Concerned about Temps I'm seeing in CPU monitoring software after recent BSOD

[GamerGuy83]

So I posted about this in another forum (Tomshardware.com), but I'd like a second opinion from folks I know watch and play with temperatures more (especially for gaming). I'll post my system specs and some screen shots at the bottom of my post related to the issue.

The situation is this I recent got a BSOD in the middle of playing a game (Sword Coast Legends) that had a hardware error code (WHEA_uncorrectable_error). After Testing I noticed I was getting some strange temperature readings in HWMonitor regarding my CPU. At first I thought perhaps my CPU fan was worn out and needing replacing, but decided to do further testing to be sure. I Went and Downloaded 4 other programs and used them to monitor my hardware to see if I got the same thing. Those programs being HWinfo, Speedfan, ASUS AI Suite III, and AMD Overdrive. Only HWinfo seemed to collaborate the strange situation I was getting with my temperatures as the rest all say that I'm fine. However, I then also noticed that the side of my case that rests directly behind the Motherboard was getting a little warm right were it sits behind my CPU any time I'm running a game. The rest of that side of the case (and the entirety of the rest of the case) remains cool as does the air blowing out of my case.

That said I then noticed that in HWMonitor, HWinfo, and Speedfan all also list 4 temperature sensors that are always in the range 101c to 104c for temperature (these appear to be sensors on the mobo though not 100% certain). They list this even at system idle 15 minutes after starting up the computer. So I am now wondering if the issue is actually to do with faulty heat sensors either in the CPU or in the Motherboard.

System specs:
ASUS Crossblade Ranger (has AMD A88x Chipset with newest 1702 Bios)
AMD A10-7850K processor 3.7 Ghz (using Stock Fan and heatsink)
EVGA Geforce GTX 750 video card (latest drivers)
Creative Soundblaster X-fi series XtremeGamer
Thermaltek Armor Series A90 Case (Improved with a 2000rpm rear exhaust and a added side in take fan 1440rpm)

Here are links to screen shots from 3 of the programs as they sit in idle.

HWinfo (idle) -

https://imgur.com/KVrJmKY

Speedfan (idle) -

https://imgur.com/7vQMrGj

AMD Overdrive (idle) -

https://imgur.com/5A6A1Nk

Here are the screenshots for wile running a game.

HWinfo -

https://imgur.com/Xx21KDx

Speedfan -

https://imgur.com/0sAQV6l

AMD Overdrive - http://imgur.com/jUwSEG4

 

[Mandrake4565]

Not sure if any programs other then AMD overdrive can read the temp sensors on the A10 Apu correctly. Also note Overdrive doesn't give you a temperature it gives you distance to TJ Max. You had 25c headroom while playing, so you're good to go. As far as the BSOD that is another issue and doesn't look to be caused by the temps.

 

[GamerGuy83]

Not sure if any programs other then AMD overdrive can read the temp sensors on the A10 Apu correctly. Also note Overdrive doesn't give you a temperature it gives you distance to TJ Max. You had 25c headroom while playing, so you're good to go. As far as the BSOD that is another issue and doesn't look to be caused by the temps.

Yes I'm aware of the difference between actual CPU temperature, and the meaning of Thermal Margin. I posted that info as it's still a decent indicator of my temperature situation, though it might be effected if I do have faulty heat sensors. Which given that my case is getting warm in the spot behind where the CPU rests i am concerned if that might be the case.

 

[PetrolHead]

However, I then also noticed that the side of my case that rests directly behind the Motherboard was getting a little warm right were it sits behind my CPU any time I'm running a game. The rest of that side of the case (and the entirety of the rest of the case) remains cool as does the air blowing out of my case.

That's the socket radiating heat to the side panel. AMD Overdrive doesn't say anything about socket temperatures, but the ones displayed by HWiNFO seem to be in the correct and safe ballpark. It shouldn't be anything to worry about, unless they're also totally unreliable. Some of the other readings on HWiNFO seem really odd indeed. Vcore 2.768 V? That can't be right. Also, if that package temperature was correct, your CPU should have shut down automatically. I'd do as Mandrake4565 suggested and trust AMD Overdrive on the package temp.

BSODs are often associated with RAM issues, so I'd run Passmark's MemTest86 to see if it finds any errors. If you can, let it finish the four passes it will do on standard settings, but note that this may take even a few days depending on your system.

 

[Mandrake4565]

Gamerguy, what I mentioned about the Overdrive software particularly "thermal margin" was not intended to offend. You would be surprised how many people we see come through here that think it's an actual temperature.

As far as the 103c readings on HWInfo they are erroneous the only temp that I see that may be accurate pertaining to the Apu is the 54 c "current temp" in HWinfo. You could try using double sided tape and mount an 80 mm fan on the backside of the motherboard, if it will fit between the panel and motherboard tray. This will help keep it a bit cooler if it is a concern.

Could you open Cpu-Z and take screen shots of the following tabs Cpu, Spd and Memory this will give us a peek under the hood so to speak and maybe we'll see something out of wack. Additionally, could you use the in forum hosting tool to post them, some people due to work restrictions cannot see pictures hosted from outside sources. Eventually if Imgur either goes out of business or your account expires they become dead links which isn't helpful to someone visiting this thread in the future. Here is how.

 

[DNRDustin]

Hwmonitor doesn't work with the apu's so I hear. I would put a fan on the socket. Run p95, run custom blend utilizing 75% available ram, and see what happens. Uninstall everything but AOD, and use that to monitor your system.

 

[GamerGuy83]

Alright here are the images for the CPU-Z stuff.

 

[Mandrake4565]

Thanks you're tRC is set a bit tight on your memory, try relaxing to from the 29 it is set at to 33 which is what the sticks call for. Also make sure you have the Dram Voltage set at 1.65.

 

[GamerGuy83]

How do I adjust the tRC on the Dram?

 

[GamerGuy83]

I went into BIOS and the DRAM's voltage was already set for 1.65v like it should be for 1600Mhz. The Row Clock Timing is set for Auto and wont allow me to change it to anything else in the Extreme Tweaker section of the ASUS BIOS menu. Which I'm assuming the Row Clock Timing is what you meant by tRC.

 

[Mandrake4565]

You may have to set the overclock tuner to manual or Dram to manual to be able to adjust the timings. It may be a non issue but I'd like to see it where the XMP profile 1600 says it should be at due to the BSOD. It's possible it could be the culprit.

 

[caddi daddi]

and just some info to make your life easier, you can open all the copys of cpu-z you like all at the same time.

amd overdrive is all i can trust for temps with my apu rigs.

 

[GamerGuy83]

You may have to set the overclock tuner to manual or Dram to manual to be able to adjust the timings. It may be a non issue but I'd like to see it where the XMP profile 1600 says it should be at due to the BSOD. It's possible it could be the culprit.

From what I see ASUS's over clocking options (at least on this board) in the bios only allows me to change certain things on the memory everything else is default to auto with out option to change.

Also I seem to have all the correct settings already for my specific type of memory being used, the CMX8GX3M2A1600C9 model. From what I see listed else where besides Amd's own page (linked below) most folks get a tRC of 29 on it though that isn't listed by AMD.
http://www.corsair.com/en-us/cmx8gx3m2a1600c9

 

[Mandrake4565]

From what I see ASUS's over clocking options (at least on this board) in the bios only allows me to change certain things on the memory everything else is default to auto with out option to change.

Also I seem to have all the correct settings already for my specific type of memory being used, the CMX8GX3M2A1600C9 model. From what I see listed else where besides Amd's own page (linked below) most folks get a tRC of 29 on it though that isn't listed by AMD.
http://www.corsair.com/en-us/cmx8gx3m2a1600c9

I've never had my hands on that specific motherboard but I'd be shocked if it didn't have option to change the timings. It's likely just a setting that needs to be changed in order to do so. As far as the tRC I'd bet you're correct that you should be able to run at 29 without an issue. Chances are it is not the reason for the BSOD, but when trying to find what could be causing a BSOD I try to eliminate all possible causes. Having the tRc set tighter then spec could cause an issue.

 

[GamerGuy83]

I've never had my hands on that specific motherboard but I'd be shocked if it didn't have option to change the timings. It's likely just a setting that needs to be changed in order to do so. As far as the tRC I'd bet you're correct that you should be able to run at 29 without an issue. Chances are it is not the reason for the BSOD, but when trying to find what could be causing a BSOD I try to eliminate all possible causes. Having the tRc set tighter then spec could cause an issue.

Your right i certainly should be able to change it, as it looks like the Bios just isn't letting me right now. There are other options in the timing overclock section that I can change but half the options are not being given. I think i know why this is though....

When I originally installed these RAM sticks my board read them as 1333 Mhz when, as you can see by their model, they are in fact 1600 Mhz. So to get my board to run the memory at what it is supposed to be. I had to manually go in to bios and force the board to run it at 1600 Mhz in the Overclock settings. Perhaps given the recent BSOD it might be said the RAM was read as that lower speed in the beginning because I simply have a defective set of RAM. It would explain why the XMP profile says tRC should be at 33, but the board isn't running it at that.

 

[Mandrake4565]

It's not likely that the sticks are bad, usually when they are you'll be having constant issues or the Pc will not boot up at all.

 

[GamerGuy83]

It's not likely that the sticks are bad, usually when they are you'll be having constant issues or the Pc will not boot up at all.

I don't so much mean bad but I mean defective, I have had BSODs in the past ever since using the sticks and more frequently since installing windows 10 months ago. All of those, however, gave me software error codes and I traced the most recent ones to a windows 10 compatibility issue in Nvidia's Geforce Experience. As soon as I removed that program my problems seemed solved for months. Now I have this BSOD which unlike the others this was the first BSOD to be hardware related with the error code.

 

[Mandrake4565]

Try running Memtest for at least 4 passes if you get any errors it's either the sticks, timings or possibly the Imc isn't getting enough voltage.

 

[GamerGuy83]

Try running Memtest for at least 4 passes if you get any errors it's either the sticks, timings or possibly the Imc isn't getting enough voltage.

Haven't run memtest yet, but I'm going to in hopes of getting a more detailed bit of info on what my problem is. However, Windows Memory Diagnostic during it's 2 passes has now detected a memory problem, it just wont tell me exactly what the problem it detected is.

 

[Mandrake4565]

Ok if windows memtest is detecting an issue then it is likely the culprit. Unfortunately, it's not going to pin point what exactly the issue is. As I said earlier it could be the sticks themselves, the timings ie the tRc being a bit too tight or the Cpu IMC needs a bit more voltage. I would start by figuring out how to change the tRC to 33 or try setting the memory to the D.O.C.P. profile 1 if the option is available and you haven't already. Maybe read through this review it may help.