PC Processors

[BiscuitBomber]

When running anything difficult on the PC including testing with Cinebench, Blender, Adobe and video editing. One of my cores is running hotter than the others as shown below. I've re-pasted but same issue is occurring has anyone else experienced this. I even ran the fans at 100% and still peaks instantly at 95 degrees just on that core. I haven't adjusted any overclocks PBO, Curve Optimizer etc. Everything is set to default minus AMD EXPO in the BIOS.

Hardware:

Processor: 7950X (Latest Chipset)
Motherboard: MSI X670E Ace (Latest BIOS)
GPU: EVGA 3080 TI FTW3
MEMORY: G.SKILL Trident Z5 Neo RGB Series AMD EXPO 32GB (2 x 16GB) 288-Pin PC RAM DDR5 6000 (PC5 48000)
PSU: EVGA 1000W P6
Cooler: 360 Phanteks AIO MPH

[BiscuitBomber · Accepted solution]

To give an update, I went ahead and swapped my CPU. New one is running warmer on all cores, another thing I noticed is my old one never peaked above 200 watts even though my BIOS was set to auto. The new one is hitting 230 watts and the per core is more evenly distributed heat wise too. Also it’s peaking almost 1000-1500  points higher with multicore on Cinebench and 100 points higher on single core. I’m not sure if I just had unlucky silicone the first time, but I’m happy everything seems to be normal now.

[misterj]

BiscuitBomber, please post a screenshot of RyzenMaster running Cinebench R23. Thanks, John.

[BiscuitBomber]

[misterj]

Thanks, BiscuitBomber. Please open a Support Ticket here. Provide both of your screenshots and ask them why they are different and if Ryzen Master (RM) will supply per core temperatures. Because the temperature shown in RM is red your processor is throttling. Please post what you learn. Thanks, John.

[BiscuitBomber]

Thanks, I completed that this morning. I was just wondering if anyone else ran into this as well. I can understand it throttling based off everything I read regarding the 7950X. I just thought it was weird that only 1 core was doing it vs the others.

[misterj]

I agree with you, BiscuitBomber. I have never trusted the free HW applications and only trust RM, so I was suspicious of your first post. However, there may be something strange wrong with your processor causing this and I hoped AMD Support would tell us. I think I saw this in an earlier post but never seen it myself. Hopefully we will get a definitive answer from AMD. Enjoy, John.

[DimkaTsv]

You have core 3 that runs at 82 degrees and core 5 that runs ar 88.5 degrees...
Ryzen 7000 boost behaviour will target max TDP OR 95 degrees temp. depeding on what will happen before.
But as CPU core frequency multiplier isn't working on per core basis it limits whole CCD frequency to same value. I am not sure about second one though.
So seems likeboost behaviour is working like intended.

To improve temps you must remove heat from CPU even faster. Like shaving a bit from IHS, using liquid metal, water cooling etc.

Core to core difference is to be expected. Some cores just generate more heat than others. Or it is harder to move heat away from some cores, because they can be surrounded by other cores for example.

[MADZyren]

You could check if there is a temperature problem by reducing CPU voltage. If all other cores start running cooler, but that one still goes to 95C there might be an issue. Also if you limit temperature to say 75C instead of default 95C, will it cause clockspeed of that core to drop more than with other cores or does the whole CCD run at remarkably lower clockspeed.

[BiscuitBomber · Accepted solution]

To give an update, I went ahead and swapped my CPU. New one is running warmer on all cores, another thing I noticed is my old one never peaked above 200 watts even though my BIOS was set to auto. The new one is hitting 230 watts and the per core is more evenly distributed heat wise too. Also it’s peaking almost 1000-1500  points higher with multicore on Cinebench and 100 points higher on single core. I’m not sure if I just had unlucky silicone the first time, but I’m happy everything seems to be normal now.

[misterj · Accepted solution]

Thanks, BiscuitBomber. It does look like you have a solution-Great!  Enjoy, John.