Error question

[niceguy123]

Can someone point me in the right direction as to why I am getting this errors? I am trying to run -a 0 -m 11300 against own hash but the attack fails in couple of minutes.
I have 537.34 driver and cuda 12.2.2 installed. OS Windows 10 Pro. Couple of rtx 3090

When I run the attack i get multiple messages in red text like:

cuMemFree(): an illegal instruction was encountered
cuEventDestroy(): an illegal instruction was encountered
cuStreamDestroy(): an illegal instruction was encountered
cuModuleUnload(): an illegal instruction was encountered

and then it quits

[niceguy123]

I forgot I am using hashcat-6.2.6+747.

Also why is hashcat building/rebuilding dictionary cache every single time?
Also why I am getting better benchmark results with opencl rather then cuda?
Also I can't get hashcat to even start on hash mode 23800. I tried with optimized and pure kernels. With cuda and opencl etc but it simply quits every time

If anyone can share some wisdom I would be very grateful!
Thank you in advance

[Snoopy]

please run hashcat -I and post output

[niceguy123]

C:\Hashcat>hashcat -I
hashcat (v6.2.6-747-g5b06ffe63) starting in backend information mode

Support for HIPRTC was dropped by AMD Adrenalin Edition 22.7.1 and later.
This is not a hashcat problem.

Please install the AMD HIP SDK

CUDA Info:
==========

CUDA.Version.: 12.2

Backend Device ID #1 (Alias: #9)
Name...........: NVIDIA GeForce RTX 3090
Processor(s)...: 82
Clock..........: 1695
Memory.Total...: 24575 MB
Memory.Free....: 23320 MB
Local.Memory...: 99 KB
PCI.Addr.BDFe..: 0000:02:00.0

Backend Device ID #2 (Alias: #10)
Name...........: NVIDIA GeForce RTX 3090
Processor(s)...: 82
Clock..........: 1860
Memory.Total...: 24575 MB
Memory.Free....: 23320 MB
Local.Memory...: 99 KB
PCI.Addr.BDFe..: 0000:03:00.0

Backend Device ID #3 (Alias: #11)
Name...........: NVIDIA GeForce RTX 3090
Processor(s)...: 82
Clock..........: 1695
Memory.Total...: 24575 MB
Memory.Free....: 23320 MB
Local.Memory...: 99 KB
PCI.Addr.BDFe..: 0000:04:00.0

Backend Device ID #4 (Alias: #12)
Name...........: NVIDIA GeForce RTX 3090
Processor(s)...: 82
Clock..........: 1695
Memory.Total...: 24575 MB
Memory.Free....: 23320 MB
Local.Memory...: 99 KB
PCI.Addr.BDFe..: 0000:05:00.0

Backend Device ID #5 (Alias: #13)
Name...........: NVIDIA GeForce RTX 3090
Processor(s)...: 82
Clock..........: 1695
Memory.Total...: 24575 MB
Memory.Free....: 23320 MB
Local.Memory...: 99 KB
PCI.Addr.BDFe..: 0000:06:00.0

Backend Device ID #6 (Alias: #14)
Name...........: NVIDIA GeForce RTX 3090
Processor(s)...: 82
Clock..........: 1860
Memory.Total...: 24575 MB
Memory.Free....: 23320 MB
Local.Memory...: 99 KB
PCI.Addr.BDFe..: 0000:07:00.0

Backend Device ID #7 (Alias: #15)
Name...........: NVIDIA GeForce RTX 3090
Processor(s)...: 82
Clock..........: 1860
Memory.Total...: 24575 MB
Memory.Free....: 23320 MB
Local.Memory...: 99 KB
PCI.Addr.BDFe..: 0000:08:00.0

OpenCL Info:
============

OpenCL Platform ID #1
Vendor..: Intel(R) Corporation
Name....: Intel(R) OpenCL HD Graphics
Version.: OpenCL 3.0

Backend Device ID #8
Type...........: GPU
Vendor.ID......: 8
Vendor.........: Intel(R) Corporation
Name...........: Intel(R) UHD Graphics 750
Version........: OpenCL 3.0 NEO
Processor(s)...: 32
Clock..........: 1300
Memory.Total...: 52343 MB (limited to 2047 MB allocatable in one block)
Memory.Free....: 26112 MB
Local.Memory...: 64 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 30.0.101.1994

OpenCL Platform ID #2
Vendor..: NVIDIA Corporation
Name....: NVIDIA CUDA
Version.: OpenCL 3.0 CUDA 12.2.138

Backend Device ID #9 (Alias: #1)
Type...........: GPU
Vendor.ID......: 32
Vendor.........: NVIDIA Corporation
Name...........: NVIDIA GeForce RTX 3090
Version........: OpenCL 3.0 CUDA
Processor(s)...: 82
Clock..........: 1695
Memory.Total...: 24575 MB (limited to 6143 MB allocatable in one block)
Memory.Free....: 23744 MB
Local.Memory...: 48 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 537.34
PCI.Addr.BDF...: 02:00.0

Backend Device ID #10 (Alias: #2)
Type...........: GPU
Vendor.ID......: 32
Vendor.........: NVIDIA Corporation
Name...........: NVIDIA GeForce RTX 3090
Version........: OpenCL 3.0 CUDA
Processor(s)...: 82
Clock..........: 1860
Memory.Total...: 24575 MB (limited to 6143 MB allocatable in one block)
Memory.Free....: 23744 MB
Local.Memory...: 48 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 537.34
PCI.Addr.BDF...: 03:00.0

Backend Device ID #11 (Alias: #3)
Type...........: GPU
Vendor.ID......: 32
Vendor.........: NVIDIA Corporation
Name...........: NVIDIA GeForce RTX 3090
Version........: OpenCL 3.0 CUDA
Processor(s)...: 82
Clock..........: 1695
Memory.Total...: 24575 MB (limited to 6143 MB allocatable in one block)
Memory.Free....: 23744 MB
Local.Memory...: 48 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 537.34
PCI.Addr.BDF...: 04:00.0

Backend Device ID #12 (Alias: #4)
Type...........: GPU
Vendor.ID......: 32
Vendor.........: NVIDIA Corporation
Name...........: NVIDIA GeForce RTX 3090
Version........: OpenCL 3.0 CUDA
Processor(s)...: 82
Clock..........: 1695
Memory.Total...: 24575 MB (limited to 6143 MB allocatable in one block)
Memory.Free....: 23744 MB
Local.Memory...: 48 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 537.34
PCI.Addr.BDF...: 05:00.0

Backend Device ID #13 (Alias: #5)
Type...........: GPU
Vendor.ID......: 32
Vendor.........: NVIDIA Corporation
Name...........: NVIDIA GeForce RTX 3090
Version........: OpenCL 3.0 CUDA
Processor(s)...: 82
Clock..........: 1695
Memory.Total...: 24575 MB (limited to 6143 MB allocatable in one block)
Memory.Free....: 23744 MB
Local.Memory...: 48 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 537.34
PCI.Addr.BDF...: 06:00.0

Backend Device ID #14 (Alias: #6)
Type...........: GPU
Vendor.ID......: 32
Vendor.........: NVIDIA Corporation
Name...........: NVIDIA GeForce RTX 3090
Version........: OpenCL 3.0 CUDA
Processor(s)...: 82
Clock..........: 1860
Memory.Total...: 24575 MB (limited to 6143 MB allocatable in one block)
Memory.Free....: 23744 MB
Local.Memory...: 48 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 537.34
PCI.Addr.BDF...: 07:00.0

Backend Device ID #15 (Alias: #7)
Type...........: GPU
Vendor.ID......: 32
Vendor.........: NVIDIA Corporation
Name...........: NVIDIA GeForce RTX 3090
Version........: OpenCL 3.0 CUDA
Processor(s)...: 82
Clock..........: 1860
Memory.Total...: 24575 MB (limited to 6143 MB allocatable in one block)
Memory.Free....: 23744 MB
Local.Memory...: 48 KB
OpenCL.Version.: OpenCL C 1.2
Driver.Version.: 537.34
PCI.Addr.BDF...: 08:00.0

[Snoopy]

plz try the following, add following options and see whether the attack starts or also fails ot the beginning
-D2 -d1
when running add devices 2 up to 7 (this sticks hashcat to your graphic cards only)
-D2 -d1,2
-D2 -d1,2,3
until it fails again, i think there could be a problem with the overall system RAM you provide as its just a fraction of your GPU RAM, i didnt find the thread right now where a dev answered how much RAM needed per GPU RAM

[niceguy123]

From what I remember it is best to have 1:1 of VRAM to system RAM.
The system I am using has 128GB installed on the motherboard. I will try to add some of the SSD as RAM and see if it makes a difference.

"plz try the following, add following options and see whether the attack starts or also fails ot the beginning
-D2 -d1
when running add devices 2 up to 7 (this sticks hashcat to your graphic cards only)
-D2 -d1,2
-D2 -d1,2,3"

Ok it does not want to start with more then 4 cards that way. And most of the times I have to try several time to start it even with 2-3 cards. It gives - "Initializing backend runtime for device #2. Please be patient..." and quits and I have to try again and again until it starts. Usually from 3rd, 4th time (This is with up to 4 cards)
With anything more it just quits on ""Initializing backend runtime for device #1. Please be patient..." every time.

Without those options -D -d all cards are starting but hashcat fails after 8-10 minutes with the errors from the first post

[niceguy123]

After all the tests yesterday it stopped working at all. I had to actually "reinstall" hashcat to run it again. Anyone any ideas?
Maybe I should switch to Linux.

[niceguy123]

After more tests it might be related to somewhat aggressive undervolting I did on the GPUs

[MrRaja]

To resolve issues with Hashcat on Windows and multiple GPUs, ensure your system has a balanced VRAM to system RAM ratio, avoid aggressive GPU undervolting, and consider switching to Linux for better GPU support.

[niceguy123]

Thanks for the reply! Yes it seems the undervolting was an issue.
Also do you have any idea on the other questions I asked?
Why is hashcat building/rebuilding dictionary cache every single time?
Why I am getting better benchmark results with opencl rather then cuda?