I would start running memtest one stick at a time to start. While CPU voltage does seem a tad high, it's actually supposed to go down under load in most cases. Crashing that early seems like ram to me. I'd start there for sure.

 

Or try a different PSU

 

I also had a 500-watt power supply
I tested that one too, but it still didn't work
changed graphic card to rx 580
changed nvme to adta 2560gb
but didnt work

 

iI tested the RAM, no issues were found

 

stuck at 0
& this also shows CLOCK_WATCHDOG_TIMEOUT

 

My 5800x would not even POST with the 500W supply I had tried.

Did you try a BIOS reset or reflashing with the the current version?

 

I updated the BIOS to the latest version, but it didn't make any difference.
I even tried older versions, but it still didn't work

 

I even tested my CPU on an X570 motherboard - still had the same issue

 

My 5900x had similar symptoms on my old build with Gigabyte d3 1.0 x570 which had a known heat issues that can cause problems if ignored, and my memory was also the main problem it died after many years.. but the board was killing SSD's and my new RAM kit. My processor is ok.. and volts was never that high.

 

During Windows installation, my system freezes between 17% to 30%

 

4x16GB 3200 is not guaranteed. Take 2 sticks out and retest. You also do not need more than 1.375 ish volts on the cpu at stock. No way it should have that applied with no overclocking.

 

All settings are on default and I did not modify them. Previously, I had two 16GB RAM sticks and a 256GB A-data SSD. Later, I upgraded to four 16GB RAM modules and a Lexar 620 NVMe 512GB SSD. After this upgrade, these problems appeared and became worse

 

4 sticks is always a bit harder on memory controllers. Pair that with the large capacity per stick, and you can have some struggles....especially when getting all 4 to run at high speeds. I think you should either try setting the ram to 2666 or removing 2 sticks and see how the machine behaves. I understand that you want to have an eyepopping amount of ram at its max speed, but are you even presenting this machine with the types of workloads that would require it? 32gb is plenty these days, unless you are using the machine for rendering or video work or computation. Are you doing those things?

I know your memtest came back clean but that isn't stressing all of the ram at once, but rather checking sections for error-free operation. You can pass memtest and still have system instability.

 

Yes, I work in architectural visualization

 

There is also the idea of DIMM placement. Your strongest ones in Bank B and weakest in A. I was only working with SR DIMMS and they would not run XMP until I shuffled them around. One pair I have would do 4000 and the other could hit 4133...but all together they would lock and hang at 3600.

 

When I ran 4 dimms I also did the swap banks test which helped and had to set max freq for ram and also in for the inf. fab for example 3600mhz ram.. 1800mhz and 1800mhz match in bios auto was failing.