[SOLVED] Confused about temp display for motherboard

lordbredd

Honorable
Jan 4, 2013
34
0
10,530
i have an asus m5a99fx pro r2.0 with an amd fx-9370 running at stock speeds installed using a nzxt kraken x61 cooler. hwmonitor gives me 2 temps for the motherboard section labeled "cpu" and "mainboard", and 2 more temps for my cpu labeled "package" and "cores". my question is what is the "cpu" label from the motherboard section reading? i had it go up to 81C today while the "package" and "cores" reading under the cpu section reached 67C at the same moment. i was playing monster hunter world and im scared i might be damaging my motherboard and/or my cpu. im in an area where the weather was 105F today and my place has no ac to help keep my rooms ambient temperature down so i know thats contributing to my high temps but i just want to know what is the max safe temp for the readings so i can quit and take a break to let my pc cool down.
 
Last edited:
Solution
So the way it works is, under a load, that number gets smaller and smaller. You do not want to ever see it go less than 10°C distance to TJmax. If it does, there is a problem. Lack of cooling, too much voltage, failing motherboard, something. On AMD, it is not a question of it being safe at that temp. If you are in the red, with no distance left to TJ max OR a negative result, you have a problem that needs fixed.

Since it is only an estimated guess at the temperature based on a formula, there is always a little room for error, and that could work against you in reality. Throttle remedies to prevent damage, don't always work as expected. Plenty of systems damaged because the owner thought the system wouldn't allow damage to occur...
You can't use HWmonitor to monitor thermals on that platform. You need to use either Core Temp with the "Show distance to TJmax" option enabled in the advanced settings, or AMD overdrive. AM3+ processors don't use a standard core/package temp monitoring scenario like other platforms including the more recent AMD Ryzen or all Intel platforms. It uses an estimated distance to TJ max.

This guide is being revised to include the more recent Ryzen AMD thermal guidelines, but the main info for prior AM3+ platforms is still there although the images are currently not being displayed. Should get you where you need to be though.

 

Paperdoc

Polypheme
Ambassador
Good info above. OP, just to help. Every CPU chip has a temperature sensor built into it and sends out the result on one of its pins to the mobo. That is the CPU temperature. Then there is at least one more sensor built into the mobo by its maker at a location they consider important and representative of the overall mobo situation. That's the mobo temperature, sometimes called System Temp. SOME mobos have additional temp sensors built into certain other components like the North Bridge chip of the Voltage Control system that you MAY use IF you are trying to concentrate a particular fan on that component.

There are almost NO mobos that can access and control temperatures on an added graphics card because there is no "standard" design for how to do that. But the graphics card itself normally has at least one temperature sensor in its GPU chip and its own automatic control system for its own cooling fan(s). To allow you to access this info and configure cooling on that card, software utilities normally are supplied with the card and installed when its Device Diver is installed. Often these are used via an icon in your system tray at bottom right of the screen. SOME third-party software systems can access that same info if they are written to work with whatever software tools the card maker is using.
 

lordbredd

Honorable
Jan 4, 2013
34
0
10,530
You can't use HWmonitor to monitor thermals on that platform. You need to use either Core Temp with the "Show distance to TJmax" option enabled in the advanced settings, or AMD overdrive. AM3+ processors don't use a standard core/package temp monitoring scenario like other platforms including the more recent AMD Ryzen or all Intel platforms. It uses an estimated distance to TJ max.

This guide is being revised to include the more recent Ryzen AMD thermal guidelines, but the main info for prior AM3+ platforms is still there although the images are currently not being displayed. Should get you where you need to be though.


i have taken this advice but now im concerned about what is the maximum safe temperature of my components. at idle using core temp with "distance to tjmax" enabled i get cpu 0 to read "55-63C to tjmax". i understand now that if it goes past tjmax the core will be throttled but will it still be safe to run at that temperature? i got amd overdrive as well and all the cpu cores are reading the same values of 36-43C while idling, but now there is tmpin1 that is at 75C. i tried reading around and i gathered that tmpin1 is the socket temp? if so then what is the maximum safe temperature for that as well?
 
So the way it works is, under a load, that number gets smaller and smaller. You do not want to ever see it go less than 10°C distance to TJmax. If it does, there is a problem. Lack of cooling, too much voltage, failing motherboard, something. On AMD, it is not a question of it being safe at that temp. If you are in the red, with no distance left to TJ max OR a negative result, you have a problem that needs fixed.

Since it is only an estimated guess at the temperature based on a formula, there is always a little room for error, and that could work against you in reality. Throttle remedies to prevent damage, don't always work as expected. Plenty of systems damaged because the owner thought the system wouldn't allow damage to occur. Damage starts happening at temperatures below the throttle figure, it just isn't instantaneous damage like what can happen when you are negatively past TJmax. Cumulative thermal fatigue can happen even when at close to TJmax temperatures for extended periods of time. It is best to consider ten degrees from TJmax AS TJmax, and stay away from that close of a tolerance.

You can simply forget about the actual temperature readings that say the CPU is "75°C" or whatever. They are not going to be right. Also, uninstall HWmonitor and never use it again IMO. Download and use HWinfo for OTHER sensors you might want to monitor.


Monitoring software

HWmonitor, Open hardware monitor, Realtemp, Speccy, Speedfan, Windows utilities, CPU-Z, NZXT CAM and most of the bundled motherboard utilities are often not the best choice as they are not always accurate. Some are actually grossly inaccurate, especially with certain chipsets or specific sensors that for whatever reason they tend to not like or work well with. I've found HWinfo or CoreTemp to be the MOST accurate with the broadest range of chipsets and sensors. They are also almost religiously kept up to date.

CoreTemp is great for just CPU thermals including core temps or distance to TJmax on older AMD platforms.

HWinfo is great for pretty much EVERYTHING, including CPU thermals, core loads, core temps, package temps, GPU sensors, HDD and SSD sensors, motherboard chipset and VRM sensor, all of it. When starting HWinfo after installation, always check the box next to "sensors only" and de-select the box next to "summary".


Run HWinfo and look at system voltages and other sensor readings.

Monitoring temperatures, core speeds, voltages, clock ratios and other reported sensor data can often help to pick out an issue right off the bat. HWinfo is a good way to get that data and in my experience tends to be more accurate than some of the other utilities available. CPU-Z, GPU-Z and Core Temp all have their uses but HWinfo tends to have it all laid out in a more convenient fashion so you can usually see what one sensor is reporting while looking at another instead of having to flip through various tabs that have specific groupings, plus, it is extremely rare for HWinfo to not report the correct sensor values under the correct sensor listings, or misreport other information. Utilities like HWmonitor, Openhardware monitor and Speccy, tend to COMMONLY misreport sensor data, or not report it at all.

After installation, run the utility and when asked, choose "sensors only". The other window options have some use but in most cases everything you need will be located in the sensors window. If you're taking screenshots to post for troubleshooting, it will most likely require taking three screenshots and scrolling down the sensors window between screenshots in order to capture them all.

It is most helpful if you can take a series of HWinfo screenshots at idle, after a cold boot to the desktop. Open HWinfo and wait for all of the Windows startup processes to complete. Usually about four or five minutes should be plenty. Take screenshots of all the HWinfo sensors.

Next, run something demanding like Prime95 version 26.6 or Heaven benchmark. Take another set of screenshots while either of those is running so we can see what the hardware is doing while under a load.

*Download HWinfo


For temperature monitoring only, I feel Core Temp is the most accurate and also offers a quick visual reference for core speed, load and CPU voltage:

*Download Core Temp
 
Solution

lordbredd

Honorable
Jan 4, 2013
34
0
10,530
im having issues running hwinfo64. makes my whole system freeze requiring a hard reset. im working on their forums to fix that atm as well. i ran prime 95 for an hour running amd overdrive and core temp side by side. i know its not long enough to really test thermal issues, but i had time constraints. amd overdrive had temps go down to 7C then stabilize around 21C while coretemp peaked at 21C but stabilized around 43C to tjmax. is amd overdrives "thermal margin core temps" monitoring something different from core temps "distance to tjmax"?
 
Last edited:

lordbredd

Honorable
Jan 4, 2013
34
0
10,530
got hwinfo to work afterall. here are my idle and load screenshots. the temperature im concerned about the most is the one in blue labeled "cpu" under the motherboard section because it reaches over 80C during load. i just dont know if it is actually reporting for the cpu when the cpu section is reporting different temperatures. if it isnt actually reporting for the cpu then what component is it reporting for and is it possibly getting damaged from these temps?
these are my readings after a cold boot and letting it idle after start for 5 min.
UwuvcL1.png

bfJm2TS.png

AR4rbAT.png

RmtgtB1.png


these next 4 are after letting prime95 run for an hour.
pqs8Owo.png

KzyNmOy.png

qOmL4e9.png

YBPuWQL.png
 
Last edited:

Karadjgne

Titan
Ambassador
From what I see, you only ran 7 workers out of 8 possible, so the temps you are getting from p95 won't be the max possible.
You can't treat FX like Intel. It simply doesn't work that way. Intel cpu's have a thermal strip embedded in the cores themselves, FX does not, it's external to the core, so right from the get-go physical temp readings will be off. FX internal core temp max is 62°C for all the FX, but because of the way temps are read, you actually seeing an accurate reading of that number is slim to none.

Most times, if thinking according to Intel, the closest number to reality for FX is package temp, not core temp. Once package temp hits @ 72°C ±, the cpu will start to throttle. Once package temp hits @ 82°C ±, you can expect sudden shut downs at any time. But what you are doing by that is Thermal Margining the cpu in your head. You see 75°C and think 'wow, close to shutdown limits, got 5-6°C left'. AMD Overdrive does that automatically, using complex algorithms that combine the vcore, current and other temps to realistically tell you exactly what you have left. But even that number in its exactness is pretty moot, what is important is what the number represents. If TM is in the 40's, you have no worries whatsoever, tons of room. If it's in the 20's, ok, there's a load but I'm still good. At 10, start to worry, at 5 be worried, if it hits 0 start saving data and consider which program will be immediately stopped, if -1 or more, forget saving, just shut stuff down.

TM gives you a working limit, without any guess work as to whether things are too hot or not. Simple.

The problem with many programs is labeling. There is no standard. One program cpu temp is actually the package temp, in another its core temp, some it's socket temp. Hwmonitor is really bad with things like tmpin, which for 90% of mobo's is 1 of 4 things, either the VRM mosfets, Northbridge, Southbridge or cpu. Which gets nuts on lga1155 or newer since there is no Southbridge, and Northbridge chipset no longer is the memory controller but the pci-express hub.
HWInfo is much better, but especially for FX, useless at reading temps as you have no idea which cpu temp is which, and which applies.

Stick to Thermal Margins, it's the only reliable indicator of temp, without having an actual temp number. It's not a quantifiable result.

Edit: just for info, core temp is immediate, it's pulled directly from the core itself, calibrated by intel. Package temp is an average of the last 256ms samples taken from core temps. Cpu temp is unreliable as normally its pulled from a diode underneath the cpu itself (those chunks dead center of the socket on the mobo) but can also be combined with sensors external to the socket, is not calibrated by anyone for accuracy and can be affected by VRM heat dumped into the ground plane. All that topped off with a guess as to what the cpu is actually at. And that's if it's not being labeled as something else.
 
Last edited:
At ten degrees margin remaining, I try to find the problem, and not allow it to get any worse.

What in the world are you talking about Karadjgne with the 7 workers, there are workers 0-7, which is 8 workers total and is totally normal.

I am seeing different thermal margins for AMD overdrive and Core Temp though. HWinfo can be ignored for anything related to the CPU. It won't be even remotely accurate. Both Overdrive (22 degree margin) and Core Temp (40 degree margin) show that under a full load there is a significant margin remaining, and that's good. While I don't like the fact that they show different margins, they both show enough margin for safety.

First, make sure that there is not a newer BIOS version available for your motherboard than the one you currently have installed. Second, make sure that both Overdrive and Core Temp are the latest versions of those programs. Overdrive is obsolete now since newer AMD products use standard thermal measurements and Ryzen master, but there are several versions out there so make sure yours is the latest version. Core Temp should be latest version if you downloaded it from the alcpu.com website and have version 1.4.
 

lordbredd

Honorable
Jan 4, 2013
34
0
10,530
One program cpu temp is actually the package temp, in another its core temp, some it's socket temp.
yeah this was my initial concern. is there any precedent with this motherboard and sensor programs where they have been correctly identified? i tried searching before making this post but couldnt find anything. if my motherboards "cpu" temp is actually the socket temp reaching 80C+ is that damaging my components?

First, make sure that there is not a newer BIOS version available for your motherboard than the one you currently have installed. Second, make sure that both Overdrive and Core Temp are the latest versions of those programs. Overdrive is obsolete now
i do have the most recent versions of all of these, but if overdrive is obsolete now is it even worth using/trusting? i was also concerned about the different temps reporting between the two programs and wonder if that could be a result of damaged components from years of thermal wear.

i understand having a 20C thermal margin, and what a thermal margin is, is good now so thank you for that Darkbreeze. with all my years of having amd components i have no idea why i never heard of thermal margins before.
 

Karadjgne

Titan
Ambassador
Hey DB, look again at the VID in the picture just below where op says 'after running Prime95 for an hour'. Core #0 is showing idle voltages, 0.8v, whereas the rest cores #1-#7 are showing load voltages of 1.4v. Core #0 clock speeds 1.4GHz, ratio 7x, cores #1-#7 running 4.4GHz, ratio 22x.

Core #0 isn't doing anything, Prime95 is running on cores #1-#7 only. That makes 7 workers, not 8. So 2x explanations, either op only ran 7 workers or a core has gone bunk and won't push that kind of load.
 
Last edited:

lordbredd

Honorable
Jan 4, 2013
34
0
10,530
that was just the instant i took the screen shot. i ran a blend torture test and when doing that the cores switch around like that if you look at the core temp readings in subsequent images. idk how or if i would stop that
 

Karadjgne

Titan
Ambassador
To answer Op, Overdrive is obsolete for only 1 reason. The FX are obsolete. Ryzens have a temp almost exactly the way Intel does theirs, so there's no real usage for Thermal Margins as such. However, since you do have an FX currently, any software from that generation of cpus is still valid for you. But as the FX really start getting phased out and replaced by newer cpus, getting ahold of valid software is going to get harder as websites quit offering it for download. So as Darkbreeze advised, grab all you can, from chipset drivers to bios to Overdrive and save a copy somewhere, so if you ever need it, you have it.

I saved all the last updates, bios etc to a DVD and stuck it in the drawer. If I have to scrap windows, or my drive dies, I have full backup utility seperate from what's installed.
 

Karadjgne

Titan
Ambassador
Never seen that, ever. Base speeds for an FX 8350 is 4GHz, boost 4.2GHz. Why voltages and speeds would drop to almost a sleep state, while still under a 100% load is beyond me. Normally usage will drop off, then voltages and speeds will follow, even on blend. Not other way around.

Prolly just me vs blend, being on mobile means blowing up zoom, and I wasn't looking at coretemp readings as those I'd trust to be accurate temps, i just looked at the more detailed info in hwinfo
 
Last edited:
Not uncommon at all. All systems with speedstep or Cool N quiet will drop to between 600-800mhz when idle unless you have the performance power profile enabled in the control panel and DON'T have the minimum processor power state manual set to 8%, which is where I like to keep it and where it lives normally on all other plans except the performance plan.

The only time ANY CPU stays at it's base clock is if all power saving features are turned off/disabled, so yes, that is normal for idle or not full load function.

The screenshots where the system IS under a full load, show 100% for all cores. Normal.
 

lordbredd

Honorable
Jan 4, 2013
34
0
10,530
if i try turning off cool n quiet or the other power limiting options in bios i get a system freeze when i run prime95 so i dont worry about those or clocking up at all. otherwise all my temp readings are safe then from what ive gathered correct?