Question Ryzen 5 2600X running very hot...

Jul 2, 2019
16
0
10
Hello everyone!

I'm new to this forum, so I hope I don't make any mistakes regarding the rules.

I built myself a new computer a few days ago and even though all seems to go pretty well (running very quietly, no errors/crashes, good FPS in different games, ...), I am concerned about the temperature of my new Ryzen 5 2600X.

My build consists of:

  • AMD Ryzen 5 2600X (using the stock Wraith Spire cooler, but cleaned the pre-applied thermal paste and used TITAN TTG-G30030 instead)
  • MSI X470 Gaming Pro Carbon
  • Corsair Vengeance RGB Pro 16GB DDR4 DIMM 2666Mhz/16 (2x8GB) Black
  • Corsair Carbide Series SPEC-DELTA RGB
  • Corsair RM650X 80+ Gold
  • AMD Radeon R9 380 4GB

The thing is, the CPU seems to run very hot while stress testing and has alot of fluctuating values. The only thing I changed on the BIOS was an upgrade to the latest version and I enabled the XML2 (?) profile for my RAM.

zWXc77T.png


vJncBRL.png


Could anyone please assist me on this matter? Also, just noticing now, the CPU and chassis #1 fans... Is that correct? Could I have mixed them up?

Thank you in advance and kind regards,

Sindarin
 
I shall give it another shot (it's some KBxxxx install that caused the error), after trying to delete it, the Ryzen Master installation gave me a message that it wasn't installed correctly or something like that.
 
These is the message and error that appears each time I try to install Ryzen Master:

Hc5sJ33.png


Q7xlxRV.png


It says: "0x80070666 - There is allready another version of this product installed. Cannot complete the installation of this version. Use the configuration screen to configure or remove the allready existing version."

Thing is, there is no other version installed.
 
Ah, found how to uninstall the correct update so Ryzen Master could install the other one and install successfully! 😊

So uhm, if I would stress test the CPU now, it would give the correct temperatures? Meaning I almost didn't blow up my CPU?

AimXBRo.png
 
This seems strange. I was under the impression the 2600x doesn't use a thermal offset. I can use HWinfo64 to get accurate temperature readings. The CPU temp will show up as Tctl/Tdie in HWI64 because the sensors read the same. This indicates that the 2600x doesn't use a temp offset according to HWI64's author:

https://www.hwinfo.com/forum/Thread-Is-CPU-Tctl-value-is-still-a-Tdie-value-on-Ryzen

Oddly, the max 108c temp that Aida64 was showing in the OP's screenshot is exactly 20c above where mine throttles which is where the temp offset would be. Could this be affected by an old bios or chipset drivers? Why would his chip have an offset while mine doesn't?

Either way the alarming temperature is clearly an incorrect reading. That being said, If you are using the stock wraith spire it's a bit inadequate for the CPU IMO. It's a great stock cooler but the out of the box boost settings push it's limits a bit farther than I'd be comfortable with long term. Especially in a poorly ventilated case. I was able to maintain higher overall game clocks by manually overclocking with the stock cooler. Even with a -100 mv offset, the stock settings would cause the CPU to boost over 4 ghz for 10-15 minutes while gaming and then down clock (because of thermals) below what I was able to achieve with a conservative all core overclock. The best solution is to get a better cooler if you want all of the performance you payed for. You can maintain over 4ghz even in stress tests with a half decent tower air cooler or 240 AIO and a few bios tweaks.
 
  • Like
Reactions: Sindarin
Actually, HWinfo and Core Temp have the offset preconfigured on current versions. So those are usable as well. HWmonitor and the other not-preferred utilities, probably don't.

I'm not sure about the specifics of excluded models from the offset issue, but I AM sure that HWmonitor, regardless of what you are using it on, tends to be half baked at best most of the time. I recommend avoiding HWmonitor, Openhardware monitor, Speccy and most of the bundled utilities as well. HWinfo, Core Temp, Ryzen Master, these are trusted utilities that tend to be accurate regardless of chipset or platform, and are kept religiously up to date.
 
  • Like
Reactions: Sindarin and rigg42
I'm not sure about the specifics of excluded models from the offset issue

https://www.guru3d.com/articles-pages/amd-ryzen-5-2600x-review,7.html

Temperature Reporting
To keep a "consistent fan policy," AMD is forcing 10C offset on only the Ryzen 2700X processors. This makes them report temperature a good 10C above what the sensor reads. The rest of the models on the AM4 Platform have the same maximum tCTL value. The primary temperature reporting sensor of the AMD Ryzen processor is a sensor called “T Control,” or tCTL for short. The tCTL sensor is derived from the junction (Tj) temperature—the interface point between the die and heatspreader—but it may be offset on certain CPU models so that all models on the AM4 Platform have the same maximum tCTL value. This approach ensures that all AMD Ryzen and thus Ryzen Threadripper processors have a consistent fan policy.

Product NameTrue Junction Temp (Tdie)tCTL Offset for Fan PolicyTemp Reported by (tCTL)
Threadripper 1950X43°C27°C70°C
Threadripper 1920X43°C27°C70°C
Ryzen 7 2700X38°C10°C48°C
Ryzen 7 270038°C0°C38°C
Ryzen 5 2600X38°C0°C38°C
Ryzen 5 260038°C0°C38°C
Ryzen 7 1800X38°C20°C58°C
Ryzen 7 1700X38°C20°C58°C
Ryzen 7 170038°C0°C38°C
 
  • Like
Reactions: Sindarin
Good enough. Regardless, it doesn't change the accuracy or relevance of HWmonitor though. Ryzen master is what should be used when possible, as it will be the most accurate of all the monitoring utilities regardless of with, without or in spite of, any other considerations.
Yeah HWmonitor is crap. I personally don't see any compelling reason to use RM over HWI64 though. I don't even install it on my builds. To each his own.
 
  • Like
Reactions: Sindarin
I sure came to the right place with the right people with my questions, although I must admit, I don't really know that much of what I'm reading all above, looking up the different terms like "offset" etc. now and trying to understand what it all means. I have run a test with Prime95 for about 20 minutes and got the first screenshot while testing, the other one below that was taken around 5 minutes after the Prime95 test.

While testing, the temperature of the CPU went up quickly, but stayed at a solid 94.75 degrees Celcius, it kept at that exact temperature for another 10-15 minutes after reaching it.

ZnK9u7m.png


CUVbCDa.png


I really hope to learn and understand what all the date you've provided me above means and I really appreciate the effort and time you guys/girls are taking to help me figure this out.

Thank you and kind regards!
 
What VERSION of Prime95 did you run? 26.6 or the latest version? If you ran the latest version, did you disable the AVX and AVX2 instruction option at the bottom of the popup menu? If not, please disable those two settings and then retest. Your temps are still too high and it's likely due to the use of AVX instructions.

Also, WHICH test in Prime did you run? Small FFT? Large FFT? Blend? Small FFT is what you SHOULD use for thermal compliance testing. If you used anything other than Small FFT option torture test then you need to retest and use the Small FFT option instead.
 
  • Like
Reactions: Sindarin
Yes, I disabled the AVX and AVX2 instruction option in the popup, I also followed your other instructions and ran the Small FFT torture test this time. The version of Prime95 I'm using is "p95v298b5.win64" (29.8). The temperatures seem to be the same though... 🙁

Before the test:

DXL6dqU.png


Settings used:

PgRtXyb.png


Quickly rises up to about the same result (within a minute) and staying at that temperature, ran for about 30 minutes:

ygdFE8W.png


I highlighted two things that changed while testing:

The values below TDC (CPU) sometimes changed to a yellow color and once I saw the "AUTO" move the green selection to "Precision Boost Overdrive" for a second, then it changed back to "AUTO" and stayed there. Meanwhile, the CPU temperature stayed 94.75 degrees Celcius.
 
This is normal for the stock CPU settings and cooler combination. As I mentioned previously, you need a better cooler. For some reason AMD decided to let the CPU run up temps this high out of the box on the stock cooler. They should have put this cooler on the 2600 and a prism on the X or tuned down the boost.
 
  • Like
Reactions: Sindarin
A negative voltage offset can help mitigate the temps some what but it won't solve the issue. I went through this song and dance recently with a 2600x. You'd actually get better temps and performance overall dialing in a 3.9 - 4 ghz manual OC.
 
  • Like
Reactions: Sindarin
Yeah, I believe you're absolutely right, but the things is that it would be kind of a small problem buying a new cooler atm... Vacation time with the kids, trying to save up for something special as a surprise vacation too at the end of the year, ... Is there maybe something I can do in the MSI BIOS/UEFI to make it more stable? As in, max performance, but not in a way that it goes into automatic overclocking mode or anything like that? I guess that's done with the offset voltages?
 
A negative voltage offset can help mitigate the temps some what but it won't solve the issue. I went through this song and dance recently with a 2600x. You'd actually get better temps and performance overall dialing in a 3.9 - 4 ghz manual OC.

Oh, allright, I got a message there was a new post while writing the previous' one... 😊 Would you mind helping me take the first little step by explaining what you mean with dialing in a 3.9 - 4 ghz manual OC. That involves changes in the BIOS/UEFI, but tbh, I don't know much about that.
 
Download Hardware info 64, Intel Burn Test, and realbench. Reset the Bios settings to default. Turn the memory XMP profile on. Set Vcore voltage to manual. Type in 1.25-1.3 volts in vcore. Find load line calibration in the power settings. Set it to a medium setting. Probably level 3 or 4 on MSI. Type a 39 into clock multiplier. Save changes and reset (F10). Open up HWI64 with sensors only checked. Open up IBT and run a test on its default settings. If it doesn't freeze, crash, give you a pop up telling you you're unstable, or hit a max CPU temp over 80 c you can run an 8 hour stress test overnight with real bench. Click the stress test tab then select 8gb from the RAM drop down and run the test. Adjust the voltage up in .025 increments or drop the multiplier down in 25 mhz increments if you aren't stable or are too hot. If the voltage/ frequency settings cause a post failure you can usually let it post cycle and after 3 times the bios will reset. Otherwise you can find the clear cmos jumper on your motherboard and short the pins while powered down. The location of the jumper is in the manual. 1.25 vcore/ 39 multiplier/ medium LLC will more than likely be completely stable.
 
  • Like
Reactions: Sindarin
Gahhhh, just get a cooler. Same as this thread.



You could even go with something cheap, and it would STILL be a lot better than that stock cooler.

PCPartPicker Part List

CPU Cooler: Deepcool - GAMMAXX 400 74.34 CFM CPU Cooler ($22.89 @ OutletPC)
Total: $22.89
Prices include shipping, taxes, and discounts when available
Generated by PCPartPicker 2019-07-03 23:07 EDT-0400
 
  • Like
Reactions: Sindarin
Thanks for your assistance, I will first try to figure it out in the BIOS/UEFI, but if that doesn't really work out, I'll definetely get another CPU cooler, didn't know they where that cheap actually...
 
That is the least recommended option that I would recommend. LOL. It will be better than the stock cooler, but it is not a terrific cooler. It is better than the Hyper 212 EVO and better than stock, but you'd be much better off if you went with something like what I recommended in the thread I linked to above, which coincidentally happens to be a fairly similar setup as yours.
 
  • Like
Reactions: Sindarin and rigg42
I have to agree with Darkbreeze that a new cooler is the best option. I don't think the manual oc is that hard to pull off but its still a band-aid for the cooler issue. If you wan't the performance you paid for when you bought your x cpu you need a better cooler. I've had good luck with the Deep Cool stuff. Good bang for the buck. I had a 2700x running full boost with good temps no problem with the RGB version of that same cooler he mentioned. I believe it has the same heatsink with a better fan and some RGB bling on the top.

Even if you do get the new cooler you can squeeze some extra performance out of the CPU by enabling precision boost overdrive and tweaking a negative offset voltage. Even if you don't enable PBO the offset voltage will optimize temps. I believe MSI added the offset voltage option in newer Bios'. Speaking of that updating your Bios is probably a good idea. You will need to stability test just like a manual overclock if you go down this road. It's certainly not completely necessary but its really not that hard. It's a good learning experience if you are into PC's.
 
Last edited:
  • Like
Reactions: Sindarin