Hwi64 showing CCD1 temp 145c on r5 3600x

[crosslhs82]

Hi Everybody And Happy New Year!!!!!!!!!!

Fisrt time poster here and having an issue that I need to know about.
Specs: R5 3600x
Asus Rog Strix X470 F Gaming
G.skill Trident Z Royal Gold 3600 Cl16 16-16-16-16-36 1.35v running At 3733 15-15-15-15-30 1.40v
Zotac Gtx 1070 Ti 8gb
Corsair H110i cooler set as exhaust on top with Noctua 3000 rpm IPPC fans
Samsung 970 evo+ 250gb M.2 Os drive
Crucial Mx500 500gb M.2
Samsung 860 evo 250gb 2.5 sata
2x Pny cs1311 120gb ssd sata
Seagate 3tb spiner
Corsair Cx650m psu
In a Thermaltake F31 Suppressor 4 case fans 2 front 1 bottom 1 exhaust so all total 3intake 3 exhaust cooling is not the issue.

So formalities out of the way.
I'm having an issue where trying to tune my New 3600x i just got from Micro Center running Occt and Hwinfo64 last version and current version showing that CCD1 is super hot on the max values and throttling but Occt isn't crashing or showing errors. Picture attached in the picture you will see the values in Red. While running Occt Large Data Set Only Hwinfo64 was running with it Rm was opened after the issue came about.
Is this an Issue or not That I need to Be Worried ABout??? Or Is it a Bug on Occt or Hwinfo64?
Any Advice or thoughts Appreciated Thanks for the time.

 

[Martin]

Happy New Year to you too!
I cannot see any extensive CCD1 temperature on the screenshot you posted. When did you record such high temperature? Did it perhaps happen when RM was running along with HWiNFO only?
The Thermal Throttling flags seem indeed to be asserted, but this might be due to some other issue, perhaps a BIOS update might fix this problem.
As long as you don't see any performance degradation, I wouldn't worry much about this.

 

[crosslhs82]

Thank You for your quick repsonse.!!!!
Rm was only opened after the supposedly throttling started.
If i would have captured the times before, it showed 206.9c on Ccd1 with a -8.0c on the cpu tdie above ccd1 lowest temp value.
Only Occt and hwinfo64 running during Occt large data set run.

 

[Martin]

That must have been some glitch in the CPU, the fact that it affected Tctl/Tdie confirms this. This has most probably caused assertion of the thermal throttling triggers too.
I can try to filter out such invalid values, but the bug needs to be fixed by AMD in the AGESA/firmware.

 

[crosslhs82]

Ok i will try to get a new screen shot for you but i Have to leave town today for my work.
It could be possibly sunday before I get back home.
At least it sounds that I don't have to have a fire extinguisher next to me. LOL
Thank You for your time I will get back to you.

 

[Zach]

Happy New Year to all!

Ok i will try to get a new screen shot for you but i Have to leave town today for my work.
It could be possibly sunday before I get back home.
At least it sounds that I don't have to have a fire extinguisher next to me. LOL
Thank You for your time I will get back to you.

Not a chance to be able to work above 100C as throttle temp is 95 for ZEN2.
Make sure you have the latest BIOS/UEFI version for you board. If you do have the latest, then you have to wait for a new one.

 

[crosslhs82]

Thank You and Happy New Year

It is on the newest bios for the mobo 5220 and newest Amd and Asus chipset drivers.
Have tried soc voltage at 1.05v and left it at auto still same thing.

Will respond when I get back home Sunday and can resimulate the condition and get a fresh screenshot.

 

[Zach]

Is this happening a lot? I mean, like every day or every boot? Every time some "heavy" load running, benchmark or stability test?
What I want to say is that if you can replicate the issue frequently enough then you may want to try the same but with lower DRAM and InfinityFabric speed.
I have experience same strange behaviour of invalid temp readings (negatives) and thermal throttle reporting (YES) at HWiNFO without real thermal issue, and even missing sometimes a whole (mainboard) section in sensors mode running DRAM/UMC/IF coupled at 1866 or 1900MHz. When at 1800 or 1833 nothing of this ever happens.

If you want to keep the speed of 1866 or 1900 at least you may try to raise some RAM/UMC/IF related voltages and see if this go away.
All this could be "just" some erratic communication between CPU and the system. And quess what? InfinityFabric is in the middle.
Whats your...
1. RAM voltage?
2. cLDO VDDP voltage?
3. cLDO VDDG voltage?
4. CPU SoC voltage I can see its 1.08 which is the default.

RAM (MEMCLK), voltage is DRAM or MEM VDDIO
UMC (UnifiedMemoryController = UCLK) voltage is cLDO VDDP
IF (InfinityFabric = FCLK) voltage is cLDO VDDG

Do not alter any voltages if you dont know what you are doing or what are the safe levels of these voltages!

When you return you can also post a screenshot of the latest version of RyzenMaster. All the info will be in that screenshot.
Like this...

 

[crosslhs82]

i uninstalled hwi64 latest version and installed the beta version and here is the screenshot of hwi64/ running Occt and ryzan master after test was ended.
I have no timeline on when the readings went to showing the throttling and - temps but was ok for the 1st hour then i went to bed.

https://imgur.com/a/TqhPKaT

IN Bios current setting
43.00 for 4.3 ghz vddcr cpu voltage @1.29375
soc auto @1.10v
Dram voltage 1.39000v Vddg v Auto no value Just lowered it more and will start retest
Vddg Iod Auto No value
all rest on auto
Dram timings control section
CLdo vddp voltage -- auto no Value

 

[crosslhs82]

hopefully better screenshot

2nd screenshot of lowered ram voltage notice the _8c cpu-t-die
still running Occt that temp is now -10c

 

[crosslhs82]

new screenshot of negitive voltages

 

[crosslhs82]

5hrs 34mins results

 

[crosslhs82]

here is a new screenshot with ram voltage set @1.35v
notice -c temp and -voltages and -204.9

closed out hwinfo64 and reopened new showing better while still running the same occt.

 

[crosslhs82]

Ryzen Master at idle just opening windows snipping tool.

 

[crosslhs82]

Occt running 3mins into run with R master open

 

[crosslhs82]

Have ran p95 since 4pm yesterday so 12 hrs and no issuse's on Hwinfo64 or p95 with Darbreeze's settings from Tom's hardware
Custom settings of 512k / 4096 / 15min fft's / 11000mb's ram and no avx. again no isssuse's p95 or hwi64.

Is it possible there is a conflict between Occt and Hwinfo64 ?

 

[crosslhs82]

full screenshot

 

[Martin]

Theoretically a conflict between OCCT is possible, but I'm not sure about this as I don't have insight into its code.
Have you observed such erratic values when running any other software than OCCT ?

 

[crosslhs82]

Thanks for responding Martin.
Occt is the only 1 I have noticed so far.
No issues on p95 as shown.
Will run more p95 and Realbench stress test over this next weekend when I get back home.

 

[crosslhs82]

have ran p95 blend with no avx's for 9hrs 16mins with no problems reported.
next will try realbench stress test.