PCI-E Gen 4 issues with ryzen 2000 and 3000's ?

[alexdpars]

So I am getting BSOD's on my PC and it is driving me nuts. It pushed me to change many parts on my PC without being able to solve the issue.

My base config was

3900x

x570-I asus rog strix

980 pro 2tb ssd

32GB of ram tridentz gskill

3060ti asus tuf

Fractal Design 650W SFX 80 plus gold

 

So I had theses blue screen that at first I thought where caused by my PCI-E riser cable so I changed it (also an upgrade went from gen 3 to gen 4). The issue seemed to disapear but came back a week later. Being tired of if I pulled everything out from my case and let it run on my desk open air. The issue went away to come back a week later. I tried just running furmark but it was hard to reproduce the issue just with furmark (I don't even know if that was happening). I've ended up being able to reproduce the issue no problems by hitting the ssd at the same time I hit the GPU with some load. So I start an unzip of cinebench, install a game on steam and run furmark. Doing that reproduce the issue each time. When I only install and decompress cinebench the issue don't happen.

 

So at first I tried an other motherboard (from my server) which is 2700x with b550 ASUS prime and the issue was still happening. I switched the PSU for a seasonic (the one from my server) and the issue took a lot longer to happen few hours instead of a few minutes) but still happened I thought it was just the PSU at that point. I also tried a 970 pro ssd I had with a fresh windows install same issue. I tried with an other GPU an old 1080 I had lying around. couldn't reproduce the issue with the 1080 in one hour. Put back the 3060ti issue happen in 5 minutes. Must be the GPU I guess ? I bought a 6700 xt put that on the mobo run ddu install drivers. Run my benchmark problems and the issue rise in a few minute. Then I thought that may be the current that is the issue. I had a ups on the basement Plug that in run a test. Everything went fine for an hour no crash. That was the issue right ? Just to confirm I stopped the PC and retry it from the wall. Took about 20 minutes but it finally failed. Felt like I finally found my issue. Then had a few blue screen memory related. I reseated the ram but then the issue came back. IDK what to do anymore right now I have basically only the SSD that is still the same as the old system but I have no idea what is happening. I will try to switch the SSD again  and see, but doing that mean that I have no more component from the old system running but I don't have much hope at this point.

 

In the event viewer I see the crash in the system logs but it says that the dump creation failed. I installed a second SSD and tried to set that as a location for the mini dump thinking that if the SSD is locked up may be dumping the info to a second drive would work but no...

 

The only thing I can see at this point is that the CPU's being older might have issues handling Gen 4 GPU's but even setting the x16 port to gen 3 didn't help.

 

Let me know what you guys think.

[alexdpars]

At first the issue was happening while gaming. Then I found the way to reproduce the issue more easily by hitting the SSD and GPU at the same time. Which is what happen while gaming but less harder.

[alexdpars]

Ho and yeah in the process I managed to brick my motherboard (waiting for a new one). After a crash my asus booted up in recovery mode asking to insert a usb key with the bios CAP file. Tried multiple ports and waiting but it never resolved the issue, Iread somewhere that I needed to reboot the PC in order for the ezflash to work but it didnt work either now when trying to post the MB is stuck on the red led.

[Somerandomtechyboi]

board is cooked

atleast you have a board on the way

 

if you still have bsod issues stresstest rams see if theyre unstable both xmp on and jedec (use something like prime95 largeffts)

 

if still bsoding and rams are reporting fine try a diff windows install on a seperate drive

 

ram and mobo are the #1 suspects when your system is doing odd stuff or being unstable, moreso the board as anything weird that isnt ram is almost always gonna be board related

[alexdpars]

But that is the second board I try. How could both board been cooked to have the same problems ?

 

I will try what you tell me but the ram is also different. I changed it with the board and cpu for the ones in my server.

[MadAnt250]

8 hours ago, alexdpars said:

So I am getting BSOD's on my PC and it is driving me nuts. It pushed me to change many parts on my PC without being able to solve the issue.

My base config was

3900x

x570-I asus rog strix

980 pro 2tb ssd

32GB of ram tridentz gskill

3060ti asus tuf

Fractal Design 650W SFX 80 plus gold

 

So I had theses blue screen that at first I thought where caused by my PCI-E riser cable so I changed it (also an upgrade went from gen 3 to gen 4). The issue seemed to disapear but came back a week later. Being tired of if I pulled everything out from my case and let it run on my desk open air. The issue went away to come back a week later. I tried just running furmark but it was hard to reproduce the issue just with furmark (I don't even know if that was happening). I've ended up being able to reproduce the issue no problems by hitting the ssd at the same time I hit the GPU with some load. So I start an unzip of cinebench, install a game on steam and run furmark. Doing that reproduce the issue each time. When I only install and decompress cinebench the issue don't happen.

 

So at first I tried an other motherboard (from my server) which is 2700x with b550 ASUS prime and the issue was still happening. I switched the PSU for a seasonic (the one from my server) and the issue took a lot longer to happen few hours instead of a few minutes) but still happened I thought it was just the PSU at that point. I also tried a 970 pro ssd I had with a fresh windows install same issue. I tried with an other GPU an old 1080 I had lying around. couldn't reproduce the issue with the 1080 in one hour. Put back the 3060ti issue happen in 5 minutes. Must be the GPU I guess ? I bought a 6700 xt put that on the mobo run ddu install drivers. Run my benchmark problems and the issue rise in a few minute. Then I thought that may be the current that is the issue. I had a ups on the basement Plug that in run a test. Everything went fine for an hour no crash. That was the issue right ? Just to confirm I stopped the PC and retry it from the wall. Took about 20 minutes but it finally failed. Felt like I finally found my issue. Then had a few blue screen memory related. I reseated the ram but then the issue came back. IDK what to do anymore right now I have basically only the SSD that is still the same as the old system but I have no idea what is happening. I will try to switch the SSD again  and see, but doing that mean that I have no more component from the old system running but I don't have much hope at this point.

 

In the event viewer I see the crash in the system logs but it says that the dump creation failed. I installed a second SSD and tried to set that as a location for the mini dump thinking that if the SSD is locked up may be dumping the info to a second drive would work but no...

 

The only thing I can see at this point is that the CPU's being older might have issues handling Gen 4 GPU's but even setting the x16 port to gen 3 didn't help.

 

Let me know what you guys think.

Wait! Just incase I missed something. Did you just check everything else but your RAM?

[alexdpars]

I switched the ram with the mobo and CPU. I had a 64gb kit on my server's mobo. I also tested the gtx 1080 again tonight and it crashes even faster... I also realized that apparently the ryzen 2700x doesn't support the gen 4 pcie so it was running in gen 3 all along. IDK what is happening anymore. The system is completely different right now and I still have the issue. I also realized that I installed the ram wrong at first. Installed it right and still have the issue.

[alexdpars]

BTW what can caused that : Dump file creation failed due to error during dump creation. ? Ram, SSD ,CPU ? 

[alexdpars]

I just installed ubuntu and ran some benchmarks on there. Took a little while but still crash after some time... I tried to get useful information from the logs but seems like the logs get erased on reboot ?

[alexdpars]

Somehow clearing the CMOS made the system way more stable... I've been able to make it crash but had to open like 15 youtube videos while benchmarking extracting files and downloading games from steam. I'll swap the motherboard today when I'll receive it and hope for the best.

[Slizzo]

What happens when you force the PCI-E mode to 3.0?

[alexdpars]

I tried that didn't change anything.

I actually found the problem. Sitting in front of my PC in pieces I was wondering... What haven't I tried to change yet ? Two things popped into my mind. PSU cable and extension cables. I tried changing the cable, same issue. Then I removed the extension sleeved cables from it and boom no issue anymore it's been a few hours and my PC is rock solid. I rebuilt it in its case and seems to work fine.

 

Thanks for everyone's help.

[QuantumSingularity]

Still no MemTest??? Do this and do at least 8 passes or 2 rounds of 4 passes if you use the free version. Had similar issues recently with a system i've built about a year ago, turned out to be the memory. The first 4 passes were ok and the verdict was PASS, but the next few were a FAIL. Replaced the memory and everything is fine, but if in your case that doesn't solve the issues, it's probably the memory controller that's gone bye bye.

[alexdpars]

I did a mem test with my servers ram now I've switched everything back to my setup. I did a benchmark all night and was fine I will do a mem test tonight that will be the final test if it passes it it will mean my PC is now 100% stable