SnuCL
#1
Re: http://hashcat.net/forum/thread-2714.html

@epixoip did you try SnuCL? Does it work at all? Just from reading the manual, it looks like it would need to be included while building ohc so maybe we can't really try it out ourselves.

Are the platform names accepted in oclHashcat 1.00? Here's an excerpt from the manual (page 10):
Table 3.1: The name and vendor of the SnuCL platforms
Code:
Platform      CL_PLATFORM_NAME   CL_PLATFORM_VERSION
SnuCL CPU     "SnuCL CPU"        "Seoul National University"
SnuCL Single  "SnuCL Single"     "Seoul National University"
SnuCL Cluster "SnuCL Cluster"    "Seoul National University"
#2
it's on my to-do list, but quickly slipping to the bottom of the list.

it certainly looks like a promising alternative to VCL. and the team seems to be a bit more serious about OpenCL middleware than MOSIX, especially since they have a cluster on the Top 500 chart with 224 7970s.

it still has the same requirements as VCL (Infiniband, etc), with one additional requirement: the program needs to be built with MPI support. so if we were to get SnuCL support in hashcat, it will have to be built with MPI support. so what will likely have to happen is another dedicated binary, e.g. sclHashcat or something.

it's slipped to the back burner for us though, because we have something far more nefarious in mind for clustering Smile
#3
Thanks, OK, I'll do some experiments with JtR and we'll see how that turns out first.

(12-12-2013, 09:31 PM)epixoip Wrote: it's slipped to the back burner for us though, because we have something far more nefarious in mind for clustering Smile
Nefarious lol, care to elaborate a little?
#4
Yes, please elaborate- VCL has been pissing me off! Sad
#5
i Hope its distributed "logic" to the end node. So we dont need the infiniband connection. Dont know why it haven’t been done before.

1UP. Would like some more info. Smile