Hi,
I recently built a brand new gaming computer and have had nothing but constant system crashes. At first it would happen randomly form the desktop with only maybe Chrome open. Eventually it stopped crashing from the desktop and now only crashes during any type of GPU activity like running Heaven, a game, or trying to stress test with a scanner like Precision X1. Crashes are always the same, both screens go black, audio stays up for up to 20 seconds, then the system restarts. 
 
Event Viewer has given me all of the following reports:
1:
The computer has rebooted from a bugcheck.  The bugcheck was: 0x00000116 (0xffff8b035a636460, 0xfffff80760827b14, 0xffffffffc000009a, 0x0000000000000004). A dump was saved in: C:\Windows\MEMORY.DMP. Report Id: 0718ba3e-d5d8-4f51-8c09-6e26b45f697d.
 
2:
The description for Event ID 14 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.
If the event originated on another computer, the display information had to be saved with the event.
The following information was included with the event:
\Device\Video3
0000(0000) 00000000 00000000
The message resource is present but the message was not found in the message table
 
3:
The description for Event ID 14 from source nvlddmkm cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.
If the event originated on another computer, the display information had to be saved with the event.
The following information was included with the event:
\Device\Video5
0000(0000) 00000000 00000000
The message resource is present but the message was not found in the message table
 
Things I have tried so far:

• Remove all OC software EVGA Precision, MSI Afterburner, Intel XTU (currently none are installed)
• Remove all RBG apps like MSI Center, Lian Li Connect, ICUE (currently none are installed)
• Remove Overclocks (was OCing CPU from XTU and sometimes from the BIOS, have tried with and without OC)
• Remove XMP from BIOS (tried with and without XMP)
• DDU and Clean Install of 471.68 (tried this multiple times, including without being connected to the internet)
• One version of this includes the following instructions: (note this didn't work for me, I couldn't find the nvlddmkm.sys file in the "drivers" folder, it is located elsewhere and doesn't allow me to modify it)
  1. On the search bar, type in Device Manager and open it.
  2. Look for Display adapters and expand the section using the drop-down arrow.
  3. Right-click on your GPU and click on Uninstall.
  4. In the confirmation window, be sure to tick the Delete the driver software for this device checkbox, then click OK.
  5. After uninstall is complete, download the latest Nvidia drivers for your GPU from . Install these drivers and reboot the computer.
  6. When your computer boots, navigate to C:\Windows\System32\Drivers and rename the nvlddmkm.sys file to nvlddmkm.sys.old 
  7. After this, open the Nvidia directory, usually found at C:\ProgramFiles\Nvidia, and look for a file named nvlddmkm.sys_ and copy it to the System32 folder.
  8. Now open the Command Prompt as administrator and type in the following: EXPAND.exe nvlddmkm.sy_ nvlddmkm.sys
  9. After the expansion is complete, copy and paste the newly created nvlddmkm.sys file to the C:\Windows\System32\Drivers folder.
• Underclock (tried -200 memory and gpu clock plus only 90% power in Precision X1) (underclocking will let me run WoW for maybe 30-60 minutes, but eventually crashes)
• Ran SFC and DISM (SFC detected errors once early in my troubleshooting and fixed them, never detected errors again)
• Ran Windows Memory Diagnostic and Memtest86 (no issues from either)
• Ran CHKDSK
• Ensured Windows was set to Never sleep or shut off display, LSPM set to Off
• Nvidia Control Panel Power Management Mode set to "Prefer Max Performance"
• Multiple Windows re-installs
• Updated BIOS and installed latest Intel/MSI drivers from the website
• Set TDR from 2 to 10 (currently set to 10)
• Of note, I previously had this card with the stock cooler in my old PC and it worked fine. After the new rig got here, I removed the stock cooler and placed the EK kit on it. I was meticulous and followed the instructions to a T to ensure I didn't damage anything.

My build:
CPU: i9-11900k
GPU: EVGA 3090 FTW3 Ultra
RAM: Corsair Vengeance RGB Pro 32GB (4x8) DDR4 3600
MB: MSI MPG Z590 Carbon EK
SSD: Samsung 980 PRO 1TB
OS: Windows 10 Home
PS: Corsair RM850X
 
The system is water cooled with EK gear and I'm also using the LINKUP Vertical GPU Mount with a PCIe 4.0 Riser Cable. I'm also using the 24 pin and 3x8 pin STRIMER Plus cables.
 
I would really appreciate help with this as I feel like I've tried everything at this point and nothing is working. I'm desperate at this point.

I've already tried removing the Strimer Cables that go to the GPU with no luck, does the one that goes to the MB matter at all? Haven't tried taking that one off yet. Unfortunately I don't have the extra tubing on hand to just remove the vertical bracket and put the GPU directly on the MB.
Also, not sure if it mattered, but I had the Strimer Cable hooked up with 2x 6+2 PCIE cables, one of which was a splitter with a second 6+2 connected. I changed that up and there are now 3x dedicated 6+2 PCIE cables hooked into it. Didn't make a difference though.

I guess I'll have to make some new tubing so I can plug right into the motherboard without the riser cable. I also saw that I had the riser connected into PCIE slot #2, so I moved it to #1, but of course that didn't change anything. I'll update sometime this week when I get a chance to re-do the loop and get it set up again.

I'd still say the issue is in the riser mount, cable &/ or  extension cables.
  ..&  Possibly the 850 psu also can't completely feed enough power under spiking amperage draws of these 30xx series cards. 

- it happen to me . same errors and my 3090ftw3 died playing New World (56Hours of play)
- It all start with Losing Display Signal..... only fix would be Hard Reset system (Power button Push).... at the beginning losing signal 1..or 2 times a day......... right whend it dies losing signal 6...7 times a day .... the These ERRORS in LoGs = games crashing
worth mentioning as i was trying to play games..... my games would crash like in 5 mins .... then i try New World an then it Popped (red led of death) = my card dead
had the card for almost a year
 
my specs : 10900k not OC   / Gskill 32gb 4000Mhz CL17 / 2tb SSD 970 evo / PSU Evga 1600w T2  (using originaal cables)
              all watercooled  (gpu active watercooled Backplate = never saw more than 65C on hot spot NOT OC'ed ever)
  saying all these to take out of equation --- inproper Psu power / cables / heat /or OC'ing

tweaver2007
Small update for everyone...
So I moved the riser cable to the correct spot and did some research on PCIE 4.0 and issues with riser cables. It seems like most people who have problems with the riser have straight up Blue/Black SOD issues at startup, like it won't display any image whatsoever until they change a setting in the BIOS. I don't think the riser cable is the issue. Another problem is that I did try to mount the 3090 normally with no riser, but the EK kit on the card is too tall and won't allow me to put the side panel glass back on the case while it is on there.
 
I have a brand new EVGA 1000W PSU waiting for me at home. I'm going to try installing that when I get back in about a week. I did a lot of reading on the draw of the 3090 and i9-11900k plus all the Strimers and RBG fans I have. I think what happens is that as soon as the 3090 ramps up during a game, it is too much for the PSU and everything shuts down. I think I'm maxing out the PSU. If that fails, I'm going to remove all the Strimers again to try and rule that out once more. Failing that...RMA I guess. 
 
Coolmanfu, when you say pigatils, are you referring to the PCIE 8 pins that split from 1 at the PSU into 2 that connect to the GPU? I did try going 3 dedicated instead of 1 + 1 pigtail, but no luck there.
 
I'm praying the 1000W Gold PSU I ordered will be sufficient. I read about other people using a i9-11900k and a 3090 using 1000W PSUs, only 1-2 people used 750 or 850, but I don't think they had as much RBG puke and extras like I do. 
 

Yes that's what I mean by pigtails. In "theory" pigtails deliver enough power for multiple 8 pin connectors on one line but in practice it's just not true and they lead to a ton of problems. For your card you should have 3 separate cables going from PSU to gpu. Also you said you ordered a 1000w gold. I really hope that works for you. The cleaner power delivery of a platinum is "better" but arguably it's impossible to measure the difference between gold and platinum.

I think your ultimate issue is that 850 is just not enough for everything your system has going on. Just out of curiosity so you use an all in one liquid cooler for your cpu?

I had the same issue, turns out it was my icue software controlling my 3090FE RGB lighting. Seems like you've ruled this out so this commend is more for others reading it in case that helps you. Had the exact same error code and upon disabling plugins in icue I've never had another issue. 

It's still worth testing without the riser.  Just don't put the side panel back on: it's not going to kill your computer to run it open for a little while.  That way you'll know for sure.
 
If installing the card directly on the motherboard doesn't help, try DDUing this driver and installing an older one.  Bug check 0x116 means that the GPU driver timed out and Windows failed to reset it.  Given that you already DDU'd once, it's probably not the driver, but it's easy enough to try an older driver just in case it helps.

tweaver2007
I'll make some new tubing to accommodate the card going directly on the MB and pray to god that works. If that doesn't fix it...I guess I'm going to have to RMA this thing. I just have no clue anymore. 

Did you test the card fully before mounting the waterblock?  Not sure about the time involved vs. messing with the loop but it might be worth re-mounting the stock cooler to see how it runs with that and no riser.  Air cooler or AIO on the CPU so you don't have to rebuild the loop.  Lots of reported problems with blocks this gen but I don't follow that closely.

tweaver2007
So I did actually have the card in my old gaming tower with the stock cooler and it worked just fine. I'll test without the riser this week once I can make some new tubes and I'll let you know.

I would also suspect the block.  Do some searching on here.

https://forums.evga.com/30803090-FTW3-waterblock-deaths-m3235633.aspx
 
https://forums.evga.com/Q...080-ftw3-m3233831.aspx
 
https://www.reddit.com/r/...k_waterblockbackplate/
 
Big one:
 
https://forums.evga.com/R...-booting-m3212511.aspx

Have you tried to disable G-Sync completely?

@tweaver207
Just to let you know that this pcie 4.0 riser cable 'Linkup - Ultra pcie 4.0x16' works perfectly with my i9 11900k / 3090 ftw3 ultra, the only difference is that I have a Aorus Master z590 mb.
Here's the link https://www.amazon.co.uk/gp/product/B08NM4W7YQ/ref=ppx_yo_dt_b_asin_title_o06_s00?ie=UTF8&psc=1
I hope you get everything fixed soon.

@kevinc313
Yeah I'm putting the original fan/cooler back on today to see what happens. Someone else posted a 80 page thread saying EK blocks were causing shorts and sometimes bricking 3090s. Hoping that didn't happen with mine.
Edit: oh that was you that shared those posts lol

that might be the rizer cable