temp decrease.

Thanks Zach but I was using the Tooltip and as I said I'm a bit of a novice and can't get my head around what any of the PRD stuff is or means. That's why I came here. Been trying to Google it but I really don't know where to start when it comes from what I should be paying attention to in relation to CPU performance or GPU/Ram for that matter.

I don't know what voltages are or where they should be etc etc all I know is that my computer for a long time doesn't 'feel' right (slight input lag, keyboard keystrokes and mouse clicks not always occurring and just general poor performance from the rig I have. I mean my son has an AMD 3700X and a 5700XT and his temps don't go above 50 whereas mine are hitting 70-80 just when I open a web-page? Even playing games my temps sit at 70-85+. I've checked reddit forums and it would seem people are saying this is normal behaviour for AMD 5950X, that it ramps up at the slightest thing because of how 'aggressive' it is but how can this be an acceptable outcome when a build that is £1500+ cheaper doesn't have such issues?

It seems ludicrous to accept that a 'super' processor hits 80 degrees while doing mundane things such as opening webpages or simple tasks, I mean isn't that what you pay for? If my son's £100 chip sits at 40 degrees doing small tasks then shouldn't a super chip sit even lower, like, say 20 degrees? How has this been normalized and accepted? I just don't get it but again I'm a novice but jeepers I thought bigger builds could cope easier with more, have I missed something because I feel like selling this shit and downgrading to actually upgrade in performance :/ Strange indeed.

 

[Zach]

I understand your frustration
I can probably help you but I'll need more info about you PC behavior, settings etc.
Temp wise dont forget that the 3700X is a 65W TDP (88W max PPT) CPU and the 5950X is a 105W TDP (142W max PPT) one. Now that is under 100% load but it can make a difference and in middle/high loads too. We will get back to that later.

Looking closely on your screenshot I noticed your DRAM voltage (1.2V) being to low for such a PC. Meaning your ram doesn't run its full speed but probably remains at default speed (2133MHz/1066MHz base clock) with no XMP/DOCP enabled. DRAM voltage is supposed to be around 1.35~1.36V for full speed support, and the board sets it there auto if the DRAM XMP profile is enabled in BIOS.

At this point I will ask you to right click on the HWiNFO system tray icon > settings and click (on) to "Snapshot CPU Polling"
By tooltip description the CPU core real time metrics (speed, voltage etc) are more accurate and I trust the developer about it.



---------------------------------------------

First of all, there is almost no chance the sluggish feel and the lag of the PC to your commands to be caused by temperatures like 80+°C.
The max operating temp of 5000series is 90°C and even up to that point the CPU just throttle its speed and voltage to control temp, and it shouldn't cause any noticeable lag except for low scores in benchmarks, games and so on. Unless maybe the throttle is too aggressive. But thats not your case as your CPU doesn't hit that point.

Lets see the RAM speeds and everything else related directly to it.

This is your CPU under the IHS (Integrated Heat Spreader)



2x8core (2xCCD) chiplets on top and the SOC (System on Chip) chiplet at bottom containing all the I/O (Input/Output) stuff like all modern CPUs have. That would be the memory controller, PCI-E/USB controllers etc.
This is a the layout of the above with the 3 main chiplets.



As you can see the DRAM on the right is directly connected to CPU SOC and the Unified Memory Controller (UMC) inside it.
UMC is connected to CPU cores through the Infinity Fabric (IF) interconnection (or DataFabric, DF)

Now in order to all work properly (about DRAM) at their highest possible of their potential, some speeds must work at the highest possible BUT also must match with each other. If not then large amounts of latency is created between DRAM and the CPU cores and/or the rest of the system.

The real DRAM speed is half of the advertised one. If you have 3200MHz sticks then its real speed is 1600MHz
So, these speeds must match:

MCLK or MEMCLK = 1600MHz (MEMory CLocK)
UCLK = 1600MHz (Umc CLocK)
FCLK = 1600MHz (infinity Fabric CLocK)

A larger screenshot of HWiNFO sensors window of your system would help a lot like the one below at full screen with the same sections visible.
Use the blue arrows at bottom left to expand to multiple columns (4 at least for widescreen). You can adjust the inner column spaces from top bar.

Points of interest in red boxes


Alternatively you also use this but its strictly for DRAM settings info and states only current values for the aforementioned DRAM related speeds. Its better to see the Max/Min too as per HWiNFO.




EDIT:
Almost forgot it...
Check if you're having any WHEA (Windows Hardware Errors Architecture)
Easiest way is with HWiNFO
Its the last sensor on same window

 

[Zach]

Also its best (theoretically) for your board to have the latest stable BIOS (from MSI)
Windows should be having the latest AM4 chipset drivers from AMD directly, not MSI.

https://www.amd.com/en/support/chipsets/amd-socket-am4/x570

There is 1 for X570/X570S

 

[fatehasfans]

Also its best (theoretically) for your board to have the latest stable BIOS (from MSI)
Windows should be having the latest AM4 chipset drivers from AMD directly, not MSI.

https://www.amd.com/en/support/chipsets/amd-socket-am4/x570

There is 1 for X570/X570S

Hi Zach,

Just wanted to say a big thanks for providing the information. Every bits helps!

Ok, so I've done what you've asked and compiled the information together. I've uploaded screenshots of my bios pre and post-XMP same with HWinfo64 screenies.


Pre XMP




BIOS Pre-XMP




Post XMP



BIOS Post-XMP


Now the thing is my temps are all over the place, start my computer and they could race to 80-90 (in the bios! see pics) then when I load the computer they will settle down to around 40-60. But the moment I move a mouse, click a button or load a webpage boom straight up to 70-80 for a minute or two then down again to 50-60 and its rinse and repeat. This is before games :'(

I've a Lian Li O11 Evo case Fractal AIO, 3 Lian Li fans sucking air in from below and an exhaust fan (the AIO fans are bracketed to the top of the case - I did mount it vertically initially but I thought that may have been the temp problem so I changed the radiator position to the standard top of case. I also recently put new paste on the CPU and re-seated it just in case and nothing.

So, I'm completely miffed. Any help would be fantastic, thanks.

 

[Zach]

Glad to help on 1 issue!

I think there is indeed something going on...
It is unacceptable for the CPU to be supplied with 1.4V while you're in BIOS and its speed is 3.6GHz

My previous R5 3600 and my current 5900X both worked at base clock speed in BIOS (like yours) BUT with 1.0V and not 1.4V

See below on the right of the screenshot


From your last 2 HWiNFO screenshots I see the CPU voltage also around 1.4V
1.394~1.400V as CPU Core Voltage SVI2 TFN reports on the first screenshot (By the way this is the one and only sensor you will looking from now on for CPU voltage) And all values are about the same (current/min/max/avg)

Is there any chance that you by mistake set the CPU Vcore to a static value at some point? (1.4V) Or did you change anything else in BIOS?

Unfortunately your sensors screenshots are taken with very short time line (or no time line at all) as I can see on the bottom right on them. The clock = (00:00:02), (00:00:00)
So I cannot tell if voltage is fluctuating as it should. You should let it write data for at least 30min before you take the shot, like mine below. Use you PC for your regular staff

Also I would like to see some sections that are folded on screenshots.

These exactly on the left:


Also you can reduce the columns into 4 from 5 so making them more spacious and the numbers are all fully visible. Some of them are hidden with dots... We really dont need to see anything past the DIMM temp sensors.
Except maybe for the last one, the WHEA.
The last (4th) column you can have it scrolled all the way down so WHEA sensor is visible.

 

[Zach]

Forgot to ask...
Is there any change in how the PC feels to you after enabling XMP profile?

Also what windows power plan are you using?

 

[fatehasfans]

Forgot to ask...
Is there any change in how the PC feels to you after enabling XMP profile?

Also what windows power plan are you using?

Hi Zach,

Again, thanks for getting back. I'll do everything you said and get back. Regarding XMP everything felt the same and as far as setting VCORE values I've not touched anything. One time I did follow a guide on how to reduce input lag...

This is the video

...but then I just set bios back to default because it didn't fix the problem I had. Other than that I run everything at stock value straight out of the box. I use the balanced power program .. again, even with windows I tend to use default/stock settings.

 

[Zach]

Well this guy likes to cripple all modern features of a todays CPU (Dynamic behavior) because probably thats how he knows things from 15+years ago, and cant get his head around the complexity or the benefits of the dynamic operation.
I recommend to all when I can to stay away from static speeds and voltages.

Clearly whatever problem you're dealing with (lag) has nothing to do with if the CPU is working dynamically or not. Both my Ryzen CPUs, former and latter, was/is working as it should with all these features enabled.

Looking forward for your info

 

[fatehasfans]

Well this guy likes to cripple all modern features of a todays CPU (Dynamic behavior) because probably thats how he knows things from 15+years ago, and cant get his head around the complexity or the benefits of the dynamic operation.
I recommend to all when I can to stay away from static speeds and voltages.

Clearly whatever problem you're dealing with (lag) has nothing to do with if the CPU is working dynamically or not. Both my Ryzen CPUs, former and latter, was/is working as it should with all these features enabled.

Looking forward for your info

Hi Zach,

I think this is what you requested. If you need anything else just let me know. Learning as I go!

Thanks again

 

[Zach]

As a matter of fact I do not see anything unusual besides the high CPU temperatures.
I can tell that you did some gaming(?) by the GPU metrics (280W GPU PPT..!!)

After some web research I found that this board (MEG ACE MAX X570 and X570S) supplies the CPU in BIOS with values like 1.40~1.50V up to 1.55V sometimes depending the CPU installed. And thats for both 3000/5000 series.
Couldn't find why or if anyone managed to turn volts down when the CPU power control and boost is on Auto.
Though... how many pictures I saw the CPU temp inside BIOS was always somewhere between 40°C and 50°C, maybe little over 50, but not 80°C like yours. Big difference!

You said that you reapplied paste under your AIO's water block.
Do you have any software you use to observe any info from the AIO?
Like water temp, fan and pump speeds.
I am asking because if the CPU gets inside windows to 70~90°C the water temp must also be high, so the fans must be high and confirm also that the pump is working at the right speed.

----------------------------------

Now since the introduction of 5000series CPUs AMD also introduced PBO2 (Precision Boost overdrive 2) that contains the so called Curve Optimizer.
This is powerful tool to control the power (W) and the performance of a 5000series CPU.

All in all you can:

1. Decrease power(W) (=lower temps also) while maintaining same performance
2. Decrease power(W) and increase performance (in this case the lower power and lower temp delta will be smaller, and also the increased performance)
3. Maintain same power(W) and increase performance (no difference in power usage = same temp)
4. Decrease both (highest power/temp reduction, but also performance)
5. Increase both (higher all = power, temp, performance)

How far all 5 can go depends from a lot of parameters. Such as CPU quality, cooling and/or motherboard capabilities maybe etc.

I strongly suggest to watch the following video thoroughly as this guy explains it and demonstrates how it can be done.
He is doing the 3rd option of the above 5 but once you understand the concept you can do what ever you want.
I'm happy to try to clarify anything after you watch it.
Personally I'm doing the 5th option within some strict boundaries. Those are mainly the CPU voltage and the CPU EDC (Amperage).
My temps have gone slightly up but going from ~70C to ~74C, on both single and multi threaded (as 100%) workload is really ok.

-----------------------------------

Off topic(?)
Whats up with that drive (ST315003...) that you have a warning on it? (Drive warning = Yes).
I can see the "Total Host Reads" and the value is extremely high and also it has done some realocation which is not a good sign. Whats the use of that drive? Is it SSD NVMe?

 

[fatehasfans]

As a matter of fact I do not see anything unusual besides the high CPU temperatures.
I can tell that you did some gaming(?) by the GPU metrics (280W GPU PPT..!!)

After some web research I found that this board (MEG ACE MAX X570 and X570S) supplies the CPU in BIOS with values like 1.40~1.50V up to 1.55V sometimes depending the CPU installed. And thats for both 3000/5000 series.
Couldn't find why or if anyone managed to turn volts down when the CPU power control and boost is on Auto.
Though... how many pictures I saw the CPU temp inside BIOS was always somewhere between 40°C and 50°C, maybe little over 50, but not 80°C like yours. Big difference!

You said that you reapplied paste under your AIO's water block.
Do you have any software you use to observe any info from the AIO?
Like water temp, fan and pump speeds.
I am asking because if the CPU gets inside windows to 70~90°C the water temp must also be high, so the fans must be high and confirm also that the pump is working at the right speed.

----------------------------------

Now since the introduction of 5000series CPUs AMD also introduced PBO2 (Precision Boost overdrive 2) that contains the so called Curve Optimizer.
This is powerful tool to control the power (W) and the performance of a 5000series CPU.

All in all you can:

1. Decrease power(W) (=lower temps also) while maintaining same performance
2. Decrease power(W) and increase performance (in this case the lower power and lower temp delta will be smaller, and also the increased performance)
3. Maintain same power(W) and increase performance (no difference in power usage = same temp)
4. Decrease both (highest power/temp reduction, but also performance)
5. Increase both (higher all = power, temp, performance)

How far all 5 can go depends from a lot of parameters. Such as CPU quality, cooling and/or motherboard capabilities maybe etc.

I strongly suggest to watch the following video thoroughly as this guy explains it and demonstrates how it can be done.
He is doing the 3rd option of the above 5 but once you understand the concept you can do what ever you want.
I'm happy to try to clarify anything after you watch it.
Personally I'm doing the 5th option within some strict boundaries. Those are mainly the CPU voltage and the CPU EDC (Amperage).
My temps have gone slightly up but going from ~70C to ~74C, on both single and multi threaded (as 100%) workload is really ok.

-----------------------------------

Off topic(?)
Whats up with that drive (ST315003...) that you have a warning on it? (Drive warning = Yes).
I can see the "Total Host Reads" and the value is extremely high and also it has done some realocation which is not a good sign. Whats the use of that drive? Is it SSD NVMe?

Hi Zach,

Thanks again for getting back.

I'll try and go through what you've sent.

Firstly, on fan software I did use Lian Li's own software (L-connect) as well as MSI's mystic light. Due to a windows reinstall (hoping to solve the temp problem) I no longer have them installed but I was only (or so I thought) using them for RBG control which I no longer care for atm (until I get everything fixed) so should I install them again?

Secondly, I've been seeing reddit threads about fan curves but its something I know nothing about, but again this makes me think 'is it really necessary to create fan curve patterns for a powerful computer so it can browse the internet without hitting 90 degrees?' seems utterly unnecessary as I think this stuff should (and used too!) work fine just straight out of the box but again any help/tips there would be fantastic.

Third, I wouldn't have a clue how to reduce power but I will watch that video, thanks.

Four, regarding the ST3 it's a mechanical hard-drive I use for storage I've seen that warning for a long time now. A friend of mine told me throw out all my mechanical hard-drives (I've two) he was saying that's where my input lag problems are stemming but to me that doesn't really make sense since I had the input lag when I was using those drives as my main drives and people then told me it was because I wasn't using SSD drives so I went and bought those and moved the mechanical drives to serve as storage yet lo and behold I still have the input lag before that someone said it was my monitor so £750 later and playing on a new Rog swift I still have input lag, in fact the reason I have a new build is because people kept telling me to replace one thing or another with something new and I still have it :'(
But if you think the drive could a problem then I'll certainly remove them.


I've attached a quick snapshot of my average use as I feel this is a more accurate representation. This is me basically browsing Ebay (only one web page loaded), as I said before the moment I move the mouse or open a program or webpage the fans go berserk and temps shoot up, when I close everything and do nothing they will come down again to settle around 50-60 but the moment I load something and/or actively browse I'm sitting at 70-80+ until I stop. So, essentially this is what things look like when I'm surfing the web on a single webpage. Even if it's normal is this right? I mean how can such a powerful CPU do this? A £600 cpu nearly melting to surf the internet even though it's backed by an RX6900XT? It makes me want to go back to my 780Ti and AMD A10 as I never had anything like this, not a peep! Even when doing heavy stuff, but browsing the internet a £2.5k PC chokes?? :/



P.S. Regarding the 1.4v you think everything is ok there?

Thanks.

 

[Zach]

Firstly, on fan software I did use Lian Li's own software (L-connect) as well as MSI's mystic light. Due to a windows reinstall (hoping to solve the temp problem) I no longer have them installed but I was only (or so I thought) using them for RBG control which I no longer care for atm (until I get everything fixed) so should I install them again?

Personal opinion (and some experience) is that any software created to control RGB can cause from minor issues up to serious problems, such as conflicts, interference or what ever else you may think.
I do not care for RGB either, never did, but now days is kind of "necessary evil". I used to install ASUS Aura with my old ROG RX580 just to keep RGB turned off. My current MSI 5700XT GamingX has RGB (Just a GamingX logo) and I did install MSI's software 2.5years ago to turn that off but quickly throw it out.
Because of HWiNFO I was able to tell that the CPU was not "resting" as before during idle times or had the same behavior, even on low loads. That software was keeping the CPU on the edge of the seat for boosting all the time. And I think ASUS's Aura did something similar.
Since my upgrading to the 5900X from 3600 I also install a new G.Skill Trident Z neo RGB ram kit. I'm not considering to install any kind of software for that. It just doing its rainbow thing and I'm not paying any attention to it.

However...

I use Gigabyte's APP Center software just to install SIV (System Info Viewer) because it lets me control my case fans, with assigned temps across the system with curves of course).
I could avoid this completely because the board's BIOS has the ability to let you associate any fan (plugged to its headers) with any temp on the board, (CPU, chipset, VRM etc) and you can create fun curves of your own, besides the 4 modes (quiet, balanced, performance, max) that already has in there. SIV settings (if present) overwrite those in BIOS.
I keep SIV because after almost 3 years I didn't notice any issue, and because its inconvenient to go into BIOS every time you want to change something.

I'm also using Corsair's iCue software because I want (at any given time) full control and ability to check my AIO's metrics. The software provides 3 different fan curves for the radiator's fan (silent, balanced, extreme) and some times I may use the low or medium, but also I can create my own curves, as many as I want. Very handy... I can also set the water pump in 2 different fixed speeds (silent, extreme).
The AIO's fan curve are in fact associated with the water temperature and not the CPU temp.

Thank fully iCue is compatible (in my case) with HWiNFO so no issues/conflicts there and I can see the AIO's sensors through it

This is right now with room AC on, and 26-27°C room ambient. Outdoor is 32°C.
Typical usage (web browsing, Steam platform downloading a game update, watching vids, and also having game vikings: war of clans running on PlariumPlay platform.







If Fractal has any kind of software to control your AIO you can give it a try.
But all this cant explain the high temp of 80 in BIOS, or the lagging you experiencing.
I would make sure again that the contact between CPU and water block is tight. Paste ok, but the mounting mostly.
As I said before, on these kind of temperatures the water temp should also be high. There in no point for the AIO fans to speed to high rpm if the water is not warming up as it should under 80+ CPU temps.

-------------------------------------

Third, I wouldn't have a clue how to reduce power but I will watch that video, thanks.

The video is made for any user and mostly for inexperienced ones.
The benefits you can see from this could be huge and I can help you also.
I strongly suggest for you to see it when you have the time.

-------------------------------------

Four, regarding the ST3 it's a mechanical hard-drive I use for storage I've seen that warning for a long time now. A friend of mine told me throw out all my mechanical hard-drives (I've two) he was saying that's where my input lag problems are stemming but to me that doesn't really make sense since I had the input lag when I was using those drives as my main drives and people then told me it was because I wasn't using SSD drives so I went and bought those and moved the mechanical drives to serve as storage yet lo and behold I still have the input lag before that someone said it was my monitor so £750 later and playing on a new Rog swift I still have input lag, in fact the reason I have a new build is because people kept telling me to replace one thing or another with something new and I still have it :'(
But if you think the drive could a problem then I'll certainly remove them.

I'm done with mechanical drives around 5 years ago. Since I've bought 5 SSDs (3 currently on this PC and the other 2 (the oldest) moved on my 2 10yo and 11yo SonyVaio laptops.
Personally I wouldn't trust this drive any longer. It could be the an issue for the system (depends on type of usage), but surely not responsible for your high temp.

 

[Zach]

P.S. Regarding the 1.4v you think everything is ok there?

Yes, as it seems its common on this board, and probably not responsible for 80C in BIOS. Something else is causing this and more and more Im thinking about waterblock mounting.

 

[fatehasfans]

Yes, as it seems its common on this board, and probably not responsible for 80C in BIOS. Something else is causing this and more and more Im thinking about waterblock mounting.

I edited my post to include a snapshot of basically how I would use my computer outside of gaming it's only 4-5 mins of browsing but it represents the normal use. Is that normal?

 

[Zach]

Yes I saw it and I have a few observations about it, makes things a bit more clear, but I will get back to you later.
I'm also making some points at my last (#15) post, so I will let you see that too.

Even though I have the 5900X and you the 5950X, the 2 CPUs are (should be) very close on behavior. They should be very similar on idling or on low-middle loads.

Here is mine (about 12 min long) for you to compare and not just temps but also watts, speeds, activity and of course voltage >> "CPU core voltage SVI2 TFN" sensor a few rows under CPU (tctl/tdie) temp

This is right now with room AC on, and 26-27°C room ambient, with Chrome opened (20+ tabs), Steam platform downloading a game update, watching vid, and also having game (vikings: war of clans) running on Plarium Play platform.
Also see my system tray icons...

 

[fatehasfans]

Yes I saw it and I have a few observations about it, makes things a bit more clear, but I will get back to you later.
I'm also making some points at my last (#15) post, so I will let you see that too.

Even though I have the 5900X and you the 5950X, the 2 CPUs are (should be) very close on behavior. They should be very similar on idling or on low-middle loads.

Here is mine (about 12 min long) for you to compare and not just temps but also watts, speeds, activity and of course voltage >> "CPU core voltage SVI2 TFN" sensor a few rows under CPU (tctl/tdie) temp

This is right now with room AC on, and 26-27°C room ambient, with Chrome opened (20+ tabs), Steam platform downloading a game update, watching vid, and also having game (vikings: war of clans) running on Plarium Play platform.
Also see my system tray icons...

View attachment 8012

Yea looking at your temps our CPUs are world's apart. I just reseated the processor and applied new paste and still everything is still just heating up at the slightest instruction. You think I should just RMA?

 

[Zach]

Not really, not yet at least. I believe more and more that this is not CPU's fault. It seems that the board is keeping the CPU too jumpy on voltage and/or boost.

Look closely my screenshot again down below from before at those red boxes and recreate a new HWiNFO to around 10-12min as you did before with the same light loads but please do enable the following option first, otherwise any comparisons will not have any meaning.

The snapshot CPU polling. (as I said in post #5 earlier) it makes a difference because it minimizes the "observer effect"

In physics, the observer effect is the disturbance of an observed system by the act of observation.[1][2] This is often the result of...



See the red boxes and take account, and compare them with the 5950X, on all aspects. (min, max, avg) All have their meaning...



It makes a world of a difference for the CPU to get less voltage at any given speed and load.
That's why I strongly suggest to go through that video about the PBO2 guide.