On Fri, 2010-08-27 at 18:25 +0200, Guillaume wrote:
> Hello all,
> 
> I would like to help computing rainbow tables and decrypt A5/1 cypher
> stream.

We have already computed a set of tables that are usable for decrypting
A5/1 - but computing supplementary tables may be useful for some
application where very high probability of recovery is needed. 


> - AMD ATI Radeon 5970 HD 2GB (2 GPU) from SFX
> - 2 TB Seagate Barracuda (raid 10)
> 
> I run Ubuntu 32bits last release (ubuntu-10.04.1-desktop-i386.iso) on a
> C2D Intel 6320 OV 2380 Mhz with 2GB DDR2 5300 RAM @ 340 Mhz dual channel
> Motherboard is ASUS P5B Deluxe (Chipset P965 / ICH8R from Intel)
> 
> Does source code Radeon ATI Stream, OpenCL compatible ?
> (latest code I got - tmto-svn - Revision 197)

Please use the code from the git repository where the code is
maintained. The svn repo will only recieve perodic syncs from git, and
is lagging quite a bit behind.

> Does both GPU on the card will be automatically used ? (with little
> change on Config / Makefile of course)

Kraken uses all cards pr default, for table generation you need to make
a small edit in GenTable.py

> If I buy a second Radeon 5970 dual GPU, does the code able to use both
> cards - without using ATI CrossFire of course ? (without changing to
> much the code...)

2x5970 is a tested configuration, but for optimum speed it was found to
be best to compute 2 tables at a time in parallel to get maximum GPU
utilization.

> Does my config seems enough to compute the tables and crack the cypher ?

For cracking A5/1 you have more than enough power but I would recommend
getting a more disks, the easiest option is to try and obtain a copy of
the precomputed tables. Otherwise you will have to spend 2-4 months
computing your own set. If you do decide to compute your own tables,
please consult me for getting tables IDs assign so we can avoid doing
duplicate work.

-
Frank

_______________________________________________
A51 mailing list
[email protected]
http://lists.lists.reflextor.com/cgi-bin/mailman/listinfo/a51

Reply via email to