Issues with Tesla K20m

[ruuand]

Hello, I'm having some issues with cudaHashcat v1.35. The error message is cuModuleLoad() 209

I'm using a Tesla K20m on a Windows Server 2008 R2. I'm using the Nvidia Driver 340.84 (http://www.nvidia.com/download/driverRes...8297/en-us) which seems to be the latest driver available for Tesla K20m.

I've been able to run cudaHascat v1.30 without any problems (However i have to change the system date to run it...).

Any ideas about how i can solve this issue ?

Thanks in advance for the help
----------------------------------------------------------------

Here is an example of hashcat execution with the error message:
cudaHashcat64.exe -t 32 -a 7 example0.hash ?a?a?a?a example.dict
cudaHashcat v1.35 starting...

WARN: NvAPI_EnumPhysicalGPUs() -6 NVAPI_NVIDIA_DEVICE_NOT_FOUND

Device #1: Tesla K20m, 4799MB, 705Mhz, 13MCU

Hashes: 6494 hashes; 6494 unique digests, 1 unique salts
Bitmaps: 10 bits, 1024 entries, 0x000003ff mask, 4096 bytes, 17/7 rotates
Applicable Optimizers:
* Zero-Byte
* Precompute-Init
* Precompute-Merkle-Demgard
* Meet-In-The-Middle
* Early-Skip
* Not-Salted
* Not-Iterated
* Single-Salt
* Scalar-Mode
* Raw-Hash
Watchdog: Temperature abort trigger disabled
Watchdog: Temperature retain trigger disabled

ERROR: cuModuleLoad() 209

[philsmd]

For cudaHashcat to work you need to match this requirement stated on the download page ( https://hashcat.net/oclhashcat/ ): 346.x or later

Now, if you look at the nvidia documentations (for instance I have found this): http://uk.download.nvidia.com/Windows/Qu...-notes.pdf

It says on page 38: Supported Product NVIDIA Tesla K20(x/c/m/Xm/s) 64-bit OS only

But unfortunately within the *same* document it says (page 11: Legacy Support for Tesla architecture GPUs):
"Beginning with Release 343, the NVIDIA graphics drivers no longer support the Tesla generation of NVIDIA GPUs."

"The Release 340 drivers will continue to support these products until April 1, 2016 ... future driver enhancements and optimizations in driver releases after Release 340 will not support these products"

So, the information within this documentation is kind of conflicting. I would recommend to download a driver that meats the minimum requirement for oclHashcat/cudaHashcat , hence >= 346 and see if the installation and the driver works nevertheless.

[ruuand]

I tried the 347.88 driver. Now i have a "cuModuleLoad() 301" error. It still works with the version 1.30 of hashcat.

------------------------------------------------------------------

cudaHashcat64.exe -t 32 -a 7 example0.hash ?a?a?a?a example.dict
cudaHashcat v1.35 starting...

WARN: NvAPI_EnumPhysicalGPUs() -6 NVAPI_NVIDIA_DEVICE_NOT_FOUND

Device #1: Tesla K20m, 4799MB, 705Mhz, 13MCU

Hashes: 6494 hashes; 6494 unique digests, 1 unique salts
Bitmaps: 10 bits, 1024 entries, 0x000003ff mask, 4096 bytes, 17/7 rotates
Applicable Optimizers:
* Zero-Byte
* Precompute-Init
* Precompute-Merkle-Demgard
* Meet-In-The-Middle
* Early-Skip
* Not-Salted
* Not-Iterated
* Single-Salt
* Scalar-Mode
* Raw-Hash
Watchdog: Temperature abort trigger disabled
Watchdog: Temperature retain trigger disabled
Device #1: Kernel ./kernels/4318/m00000_a1.sm_35.64.ptx


ERROR: cuModuleLoad() 301

[philsmd]

are you sure that you extracted the oclHashcat archive (7z) correctly?

Please double-check that the files:
kernels/4318/m00000_a1.sm_35.64.ptx
kernels/4318/markov_le_v4.64.ptx

(main kernel folder is "kernels" under the extracted main directory of cudaHashcat, it includes the subdirectory 4318 which includes the ptx files)

I ask this question because your error message says:
CUDA_ERROR_FILE_NOT_FOUND = 301

"file not found" normally is shown whenever the kernel files do not exist (in the correct location), but it might also have other reasons in your particular case.

Please confirm if you can find those files

[ruuand]

I tried extracting the archive again: and it solved the problem

Everything seems to work fine, thanks a lot for your help.

[Rolf]

Classic 7z e instead of 7z x problem.
Closed.