I have a new CyberpowerPC with EVGA RTX 3080 card. Under low load, everything is fine. If I run Unigine Heaven 4.0 for around 10 minutes, the machine eventually crashes, even with EVGA fan curve set to "Aggressive". GPU temperature never exceeds 60 deg C before the crash, and CPU temperature also remains at moderate levels. Based on what I'm reading here and, in Tom's Hardware ("Nvidia RTX 3080 crashes caused by capacitors, says EVGA") and elsewhere, these graphics cards are defective in their design.  I've included a crash dump analysis below my signature.  I can get the machine to crash other ways, e.g. by trying to run EVGA VF Curve Tuner (without any overclocking). I have the latest NVidia driver for Windows 10, and latest BIOS for my motherboard (Asus Prime X570-P). Will EVGA be fixing this, either by replacing the hardware or by a driver update?
 
Dave Oshinsky
 
Crash dump analysis for Unigine Heaven 4.0 crash:
 

Microsoft (R) Windows Debugger Version 10.0.19041.685 AMD64
Copyright (c) Microsoft Corporation. All rights reserved.


Loading Dump File [C:\Windows\MEMORY.DMP]
Kernel Bitmap Dump File: Kernel address space is available, User address space may not be available.

Symbol search path is: srv*
Executable search path is:
Windows 10 Kernel Version 19041 MP (24 procs) Free x64
Product: WinNt, suite: TerminalServer SingleUserTS
Built by: 19041.1.amd64fre.vb_release.191206-1406
Machine Name:
Kernel base = 0xfffff806`69800000 PsLoadedModuleList = 0xfffff806`6a42a390
Debug session time: Thu Mar  4 15:50:46.905 2021 (UTC - 5:00)
System Uptime: 0 days 0:00:21.559
Loading Kernel Symbols
...............................................................
......Page 804e74 not present in the dump file. Type ".hh dbgerr004" for details
..........................................................
........................................................
Loading User Symbols
PEB is paged out (Peb.Ldr = 000000f0`25553018).  Type ".hh dbgerr001" for details
Loading unloaded module list
.......
For analysis of this file, run !analyze -v
7: kd> !analyze -v
*******************************************************************************
*                                                                             *
*                        Bugcheck Analysis                                    *
*                                                                             *
*******************************************************************************

WHEA_UNCORRECTABLE_ERROR (124)
A fatal hardware error has occurred. Parameter 1 identifies the type of error
source that reported the error. Parameter 2 holds the address of the
WHEA_ERROR_RECORD structure that describes the error conditon.
Arguments:
Arg1: 0000000000000000, Machine Check Exception
Arg2: ffffc182cfe52028, Address of the WHEA_ERROR_RECORD structure.
Arg3: 00000000bc800800, High order 32-bits of the MCi_STATUS value.
Arg4: 00000000060c0859, Low order 32-bits of the MCi_STATUS value.

Debugging Details:
------------------


KEY_VALUES_STRING: 1

    Key  : Analysis.CPU.Sec
    Value: 1

    Key  : Analysis.DebugAnalysisProvider.CPP
    Value: Create: 8007007e on OSHWIN2021

    Key  : Analysis.DebugData
    Value: CreateObject

    Key  : Analysis.DebugModel
    Value: CreateObject

    Key  : Analysis.Elapsed.Sec
    Value: 1

    Key  : Analysis.Memory.CommitPeak.Mb
    Value: 73

    Key  : Analysis.System
    Value: CreateObject


BUGCHECK_CODE:  124

BUGCHECK_P1: 0

BUGCHECK_P2: ffffc182cfe52028

BUGCHECK_P3: bc800800

BUGCHECK_P4: 60c0859

BLACKBOXBSD: 1 (!blackboxbsd)


BLACKBOXNTFS: 1 (!blackboxntfs)


BLACKBOXWINLOGON: 1

PROCESS_NAME:  YourPhone.exe

STACK_TEXT:  
ffffaa81`63ac6938 fffff806`69cb39aa : 00000000`00000124 00000000`00000000 ffffc182`cfe52028 00000000`bc800800 : nt!KeBugCheckEx
ffffaa81`63ac6940 fffff806`655d15b0 : 00000000`00000000 ffffc182`cfe52028 ffffc182`cc9f7e40 ffffc182`cfe52028 : nt!HalBugCheckSystem+0xca
ffffaa81`63ac6980 fffff806`69db552e : 00000000`00000000 ffffaa81`63ac6a29 ffffc182`cfe52028 ffffc182`cc9f7e40 : PSHED!PshedBugCheckSystem+0x10
ffffaa81`63ac69b0 fffff806`69cb52d1 : ffffc182`d63d8c00 ffffc182`d63d8c00 ffffc182`cc9f7e90 ffffc182`cc9f7e40 : nt!WheaReportHwError+0x46e
ffffaa81`63ac6a90 fffff806`69cb5643 : 00000000`00000007 ffffc182`cc9f7e90 ffffc182`cc9f7e40 00000000`00000007 : nt!HalpMcaReportError+0xb1
ffffaa81`63ac6c00 fffff806`69cb5520 : ffffc182`cc904508 00000000`00000000 ffffaa81`63ac6e00 00000000`00000000 : nt!HalpMceHandlerCore+0xef
ffffaa81`63ac6c50 fffff806`69cb4a65 : ffffc182`cc904508 ffffaa81`63ac6ef0 00000000`00000000 00000000`00000000 : nt!HalpMceHandler+0xe0
ffffaa81`63ac6c90 fffff806`69cb7225 : ffffc182`cc904508 00000000`00000000 00000000`00000000 00000000`00000000 : nt!HalpHandleMachineCheck+0xe9
ffffaa81`63ac6cc0 fffff806`69d0c959 : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : nt!HalHandleMcheck+0x35
ffffaa81`63ac6cf0 fffff806`69c04bfa : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : nt!KiHandleMcheck+0x9
ffffaa81`63ac6d20 fffff806`69c048b7 : 00000000`00000000 00000000`00000000 000000f0`25efe458 00000000`00000000 : nt!KxMcheckAbort+0x7a
ffffaa81`63ac6e60 00007ff8`40d5e239 : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : nt!KiMcheckAbort+0x277
000000f0`25efe310 00000000`00000000 : 00000000`00000000 00000000`00000000 00000000`00000000 00000000`00000000 : 0x00007ff8`40d5e239


MODULE_NAME: AuthenticAMD

IMAGE_NAME:  AuthenticAMD.sys

STACK_COMMAND:  .thread ; .cxr ; kb

FAILURE_BUCKET_ID:  0x124_AuthenticAMD_PROCESSOR_BUS_L1_SRC_IRD_I_NOTIMEOUT

OS_VERSION:  10.0.19041.1

BUILDLAB_STR:  vb_release

OSPLATFORM_TYPE:  x64

OSNAME:  Windows 10

FAILURE_ID_HASH:  {a44f54e5-033f-6603-b3e9-4b506acba551}

Followup:     MachineOwner
---------

Thanks for moving my post to a separate thread so quickly. I have opened a support ticket with CyberpowerPC by emailing them. They have not yet replied. I will update this thread once I have more information.
 
For now, this looks like part of a larger pattern of multiple problems with the EVGA RTX 30 series cards, and maybe more generally with all RTX 30 series cards. I was hoping to derive some insights from those on this forum, who have more experience with these problems than I do. I realize that the drivers are from NVidia, but my current thinking is that this probably a hardware problem. Should I also post on NVidia forums?
 
Dave Oshinsky

Which version of bios is it? 
What PSU do you have? 
Are you plugged into an UPS?
 
EDIT: I would also do a fresh install of the drivers using DDU. There are tutorials online, or you can check out this YT video. 
https://www.youtube.com/watch?v=bP-2B14Nckw
 
 

 
Wait...You list the BIOS as Prime X570-P Verison 3405....but then lost your  motherboard as a Gigabyte X570 UD ATX ...whose current BIOS is
F33c (this is a beta tho, might want to stick with F32 for now) as of Feb 2021 (for Rev 1)....
 
The X570-P Prime is Asus?
 
 
Can you verify parts/bios revs?
 
 
This is the Gigabyte: https://www.amazon.com/Gigabyte-X570-UD-PCIe4-0-Motherboard/dp/B07WL68L6M/ref=sr_1_2?dchild=1&keywords=Gigabyte+X570+UD&qid=1614965098&sr=8-2
This is the Asus:  https://www.amazon.com/Asus-Prime-X570-P-Ryzen-Motherboard/dp/B07SW925DR/ref=sr_1_1?dchild=1&keywords=Asus+X570-P+Prime&qid=1614965136&sr=8-1
 

I guess they substituted ASUS for Gigabyte (same X570 chipset). The BIOS that came with the PC was a version from 08/2020.  I saw crashes under load with both the older BIOS, and the current version 3405. So the latest v3405 didn't make matters worse, but also didn't fix anything.  I've seen a number of web pages devoted to crashes with these cards, e.g. .  I don't know if this kind of hardware issue is what is causing the problems with my system. The driver I'm running is the latest from NVidia (27.21.14.6172 dated 2/23/2021).
 
C:\Windows>wmic baseboard
Caption     ConfigOptions                                                             CreationClassName  Depth  Description  Height  HostingBoard  HotSwappable  InstallDate  Manufacturer           Model  Name        OtherIdentifyingInfo  PartNumber  PoweredOn  Product       Removable  Replaceable  RequirementsDescription  RequiresDaughterBoard  SerialNumber     SKU  SlotLayout  SpecialRequirements  Status  Tag         Version   Weight  Width
Base Board  {"Default string", "Default string", "Default string", "Default string"}  Win32_BaseBoard           Base Board           TRUE          FALSE                      ASUSTeK COMPUTER INC.         Base Board                                    TRUE       PRIME X570-P  FALSE      TRUE                                  FALSE                  200873094000338                                        OK      Base Board  Rev X.0x


C:\Windows>

ya know, I do have to ask
 
What version of the AMD chipset driver do you have? You might want to try updating that
 
https://www.amd.com/en/support/chipsets/amd-socket-am4/x570
 

My guess - Cyberpower has the drivers all borked up ... from AMD chipset / AGESA / &  board BIOS 
plus probably a dozen other settings  in Windows & the Nvidia drivers too. 

AMD Chipset driver amd_chipset_software_2.13.27.501 install is failing:
MSI (c) (B0:E0) [14:37:55:089]: Note: 1: 1729
MSI (c) (B0:E0) [14:37:55:089]: Product: AMD_Chipset_Drivers -- Configuration failed.
MSI (c) (B0:E0) [14:37:55:089]: Windows Installer reconfigured the product. Product Name: AMD_Chipset_Drivers. Product Version: 2.13.27.501. Product Language: 1033. Manufacturer: Advanced Micro Devices, Inc.. Reconfiguration success or error status: 1603.


When I look in Windows device manager, every AMD-related driver I have checked is up-to-date when I check for upgrades. I'm thinking that Windows has configured the AMD-related drivers properly. AMD chipsets are so common these days, and Windows is really good about handling a wide variety of common hardware.
 
This quote from the amd posting about WHEA errors is really interesting:
 
"There have been a ton  of the WHEA errors reported after bios updates since mid this year. Definitely something going on there. However this error is also a very generic error and phsyical hardware issues with CPU, GPU, Memory, power etc can all cause the same errors. As well as registry corruption and malware can do it too. So it can be very hard to pin down the errors."
 
I may have to try a completely different graphics card in this system to see if the problems persist. I have a Sapphire Pulse Radeon RX 580 I can take out of another system. If that works without any glitches, I would have to think that the crashes are related to problems with the RTX-3080 card. Life on the leading edge can be a risky proposition.

I would also recommend the fresh load of windows along with a usb full of the drivers you download to support the hardware you have installed.  Although you got what looks to be a solid machine comprised of good to decent parts, it all comes down to how they installed it and what they loaded on it.  It wouldn't cost you much for a fresh reload and drivers reinstallation other than your time as I'm sure you have a usb laying around you could fill with your drivers.  You will need a unused USB to create a boot drive with the operating system on it though.  Good luck and let us know how it goes.  

Check if the GPU is connected in the same rail with other components in the PSU. The GPU must be connected in a different rail by itself. 
 
Like many stated performed a fresh install of windows 10.  Once booting into windows, download the new drivers from the Nvidia website, once they are downloaded, uninstalled the drivers, but rebooting into safe mode unplugged the LAN cable from the computer and uninstall the drivers on Safe Mode. Once the drivers are wiped out, restart the PC, install the new drivers and during the installation plug the LAN cable. When the drivers are finished installing reboot.
Go to the bios and see if the GPU is set up correctly

daveoshinsky
I uninstalled the NVidia driver using DDU, and let Windows do an automatic reinstall of 457.71 NVidia driver. I noticed that the download from NVidia directly is a later driver, 461.72, but left the earlier driver 457.71 from Microsoft store. I also removed the tempered glass from the PC for better ventilation. Unigine Heaven 4.0 now seemed to run longer, with the GPU reaching close to 80 deg C I believe, but it eventually crashed again in the same way as earlier.
 
I earlier attached a small photo of the PCI-E connections to the GPU. Without taking the machine apart further, can I tell whether the GPU is wired properly to the PSU, on its own rail, based on the photo?  Or do I have to further disassemble the PC to clearly see the PSU, which is hidden inside a metal housing?

you need to see it. I have a feeling the GPU rail is being shared. You have to turn off completely, (unplugged the outlet cable) and look it up yourself

post approved

I would first check the power supply. 

The PSU is this model:  Apevia ATX-PR1000W 1000W 80 Plus Gold ATX12V Power Supply. It's no longer listed on their web site, with similar PR1000W remaining. I didn't see any UL number on the label (I'm in USA). I've sent an email to Apevia asking about this PSU.
There are 3 separate PCIe cables running from the GPU back to the PSU, so I don't see any "cheating" where GPU and motherboard share power cables, or the 3 GPU cables combine at any point. This may just be a single rail design.
 
Regarding WHEA representing hardware instability, I'm still trying to narrow it down. It always happens when GPU is under heavy load for a while. CyberpowerPC may be of some help eventually.

PR1000W isn't exactly the same model number. I'm not sure whether it's identical with ATX-PR1000W.
 
I'm curious how you posted a URL here. The forum SW always removes a URL when I copy & paste it.