PC Processors

[vivraan]

Computer Type: Desktop, Custom build.

GPU: RTX 2070 Super, GDDR6 8GB, don't know whether SLI/Crossfire

CPU: AMD Ryzen 7 3700X, no varying overclocking settings.

Motherboard: MSI B450 Tomahawk Max

RAM: 2x16GB Corsair something - RAM supports DDR4-3000MHz 16-20-20-38 A-XMP, and defaults to 2133MHz-15-15-15-36 - A2 and B2.

PSU: Provide the model number and its rated output level.

Case: https://www.amazon.in/Zebronics-Typhoon-Gaming-Computer-Black/dp/B076SL8GC9Stock Cooling.

Operating System & Version: I.E. Windows 10 Pro Version 1909 , Ubuntu 17.04, etc.

GPU Drivers: NVidia GeForce Game Ready 451.85 hotfix.

Chipset Drivers: AMD Chipset Software 2.07.14.327

Background Applications: Chrome, Unity or Godot Engine, File Explorer

Description of Original Problem:

With or without OC, irrespective of undervolting VCore to 1.35V, the idle temperature is always 50-55°C, spiking 60°C (2-6% utilisation, Chrome in the background) can easily reach 85°C during heavy load, and has recorded doing a 100°C in some HWinfo64 logs which were made while performing benchmarks.

I haven't tested the temperature during gaming, but I feel a warm breeze at my feet. Restarting the CPU solves the stuttering temporarily. Even the second monitor blacks out for a second at a time.

Now coming to the prime95 torture tests: Blend testing without changing any settings crashes the PC (HWinfo64 logs reveal overheating towards the end) within 3 minutes, but applying A-XMP seems to solve the monitor blackouts, and the PC works okay on the Small FFT and Large FFT tests reaching 85°C.

I have the latest AMD Chipset drivers installed as well. Should I change the thermal paste or cooling?

Troubleshooting:

Initial CPU ID validation

Most recent CPU-ID validation

HWinfo64 logs that I generated while running the test initially and also while testing other BIOS se...

My HWBOT Profile, recording recent benchmarks performed.

EDIT: Tested Outer Worlds, goes between 77-83°C.
EDIT 2: A-XMP DOES NOT SOLVE BLACKOUTS.

[black_zion]

You should be using Ryzen Master to monitor temperatures and such, though both will display the incorrect voltage if you manually set it.

85*C under heavy load is to be expected with the stock cooler and Turbo enabled, it's AMD's way of ensuring maximum performance is safely extracted, as the thermal maximum is 95*C. Idle temperature is very high, but then again if you're measuring the idle temperature while using an internet browser, especially Chrome, elevated temperatures are to be expected.

Prime95 and Ryzen processors don't mix well, and Prime95, along with all other artificial power virus tests, have long been known to cause false positives for instability due to the way they work, only made worse when it started using AVX. I've seen mine fail it even at stock settings on my 1800X and now on my 3700X and I have a much more powerful than stock cooler, but I've also hammered it with AIDA64 cache tests and had it grinding away for 48 hours straight on video encoding without a hiccup with my overclock and undervolted settings (4.3ghz 1.2v) which I run every day.

[vivraan]

That's good to know, but how should I address the stuttering and blackouts problems? It seems that using the A-XMP profile does away with the screen blackouts, but the stutter is ever-present, just not as frequent.

[vivraan]

Again, apologies for the late response.

While you had your stock cooler on, were there times when your PC stuttered, or the temps ever touched 100°C?
I tried undervolting VCore but going below 1.35V always caused PC crashes owing to memory controller problems.
It's also obvious that affecting my ambient via air conditioning helps the CPU temps as well, dipping by as much as 7-8°C. A new cooler might be needed.

[vivraan]

Of course, all temps reported rn are via Ryzen Master.

[vivraan]

Ambient seems to be affecting my PC - it rained today, so Ryzen Master repports 42-44°C.

[vivraan]

Well, I elected to have both my case and cooling changed. What cases would you prefer?

[vivraan]

EDIT: added that some tests caused the processor to spike at 100°C without failing.