Spikes in SOC voltage 7800X3D

[Blue Panadol]

Hey guys, so a few weeks ago i manually set my SoC voltage to 1.2V since i noticed it was a bit high on auto (1.25V), however ever since playing bo6 i noticed weird large spikes even though I manually set the voltage. Is there a reason for this? I believe it should have something to do with the Load line calibration but I'm not sure how to stop it from spiking to such a high level. Any help is appreciated.


I also noticed some weird behavior with my FCLK and UCLK frequencies, they are set to auto in the BIOS and the frequency isn't flat 2000MHz, 3000MHz.
I should also note that the motherboard is the gigabyte B650 Eagle AX, bios version F4

 

[Martin]

This is a different case than the double-value spikes. FCLK/UCLK are based on bus clock (BCLK), so if that's not exactly 100 MHz, then you'll see a similar difference.

 

[Blue Panadol]

This is a different case than the double-value spikes. FCLK/UCLK are based on bus clock (BCLK), so if that's not exactly 100 MHz, then you'll see a similar difference.

I see, so then would BCLK also cause this ~6% increase in SoC voltage ?

 

[Blue Panadol]

I did a bit of research and I'm seeing that the variations in BCLK is due to spread spectrum. I haven't encountered the option to disable spread spectrum but from what i understand the function itself shouldn't be an issue. I still don't understand why my SoC voltage, VDD Miscellaneous, and global Freq. Limit has this weird spike of ~6.5% I'm hoping this is a reading error and not an actual voltage spike. Any feedback on this is appreciated, thanks.

 

[Zach]

This is a different case than the double-value spikes. FCLK/UCLK are based on bus clock (BCLK), so if that's not exactly 100 MHz, then you'll see a similar difference.

I see, so then would BCLK also cause this ~6% increase in SoC voltage ?

It’s been a while (months) since I’ve seen the double misreported values.
But occasionally I have the plus ~6% mostly on FCLK, UCLK, MCLK. Very rarely on SoC voltage too.
It must be a bug like the double ones. It’s not from BCLK for sure.

R9 5900X

 

[winston1989]

Even my 9800x3d x670e, 6000cl28 having this exact issue. Sometimes it takes few hours for it to happen, sometimes half a day. Currently testing chipset drivers cause it's the last thing on my mind that can cause issues.

 

[Martin]

V

Thread 'HWiNFO64 CPU Die (Average) Spike?'

Apr 24, 2022

As you can see from the image below, I had HWiNFO64 open and minimized to tray for a while. Wasn't playing games or anything but when I opened it back up it showed that the Maximum CPU Die (Average) was 104.2c. How could this be possible? As you can see from my other temps from the sensors, everything appears to be fine. Is this a bug or?

• Venoxium
• 7800x3d cpu die cpu die (average) cpu die hot temperature
• Replies: 361
• Forum: HWiNFO® 32 or HWiNFO® 64 Bug Reports

 

[winston1989]

Looks like a power spike, the only time i saw such spikes is when using y-cruncher and overclocked ram

 

[winston1989]

try to eliminate hwinfo, see if you see such behavior with lets say OCCT

 

[winston1989]

V

Thread 'HWiNFO64 CPU Die (Average) Spike?'

Apr 24, 2022

As you can see from the image below, I had HWiNFO64 open and minimized to tray for a while. Wasn't playing games or anything but when I opened it back up it showed that the Maximum CPU Die (Average) was 104.2c. How could this be possible? As you can see from my other temps from the sensors, everything appears to be fine. Is this a bug or?

• Venoxium
• 7800x3d cpu die cpu die (average) cpu die hot temperature
• Replies: 361
• Forum: HWiNFO® 32 or HWiNFO® 64 Bug Reports

Intersting that 5900x max temp is 90 then it will throttle, but on hwinfo it didn't detect thermal throttle

 

[winston1989]

I've read the topic, false reading they say. In my case I can see massive stutters when spiking happens.

 

[winston1989]

Chipset driver removal didn't help and I have a correction on my previous statement regarding freezing, it was due to dxvk api and not the spike related.