Hard crash

[InVasMani]

Good advice try to ensure it's RAM rather than the GPU or CPU. If it isn't that try to narrow it down between the other two.

 

[BlaezaLite]

I'll try the XMP, but I did not remove the memory when installing the new CPU.

 

[Waldorf]

funny how many here forgot that amd wise, XMP is considered "ocing".
so yeah, lets drop from 3600 down to 2133 (ignoring timings here), and "enjoy" hw/gaming ...

@Blaeza
havent used any memtest in recent years, as they dont pick up on oc faults,
so use something like TM5, or hci to test.

drop CMD to 2, better set xmp profile, but adjust main timings and Vs manually, soc to 1.1v

 

[BlaezaLite]

Every restart is now a guaranteed BSOD by the looks of it. Tried to enable XMP and nothing changed, so manually set it to 3200/1600.

funny how many here forgot that amd wise, XMP is considered "ocing".
so yeah, lets drop from 3600 down to 2133 (ignoring timings here), and "enjoy" hw/gaming ...

@Blaeza
havent used any memtest in recent years, as they dont pick up on oc faults,
so use something like TM5, or hci to test.

drop CMD to 2, better set xmp profile and adjust main timings and Vs manually,
soc to 1.1v

Sorry, it was Testmem 5. I've come over all tired and depressed.

 

[Frick]

Just so you all know, I DID verify the memory overclock was fine with memtest 5, longest cycle. When I got my 6900XT, I saw all these buttons with overclock CPU and GPU on them and I wanted all the performance from them so I let Adrenaline overclock them. Please give me some credit, I'm not completely devoid of all thought...

Memory can be unstable and even defective even if memory diagnotics show the memory to be ok.

 

[Waldorf]

if not on the latest, reset bios, update, and start fresh.

but with those corsair sticks being DR, they usually dont reach same clocks/timings on Ryzen like the SR kits can,
at least for the non CL14/+4000 kits, and usually will need a little more juice.

leave all on auto, switch soc to 1.1, vddp to 900mV, vddg 950mV, vram to 3600/T2@1.35, gear/power down disabled,
so if that is stable.
past that, you would want to know what chips the kit has, as they differ in needs (voltages) and "where" they can go past jedec/xmp settings (clocks vs timings),
as sometimes they might not go much past 3200, but can do much lower timings (than expected), negating the perf "hit" of lower clocks (3200 vs 3600),
especially since amd (used to) prefer timings (over clock gains).

 

[BlaezaLite]

So, all I've done is reduce the frequency to 3200MHz and flck to 1600. No crashes since.

 

[wheresmycar]

So, all I've done is reduce the frequency to 3200MHz and flck to 1600. No crashes since.

Glad you got it sorted! Memory OCs can be a real pain in the ASS and a really frustrating one when focusing on the software side of things. For the last few years, if any issues arise with odd crashes I now always first jump to setting the BIOS to default (2166Mhz) and take things from there (iron the mem monkey out!!)

 

[BlaezaLite]

Glad you got it sorted! Memory OCs can be a real pain in the ASS and a really frustrating one when focusing on the software side of things. For the last few years, if any issues arise with odd crashes I now always first jump to setting the BIOS to default (2166Mhz) and take things from there (iron the mem monkey out!!)

I'm not sure if it's sorted, but my computer hasn't crashed at all today. Just have 3200 18-21-21-21-42 memory speed.

 

[wheresmycar]

I'm not sure if it's sorted, but my computer hasn't crashed at all today. Just have 3200 18-21-21-21-42 memory speed.

I guess we'll have to wait and see

I'm just curious when you had it @ 3600Mhz.. was this is predefined XMP profile or manual OC? Was fclk set to 1800? Voltage?

EDIT:

Just checked your boards mem support spec and 3600Mhz isn't supported (XMP profiles) but it does support 3466Mhz.

 

[BlaezaLite]

I guess we'll have to wait and see

I'm just curious when you had it @ 3600Mhz.. was this is predefined XMP profile or manual OC? Was fclk set to 1800? Voltage?

No it wasn't XMP. A member on here @The King and another guy both basically gave me instructions bit by bit. The fclk was at 1800 and dram was 1.4v. Originally I was at 3800/1900 but I kept getting whea errors, so I lowered it to 3600/1800. I'm thinking that MW2 was at fault and not the memory, but again I don't know enough to be able to make that judgement.

Edit
Surely the supported memory would be the same as gen 3 ryzen? At a guess...

 

[InVasMani]

If you do crash try raising tRCD values by 1 on each of both and possibly tRP also you could drop tREFI to 13770, but probably start with trying the others and try tREFI as a fall back if the other things don't work. Those things should relax the memory kit further a bit if it's mostly stable. Depending on how stable it is your best solution is probably raising tCL by 1 or reducing the memory frequency a bit slowly until the problem goes away.

 

[BlaezaLite]

If you do crash try raising tRCD values by 1 on each of both and possibly tRP also you could drop tREFI to 13770, but probably start with trying the others and try tREFI as a fall back if the other things don't work. Those things should relax the memory kit further a bit if it's mostly stable. Depending on how stable it is your best solution is probably raising tCL by 1 or reducing the memory frequency a bit slowly until the problem goes away.

I think it was stable... MW2, not so much... I enjoyed multiplayer so much too

This is the problem, it happened once and I panicked thinking my pc was going to die or something ridiculous. I don't mean to but my mental disorders make me this way.

 

[Solaris17]

I don't know enough to be able to make that judgement

your error log indicates volmgr (Volume Manager) dying before the system died. To me that would point to RAM or storage since they are pretty tied to one another. Given that you gave no indication of a disk issue my assumption would have been to look at anything else memory related (RAM) because instability of the volume management service would affect the entire system and generally a culmination of stacking problems.

Thats how I read that anyway. You may even be able to pinpoint it even more looking further, this is what I meant by errorlog context. We already knew the system went down right? So finding that error means nothing. What came before paints the picture. A classic case of cause and effect.

 

[BlaezaLite]

your error log indicates volmgr (Volume Manager) dying before the system died. To me that would point to RAM or storage since they are pretty tied to one another. Given that you gave no indication of a disk issue my assumption would have been to look at anything else memory related (RAM) because instability of the volume management service would affect the entire system and generally a culmination of stacking problems.

Thats how I read that anyway. You may even be able to pinpoint it even more looking further, this is what I meant by errorlog context. We already knew the system went down right? So finding that error means nothing. What came before paints the picture. A classic case of cause and effect.

Is there a way to copy the error log, then I can just post it on here so you can all have a look?

 

[Solaris17]

Is there a way to copy the error log, then I can just post it on here so you can all have a look?

Tbh im too lazy to look and I’m sure unless it was formatted and put in spoilers everyone would prefer you didn’t.

In this case though it doesn’t matter. that was just an example of how I would have approached that in the hopes that it provided insight so you could do it yourself one day.

Im with everyone else and came to the same conclusion I think it’s your ram.

If it’s still doing it; even at stock. I would then assume that volmgr has crashed first everytime and that at this point there is a good chance data is corrupt. If it was still happening at stock, I would then verify game data. Run a check disk on the drive to look for file system problems then try to play again.

 

[InVasMani]

Having Steam if you bought the game on that service validate the integrity of the game files never hurts either on the corruption subject.

 

[BlaezaLite]

I'm on battle.net.

 

[Frick]

How many RAM modules do you have? I had all sorts of random BSODs, but since running only one stick downclocked it has been stable.

 

[BlaezaLite]

I have 2x16GB Corsair Vengeance. I thought they looked nice, I now know that they aren't very good, lol.

I've just found this, because I have a smooth brain...

 

[The King]

I have 2x16GB Corsair Vengeance. I thought they looked nice, I now know that they aren't very good, lol.

I've just found this, because I have a smooth brain...
View attachment 275146

WHEA 18 in this case seems to be CPU related or CPU V-core.

Try disabling PBO in the BIOS. Although on AUTO it is supposed to be stable.

Are you running any boost override?

The only time I have seen WHEA 18 on my system is with an unstable Boost override, (CO) Curve optimizer or too high CPU OC.

 

[BlaezaLite]

WHEA 18 in this case seems to be CPU related or CPU V-core.

Try disabling PBO in the BIOS. Although on AUTO it is supposed to be stable.

Are you running any boost override?

The only time I have seen WHEA 18 on my system is with an unstable Boost override, (CO) Curve optimizer or too high CPU OC.

I've done that as well. Disabled Auto OC in Adrenaline. Just did the Antta777 in testmem 5 no errors at these settings...

Yes, I know it's at 3800, I've got pissed off with having to run at slow speeds, so thought I'd try this and it seems happy!

 

[The King]

If you see more WHEA 18s then you know its not memory related.

Adrenaline Auto OC, will only OC the GPU only? Have not used that so not sure.
If that also OC the CPU somehow then you may have found something there.

Something is causing the WHEA 18 so you still need to find out what is causing that.

 

[Mussels]

So I was playing multiplayer MW2 and then it crashed and reset my pc. As in back to before the bios screen like I'd just turned it on. None of my RGB turned off though. Any thoughts? Also whats the command to see all your whea errors?

You've got ryzen with LPX memory, so this is a well known black screen crash as the SoC resets and takes the system with it


Raise DRAM to 1.4v, SoC to 1.15V and you should be fine

Setting RAM timings to even numbers will fix any remaining issues, so if its 15-17-19-21 you'd add +1 and do 16-18-20-22
Gear down mode is meant to fix that, but it doesnt work below 2667 and doesn't solve issues with your infinity fabric clocking too high (the SoC voltage solves that side of it)

LPX ram had issues because of those two combining, low clocked LPX didnt trigger GDM and high clocked needed the higher voltages
GDM runs the odd timings but every second clock, it's often referred to as as being in-between the odd and even values in speed so that 15 becomes 15.5


edit: yeah seeing the latest post, 3800 and odd timings is the key
Fix the timings and check the voltages


VSOC to 1.15v
18-22-22-42 (you can tune those later if possible, most hynix ram clocks high but wont tighten timings)


The key here is that the crashes with SoC voltages happen at low loads and idle, so it always passes stress tests

 

[BlaezaLite]

If you see more WHEA 18s then you know its not memory related.

Adrenaline Auto OC, will only OC the GPU only? Have not used that so not sure.
If that also OC the CPU somehow then you may have found something there.

Something is causing the WHEA 18 so you still need to find out what is causing that.

Yeah, I will keep checking.

Adrenaline lets you oc the CPU and GPU separately. You should give it a whirl.

If I don't get the whea errors I know that it was caused by being oc'd, I think.

You've got ryzen with LPX memory, so this is a well known black screen crash as the SoC resets and takes the system with it


Raise DRAM to 1.4v, SoC to 1.15V and you should be fine

Setting RAM timings to even numbers will fix any remaining issues, so if its 15-17-19-21 you'd add +1 and do 16-18-20-22
Gear down mode is meant to fix that, but it doesnt work below 2667 and doesn't solve issues with your infinity fabric clocking too high (the SoC voltage solves that side of it)

LPX ram had issues because of those two combining, low clocked LPX didnt trigger GDM and high clocked needed the higher voltages

Dram is at 1.4v as per the The King's instructions and I don't know what the SoC is at, but it'll be 1.15 very shortly. Again, thanks to everyone that has helped me.