Driver crashes, kernel exceptions under heavy workload P4000 (multiple users)

[i]Hello,

My organisation purchased 4x P4000 Quadro devices and have found them to be somewhat unstable under heavy workloads. We have updated all 4 machines to the latest graphics driver available, however when processing large renders in ArchiCAD or Unity users are getting a kernel exception error (Error code 3 subcode 2).

This is replicating throughout our environment. Under similar workloads, our previous graphics card (M2000) did not generate these errors.

Troubleshooting attempted:

  • Update to latest version of Quadro driver
  • Full uninstall and reinstall of graphics driver
  • Replication of fault within freshly built windows 7 64bit environment
  • Replacement of PSU in case card was getting power starved

A screenshot of this error can be provided on request. The machines we have these cards installed in do vary in specs, such as:

Device 1 running i7 6700k, 64gb DDR4, SSD / HDD / ASUS gaming motherboard, 700w SeaSonic PSU
Device 2 running i7 4700k, 64gb DDR3, SSD / HDD, ASUS gaming motherboard, 700w SeaSonic PSU
etc

All chipset drivers are installed and update to latest version, including updating to latest bios revisions.
Windows has been patched to the most recent release, all devices running Windows 7 64bit.
[/i]

Hi dylan.kellner,

Could you provide us the dataset generating this crash? Also, are there any dump files available?

Thanks,
Ryan Park

Hi Ryan,

The files impacted are large (1gb+) ArchiCAD files containing intellectual property, so cannot be uploaded. Where does Nvidia store the dump files? As it’s not a windows BSOD or equiv, there are no small/full dump files available.

Thanks

Hi dylan.kellner,

let me connect you with a system partner that could provide you further assistance. I’ll pm you for details.

Thanks,
Ryan Park