OpenCL beta for Linux

[Bitweasil]

http://blog.cryptohaze.com/2012/05/huh- ... pencl.html

Go try it out.

Supports CUDA and OpenCL and SSE2, all at once.

[Picch]

Even when specifying the --nocuda option, does the Multiforcer still look for CUDA to some degree? I tried running this in a VM to see the CPU performance, but even with the --nocuda flag it dies complaining about missing CUDA libraries.

[Bitweasil]

Yeah, slightly.

It should complain, but not die. It's the OpenCL side complaining, I think.

It won't run at all?

[blazer]

nice. I think the "--bfi_int" also works on 5xxx cards.

I'll happily sit and wait patiently for the win ver.

[Picch]

I'm actually an idiot (it's been one of those weeks) it complains about libOpenCL.so.1 not CUDA. Since this is an Ubuntu 12.04 VM (ie no CUDA or ATI ), I downloaded and installed the Intel OpenCL SDK but it complained about missing libOpenCL.so.1.

I then symlinked libOpenCL.so.1 to libOpenCL.so and got a little further, but not by much. When attempting to run a cracking job it says:

Code: Select all

./New-Multiforcer: /usr/lib/libOpenCL.so.1: no version information available (required by ./New-Multiforcer)
Error: (GRTOpenCL.cpp:231)Device Not Found

My program flags are as follows:

Code: Select all

./New-Multiforcer --nocuda -h NTLM -f test_hashes/Hashes-NTLM-Full.txt -c charsets/charsetall

For what it's worth the processor is a new i7-3770K Ivy Bridge.

I should just stop being lazy and get everything going on my Linux partition heh.

[Bitweasil]

Oh yeah. Intel OpenCL won't work right now. I only select GPUs with OpenCL.

Try passing it --nocuda --noopencl and it should just use the sse2 CPU kernels.

[Picch]

Code: Select all

+-------------------------------------------------------------------------------------------------------+
|'p' to pause                              Cryptohaze MFN 1.31                              'q' to quit |
|                                           Mode: Standalone                                            |
|Hash type     :      NTLM |                 Passwords Found                    |                       |
|Current PW len:         6 |                                                    |  0: GPU: 2.63B/s      |
|Total hashes  :        13 |  G0o|)                                             |  1: CPU: 130.52M/s    |
|Cracked hashes:         6 |  GoOd!                                             |                       |
|Total time    : 00:01:57  |  1234                                              |                       |
|WUs: 7/43 (16.3%)         |  P4s$                                              |                       |
|                          |  Bwah                                              |                       |
|--------------------------|  bad                                               |                       |
| Td 0: CID 55211.         |                                                    |                       |
| Td 1: CID 22714.         |                                                    |                       |
| Starting pw len 6        |                                                    |                       |
| Td 1: out of WU.         |                                                    |                       |
| Waiting for workunits... |                                                    |                       |
| Td 0: out of WU.         |                                                    |                       |

As advertised, it works and it's fast. GTX580 overclocked and i7-3770K overclocked to 4.2ghz

[blazer]

2.6Bln seems pretty quick (not sure how much you oced) looks comparable to pluscat.

GW bitweasil.

[Picch]

Here's my 580 Specs:
Core Clock: 900mhz
Shader Clock: 1800mhz
Memory Clock: 4212mhz

[Bitweasil]

2.6Bln seems pretty quick (not sure how much you oced) looks comparable to pluscat.

GW bitweasil.

It's at least in the ballpark of hashcat. He usually has a slight edge, but it's no longer embarrassingly faster. And I'm open source, so if you have ideas to speed it up, please, make use of them and submit patches!