help use hashcat
#2
Hi nashcat,

This is a classic GPU drop / TDR crash. When one card crashes mid-job, the CUDA context breaks and all subsequent CUDA/NVML calls fail—which causes that massive cascade of unknown error messages during cleanup.

You can actually see the culprit right before the crash in your status log:
Hardware.Mon.#7..: Temp: 46c Core: 139MHz Mem: 810MHz Bus:1

Device #7 crashed and dropped down to idle clocks.

A few things to check on your rig:

Riser & BIOS (Most likely):
Swap out the PCIe riser and power cable for GPU #7. Also, go into your motherboard BIOS and force the PCIe slot speeds to Gen 2 or Gen 1 (instead of Auto/Gen 3). This fixes a lot of bus instability on multi-GPU setups.

Windows TDR Delay:
KeePass (-m 13400) with 500k iterations is a heavy workload. Windows resets the display driver if a GPU kernel execution takes longer than 2 seconds. In your registry (HKLM\SYSTEM\CurrentControlSet\Control\GraphicsDrivers), set the DWORD TdrDelay to 60 (Decimal) and reboot.

Isolate the GPU:
Try running the attack without GPU 7 (-d 1,2,3,4,5,6) or test GPU 7 by itself (-d 7). That will quickly tell you if it's hardware/power related to that specific card or slot.

Hope this helps!
Reply


Messages In This Thread
help use hashcat - by 98713271@ - 08-02-2026, 12:19 AM
RE: help use hashcat - by tH3_z3R0 - 08-13-2026, 09:39 PM