PC very slow to boot and start windows, random crashes

[liberteus]

So, that's a follow-up from the situation described here: 

Symptoms:

- Got crashes during high GPU load (AI stuff or gaming)

- One more crash led to slow boot times (see below for more)

 

What I tried:

• removed all mods/reshade from games
• reverted GPU oc to factory
• reinstalled latest nvidia drivers with a DDU
• launched furmark to see if the same problem occured, let it run for 15mn no trouble at all
• did a 4hrs memtest (booting from USB key) which gave a pass (no errors)
• used OCCT to stress test power for 1hr without issue
• used OCCT to test everything GPU related (VRAM, 3D Standard, 3D adapative) with instructions to use up to 100% of VRAM of GPU cores
• used OCCT to stress test CPU for an 1hr without issue
• used prime95 to stress test CPU (it's ongoing as I'm writing, 15mn into it, going well with CPU at 78C max, no sign of trouble)
• using vram test mentionned in first reply below (5mn test passed, going for longer 
• reseated the drives (multiple times)
• reseated the GPU (multiple times)
• updated bios to latest

reseating hardware "seems" to fix it temporarily. 

 

What I think it may be:

- faulty motherboard (i can replace it although it's am4 itx, but I can grab a new one if needed)

- defective GPU (i'm planning to replace it because 10gb is not enough these days)

 

Here's my config:

• win 10 pro  / latest nvidia drivers
• 3080 10gb dell oem
• 32gb ram
• 2 nvme drives (they're alright according to SMART data)
• gigabyte ITX motherboard X570
• cpu: 5700x3d + 240mm AIO
• case is a NR200
• PSU: 750W corsair SFX
• 3 screens attached, 3440x1440 and a 2560*1440 as well

 

More info:

Basically at the time I reseated the drives and everything was smooth for a while. I bought a new drive since then, properly installed it (replaced one of the 2 nvme basically). And I got new crashes now and then, always same error (see previous post, it's WHCEA uncorrectable error or something).

I thought it may be something related to PCIe lanes (because last times reseating the drives temporarily fixed it), so I reseated not only the drives but also GPU (3080 10gb). And all went smooth for like... a few week. Then got some more crashes. So, anyways, after one more crash now the PC boots very slowly (like, it was 10s before, now it's like... 40?). And it's a consistent behaviour: it takes forever to boot now. Other than that I dont detect any other symptom, once in windows everything is just as it should be and I can play.

 

[cat_B0T_Gaming]

3 minutes ago, liberteus said:

So, that's a follow-up from the situation described here: 

Symptoms:

- Got crashes during high GPU load (AI stuff or gaming)

- One more crash led to slow boot times (see below for more)

 

What I tried:

• removed all mods/reshade from games
• reverted GPU oc to factory
• reinstalled latest nvidia drivers with a DDU
• launched furmark to see if the same problem occured, let it run for 15mn no trouble at all
• did a 4hrs memtest (booting from USB key) which gave a pass (no errors)
• used OCCT to stress test power for 1hr without issue
• used OCCT to test everything GPU related (VRAM, 3D Standard, 3D adapative) with instructions to use up to 100% of VRAM of GPU cores
• used OCCT to stress test CPU for an 1hr without issue
• used prime95 to stress test CPU (it's ongoing as I'm writing, 15mn into it, going well with CPU at 78C max, no sign of trouble)
• using vram test mentionned in first reply below (5mn test passed, going for longer 
• reseated the drives (multiple times)
• reseated the GPU (multiple times)
• updated bios to latest

reseating hardware "seems" to fix it temporarily. 

 

What I think it may be:

- faulty motherboard (i can replace it although it's am4 itx, but I can grab a new one if needed)

- defective GPU (i'm planning to replace it because 10gb is not enough these days)

 

Here's my config:

• win 10 pro  / latest nvidia drivers
• 3080 10gb dell oem
• 32gb ram
• 2 nvme drives (they're alright according to SMART data)
• gigabyte ITX motherboard X570
• cpu: 5700x3d + 240mm AIO
• case is a NR200
• PSU: 750W corsair SFX
• 3 screens attached, 3440x1440 and a 2560*1440 as well

 

More info:

Basically at the time I reseated the drives and everything was smooth for a while. I bought a new drive since then, properly installed it (replaced one of the 2 nvme basically). And I got new crashes now and then, always same error (see previous post, it's WHCEA uncorrectable error or something).

I thought it may be something related to PCIe lanes (because last times reseating the drives temporarily fixed it), so I reseated not only the drives but also GPU (3080 10gb). And all went smooth for like... a few week. Then got some more crashes. So, anyways, after one more crash now the PC boots very slowly (like, it was 10s before, now it's like... 40?). And it's a consistent behaviour: it takes forever to boot now. Other than that I dont detect any other symptom, once in windows everything is just as it should be and I can play.

 

Try reducing to one monitor, go into BIOS and turn off PBO enhancement mode if it is on, maybe undervolt your GPU.

[liberteus]

1 minute ago, cat_B0T_Gaming said:

Try reducing to one monitor, go into BIOS and turn off PBO enhancement mode if it is on, maybe undervolt your GPU.

Thanks, will try and post an update.

[liberteus]

41 minutes ago, liberteus said:

Thanks, will try and post an update.

My gpu is already undervolted using afterburner. Doesnt go over 975mv IIRC, clocked at 2010mhz. Going into bios to turn off PBO. How is that related to boot time though?

[strange13930]

1 hour ago, liberteus said:

So, that's a follow-up from the situation described here: 

Symptoms:

- Got crashes during high GPU load (AI stuff or gaming)

- One more crash led to slow boot times (see below for more)

 

What I tried:

• removed all mods/reshade from games
• reverted GPU oc to factory
• reinstalled latest nvidia drivers with a DDU
• launched furmark to see if the same problem occured, let it run for 15mn no trouble at all
• did a 4hrs memtest (booting from USB key) which gave a pass (no errors)
• used OCCT to stress test power for 1hr without issue
• used OCCT to test everything GPU related (VRAM, 3D Standard, 3D adapative) with instructions to use up to 100% of VRAM of GPU cores
• used OCCT to stress test CPU for an 1hr without issue
• used prime95 to stress test CPU (it's ongoing as I'm writing, 15mn into it, going well with CPU at 78C max, no sign of trouble)
• using vram test mentionned in first reply below (5mn test passed, going for longer 
• reseated the drives (multiple times)
• reseated the GPU (multiple times)
• updated bios to latest

reseating hardware "seems" to fix it temporarily. 

 

What I think it may be:

- faulty motherboard (i can replace it although it's am4 itx, but I can grab a new one if needed)

- defective GPU (i'm planning to replace it because 10gb is not enough these days)

 

Here's my config:

• win 10 pro  / latest nvidia drivers
• 3080 10gb dell oem
• 32gb ram
• 2 nvme drives (they're alright according to SMART data)
• gigabyte ITX motherboard X570
• cpu: 5700x3d + 240mm AIO
• case is a NR200
• PSU: 750W corsair SFX
• 3 screens attached, 3440x1440 and a 2560*1440 as well

 

More info:

Basically at the time I reseated the drives and everything was smooth for a while. I bought a new drive since then, properly installed it (replaced one of the 2 nvme basically). And I got new crashes now and then, always same error (see previous post, it's WHCEA uncorrectable error or something).

I thought it may be something related to PCIe lanes (because last times reseating the drives temporarily fixed it), so I reseated not only the drives but also GPU (3080 10gb). And all went smooth for like... a few week. Then got some more crashes. So, anyways, after one more crash now the PC boots very slowly (like, it was 10s before, now it's like... 40?). And it's a consistent behaviour: it takes forever to boot now. Other than that I dont detect any other symptom, once in windows everything is just as it should be and I can play.

 

is your ram running in xmp or expo? if so you could try turning it off

[jmwhite33]

@liberteus

Being a tower case, how much weight between cpu cooler and GPU do you think the MOBO is contending with?

I've seen situations where microcracks develop in the board and at first only cause issues as everything heats up and expands causing a incomplete circuit, but overtime it gets really bad, pretty much like you are experiencing.

I'm gonna suggest something that may sound weird, lay the case on it's side so MOBO is horizontal and cards are vertical and see if any difference. A long shot I know. I think folks often don't consider the strain the weight puts on the components when troubleshooting.

 

 

 

[cat_B0T_Gaming]

3 hours ago, liberteus said:

My gpu is already undervolted using afterburner. Doesnt go over 975mv IIRC, clocked at 2010mhz. Going into bios to turn off PBO. How is that related to boot time though?

It's not, it might fix the random crashes tho

[liberteus]

9 hours ago, jmwhite33 said:

@liberteus

Being a tower case, how much weight between cpu cooler and GPU do you think the MOBO is contending with?

I've seen situations where microcracks develop in the board and at first only cause issues as everything heats up and expands causing a incomplete circuit, but overtime it gets really bad, pretty much like you are experiencing.

I'm gonna suggest something that may sound weird, lay the case on it's side so MOBO is horizontal and cards are vertical and see if any difference. A long shot I know. I think folks often don't consider the strain the weight puts on the components when troubleshooting.

CPU cooler weighs nothing, it's an AIO. 3080 isnt the worst offender, it's a short 300mm card (DELL OEM one). If the mobo is broken, I'll just buy another one. rather spend 300CAD than waste good cpu/ram etc. I may try that.

[liberteus]

6 hours ago, cat_B0T_Gaming said:

It's not, it might fix the random crashes tho

I removed PBO and the card is always undervolted.

[liberteus]

9 hours ago, strange13930 said:

is your ram running in xmp or expo? if so you could try turning it off

Turning that off now. It,s going to be slower but heh. If I can pinpoint where the problem is, I can spend money and fix it. RAM is cheap, I wouldnt care changing it

[liberteus]

Update:

turned off pbo, xmp, gpu is undervolted (been the case forever though)

 

So far, no crashes but sometimes I get no crashes for days and then it starts acting up again. I noticed that computer is also very slow to shutdown. I see absolutely no difference launching applications, or in games. None whatsoever. So strange. And the bios itself seems very slow.

[strange13930]

9 hours ago, liberteus said:

Update:

turned off pbo, xmp, gpu is undervolted (been the case forever though)

 

So far, no crashes but sometimes I get no crashes for days and then it starts acting up again. I noticed that computer is also very slow to shutdown. I see absolutely no difference launching applications, or in games. None whatsoever. So strange. And the bios itself seems very slow.

you could update your chipset drivers.

[Mark Kaine]

21 hours ago, liberteus said:

the card is always undervolted.

I kinda think that's your issue, did you think about running the whole system at *default* settings before claiming hardware issue?  Just saying cause it's super common to have such issues when overclocking, undervolting etc.

 

Doesn't mean it can't work but a lot of times it simply doesn't. 

Also this is a super aggressive undervolt too , 2010mhz on 3xxx?

 

Yeah, my card (3070) can do that, in a pretty small number of games...

 

Tldr: try running everything at default (except: maybe power limit the GPU, I understand power draw is an issue on 3080 gpus specifically)

[liberteus]

On 5/13/2025 at 5:01 PM, Mark Kaine said:

I kinda think that's your issue, did you think about running the whole system at *default* settings before claiming hardware issue?  Just saying cause it's super common to have such issues when overclocking, undervolting etc.

 

Doesn't mean it can't work but a lot of times it simply doesn't. 

Also this is a super aggressive undervolt too , 2010mhz on 3xxx?

 

Yeah, my card (3070) can do that, in a pretty small number of games...

 

Tldr: try running everything at default (except: maybe power limit the GPU, I understand power draw is an issue on 3080 gpus specifically)

I wonder why the slow boot though? Seems like things are more stable now without XMP and PBO, wonder why these issues are very recent when I used my system like that for about 4years without trouble 

GPU is 975mv @ 2010mhz, mem is oced 250mhz. It worked flawlessly like with for years without issue, and I play only 2 games anyways that are GPU intensive (7days2die, worst optimized game ever; No Man's Sky).

 

Anyways, since it's rather stable without XMP and PBO and I cant honestly see the difference, I'll keep it as it is for now.

[Mark Kaine]

17 minutes ago, liberteus said:

I wonder why the slow boot though? Seems like things are more stable now without XMP and PBO, wonder why these issues are very recent when I used my system like that for about 4years without trouble 

GPU is 975mv @ 2010mhz, mem is oced 250mhz. It worked flawlessly like with for years without issue, and I play only 2 games anyways that are GPU intensive (7days2die, worst optimized game ever; No Man's Sky).

 

Anyways, since it's rather stable without XMP and PBO and I cant honestly see the difference, I'll keep it as it is for now.

Might be degrading a bit or maybe a windows or bios update messed something up?

 

Hard to say, I was just pointing out issues like that often come from overclocking etc. so it's often best to use stock settings while troubleshooting something.

 

And yes, you still don't really know what's wrong but at least it's fixed for now.

 

 

Re: slow boot, what's the BIOS time in taskmanager?

[liberteus]

7 minutes ago, Mark Kaine said:

Might be degrading a bit or maybe a windows or bios update messed something up?

 

Hard to say, I was just pointing out issues like that often come from overclocking etc. so it's often best to use stock settings while troubleshooting something.

 

And yes, you still don't really know what's wrong but at least it's fixed for now.

 

 

Re: slow boot, what's the BIOS time in taskmanager?

Symptoms occured without any previous obvious update (i.e. piece of software, CPU, ram, bios...). It looked like hardware failure rather than software. 

And yeah, always reset everything to default will definitely rule out aggressive oc'ing and other issues.