BLUE SCREEN OF DEATH (HAPPENING QUITE FREQUENTLY NOW) [VIDEO MEMORY MANAGEMENT INTERNAL]

[Daily Drew]

Hey guys, been a lot through a lot of different customer supports lately for this but here goes nothing.
 

I am having an issue with a blue screen VIDEO_MEMORY_MANAGEMENT_INTERNAL Troubleshooting so far: - reseating ram/cpu/gpu - fresh reinstall windows/bios - reinstalling drivers for MB and CPU/GPU- Installed windows 11, thinking it was a Windows 10 bug but no same thing. - did a MEM Test with that program to check if the ram is all good and it did not find any issues. It started happening while starting a render but now it happens just out of nowhere at anytime. (I use mt machine only for work, edits in premiere/ae/blender etc.  I had a deep dive with NVIDIA customer support and we found a debug code: LOG FILE ISSUE WE FOUND SO FAR: nt ID 0 from source nvlddmkm cannot be found. PC SPECS: (ONLY THING OVERCLOKCED IS THE RAM TO IT'S 36MHZ POTENTIAL)  WINDOWS 11 MB GIGABYTE AORUS MASTER X570  BIOS: F31O RAM CORSAIR 128 GB DDR4 GPU NVIDIA RTX 3070 CPU RYZEN 9 5950X PSU CORSAIR 750W 1TB GIGABYTE AORUS SSD 8TB HDD SEAGATE 4TB HDD WD  EXTERNAL HDD 4TB WD (3 OF THOSE CONNECTED AT ALL TIMES) 12 TB NAS WD CLOUD ON THE LOCAL NET (IF THAT MATTERS AS INFO)

I have not tried a different GPU because i don't have any on hand
I bought recently a new 1000 W PSU that I have not tested yet. 


I believe LINUS said something about the GPU slowly dying if this happens, but i got my gpu brand new from a site and never used it for mining.

If somebody has any clue that would be very helpful.
Thank you all ! 
 

[Radium_Angel]

Upload the Windows crash dumps here (typically in c:\windows\minidumps, but not always) and I can deconstruct them and tell you what's going on.

@ myname when responding so I know you've posted.

 

And when troubleshooting, turn off XMP and ALL overclocking, until we get to the root of the issue.

[Daily Drew]

@Radium_Angel Hey thank you in advance for replying! 

Attached to this message is all the minidumps in the folder. 
Thank you once again! 

110922-12921-01.dmp110822-14000-01.dmp103122-13187-01.dmp110422-13750-01.dmp
110722-15046-01.dmp

[Quickstrike]

51 minutes ago, Daily Drew said:

 

PC SPECS: (ONLY THING OVERCLOKCED IS THE RAM TO IT'S 36MHZ POTENTIAL)  WINDOWS 11 MB GIGABYTE AORUS MASTER X570  BIOS: F31O RAM CORSAIR 128 GB DDR4 GPU NVIDIA RTX 3070 CPU RYZEN 9 5950X PSU CORSAIR 750W

I bought recently a new 1000 W PSU that I have not tested yet. 

 

How many sticks Of RAM and in what slots.

-> ONLY THING OVERCLOKCED IS THE RAM TO IT'S 36MHZ POTENTIAL
You mean 3600MHZ?
Is that the DOCP settings you activated or are you actually overclocking?

Also how are you cooling that monster? Case plus cpu cooler?, how many fans?

FInally, you are cutting it close with a 750W
3070 draws 220 to 250
100 W for mobo and overhead, 

142 W for CPU
18 W for 2 disks drives 
5 W per fan...
Is there a pump in there? (up to 50W)
RGB?

How is you GPU connected to the PSU? Single cable or two cables?

You might be crashing because of insufficient power delivery

 

[Radium_Angel]

11 minutes ago, Daily Drew said:

@Radium_Angel Hey thank you in advance for replying! 

Attached to this message is all the minidumps in the folder. 
Thank you once again! 

110922-12921-01.dmp110822-14000-01.dmp103122-13187-01.dmp110422-13750-01.dmp
110722-15046-01.dmp

I'll check 'em out, give me a few. @Quickstrike brings up some excellent points. What are the CPU/GPU temps at idle and under load? (CoreTemp and GPUTemp are two good small programs that can help with that)

 

[Daily Drew]

11 minutes ago, Quickstrike said:

How many sticks Of RAM and in what slots.

-> ONLY THING OVERCLOKCED IS THE RAM TO IT'S 36MHZ POTENTIAL
You mean 3600MHZ?
Is that the DOCP settings you activated or are you actually overclocking?

Also how are you cooling that monster? Case plus cpu cooler?, how many fans?

FInally, you are cutting it close with a 750W
3070 draws 220 to 250
100 W for mobo and overhead, 

142 W for CPU
18 W for 2 disks drives 
5 W per fan...
Is there a pump in there? (up to 50W)
RGB?

How is you GPU connected to the PSU? Single cable or two cables?

You might be crashing because of insufficient power delivery

 

I have only the XMP profile set to auto (if this is called ram overclocking)
But i've turned it off for a while to test if that was the issue but it still persisted.
I have 4 sticks (32gigs each) of this exact ram 
No rgb - a tiny little on the Motherboard
6 fans including the be quiet CPU fan, no water pumps

Single cable GPU

Right now 

[Radium_Angel]

@Daily Drew Ok, I'm going to post the important parts of the crash files, and then my thoughts from there.

 

__

DRIVER_IRQL_NOT_LESS_OR_EQUAL (d1)
An attempt was made to access a pageable (or completely invalid) address at an
interrupt request level (IRQL) that is too high.  This is usually
caused by drivers using improper addresses.
If kernel debugger is available get stack backtrace.
Arguments:
Arg1: ffffa68144370858, memory referenced
Arg2: 0000000000000002, IRQL
Arg3: 0000000000000001, value 0 = read operation, 1 = write operation
Arg4: fffff805635d62b3, address which referenced memory
 

DUMP_FILE_ATTRIBUTES: 0x1008
  Kernel Generated Triage Dump

BUGCHECK_CODE:  d1

BUGCHECK_P1: ffffa68144370858

BUGCHECK_P2: 2

BUGCHECK_P3: 1

BUGCHECK_P4: fffff805635d62b3
 

SYMBOL_NAME:  nvlddmkm+8d62b3

MODULE_NAME: nvlddmkm

IMAGE_NAME:  nvlddmkm.sys

STACK_COMMAND:  .thread ; .cxr ; kb

BUCKET_ID_FUNC_OFFSET:  8d62b3

FAILURE_BUCKET_ID:  AV_nvlddmkm!unknown_function

OSPLATFORM_TYPE:  x64

OSNAME:  Windows 10

FAILURE_ID_HASH:  {7eea5677-f68d-2154-717e-887e07e55cd3}
___

2nd crash dump is identical

___

VIDEO_MEMORY_MANAGEMENT_INTERNAL (10e)
The video memory manager encountered a condition that it can't recover from. By crashing,
the video memory manager is attempting to get enough information into the minidump such that
somebody can pinpoint what lead to this condition.
Arguments:
Arg1: 0000000000000017, Unexpected system command failure.

 

DUMP_FILE_ATTRIBUTES: 0x1008
  Kernel Generated Triage Dump

BUGCHECK_CODE:  10e

BUGCHECK_P1: 17

BUGCHECK_P2: ffffffffc000000d

BUGCHECK_P3: ffffaf8a85c49000

BUGCHECK_P4: 0
 

SYMBOL_NAME:  dxgmms2!VIDMM_WORKER_THREAD::FlushEvictQueue+12a

MODULE_NAME: dxgmms2

IMAGE_NAME:  dxgmms2.sys

IMAGE_VERSION:  10.0.22621.675

STACK_COMMAND:  .thread ; .cxr ; kb

BUCKET_ID_FUNC_OFFSET:  12a

FAILURE_BUCKET_ID:  0x10e_17_dxgmms2!VIDMM_WORKER_THREAD::FlushEvictQueue

OSPLATFORM_TYPE:  x64

OSNAME:  Windows 10

FAILURE_ID_HASH:  {8b70aac8-c9b8-15d4-254c-a3e0828fa53b}

Followup:     MachineOwner
 

___

4th crash dump same as the above one

___

5th crash dump same as the 3rd one

 

______

I"m with @Quickstrike on this one. Either your system can't provide the power needed, or it's overheating (can the motherboard handle that much RAM BTW?)

[Quickstrike]

OK,
My two main hypothesis
#1- Your kits are mismatched (Yes they are labelled the same but one could be a lower binned) and unstable (try running only 2 dimms). I had the exact some issue with two Trident-Z 16 gig kits (didn't crash but wouldn't oc)

#2- You are at 630W draw on a 750W, over the 80% mark and your PSU is not outputting enough juice, hence bleu screen

 

[Quickstrike]

7 minutes ago, Radium_Angel said:

______

I"m with @Quickstrike on this one. Either your system can't provide the power needed, or it's overheating (can the motherboard handle that much RAM BTW?)

Thanks,
Good job parsing the logs.

BTW X570 can handle 128 gig at 4400mhz

[Daily Drew]

First of all, I just want to say how crazy thankful I am for this forum and for you guys.
1 Month in threads emailing different manufactures even windows lol, has not got me as far as you did that fast!

I've run a bit of cinebench before and the pc was not going over 85-ish ( I was monitoring it on APP center basic gigabyte app)
I've recently even re-pasted my cpu, because i went trough that troubleshooting as well. 

You may see attached idle temps right now.

I've never thought about the memory capacity on the Mobo but I've just checked the site so it's good there:
 

1. 4 x DDR4 DIMM sockets supporting up to 128 GB (32 GB single DIMM capacity) of system memory
   
   I will change the PSU with my new 1000W Seasonic and let you know how it goes.
    

[Daily Drew]

2 minutes ago, Quickstrike said:

Thanks,
Good job parsing the logs.

BTW X570 can handle 128 gig at 4400mhz

Triple checked that and you're correct, the 2 packs of 32 gigs even though exactly the same they have different serial numbers for each pack.

[Daily Drew]

Hey @Quickstrike @Radium_Angel I want to let you know that since I've change the PSU from  750W to 1000W I have not had any blue screens since the last 48H.

Big THANK YOU to you guys and this community! 

I will further update this thread if there is any other situations in the future.
Here's a small benchmark I did with the the two programs you offered on top. (Which shows that I need a new cooling solution) 

[Daily Drew]

Hey guys, just an update on the card, it turns out it was a bad unit, and send it for RMA and NVIDIA was kind enough to send a new one.