On Sat, Feb 06, 2010 at 02:59:18AM -0800, Sylv1 wrote:
> Hi all, 
> i would like to have a lookup tool in order to query the tables i downloaded 
> in torrent form.The lookup is a local one, i want to look in tables i have on 
> my computer (linux 32 bits).

the lookup tool is being worked on. with the latest changes in table layout
there have come up some extra issues. and you wouldnt want a lookup tool
that crashes regularly.

> I dont know where to start.The lookup is a kind of recomputing lines of the 
> table in order to retrieve the initial state.Does anybody know if it is 
> possible to reuse the initial computing tool in order to lookup in the table? 
> Is it at least possible to reuse some methods of the code in order to perform 
> the lookup?My problem is that i have no idea of how to develop things for 
> CUDA stuff. And if we want to get a fast lookup, i think it should use the 
> CUDA capability. Any idea where i have to begin?

There are quite some differences between lookup and generation.
You could in theory use the generator, but you would have to write a bunch
of shell scripts to use it correctly and it would take ages to do a lookup.
Its not really an option. All A5/1 computations are executed on the CUDA
device during lookup.

> Another question is that, since we change the reduction function for each 
> distingushing point, does the first message sample used to lookup have to be 
> a distinguising point?If yes that would mean that the probability of success 
> would be decrease since we are not sure to get access to a 64bits sample that 
> is a DP.
> Thanks for the helpCheersSylvain
> 

The lookup data does not have to be a DP. When you do a lookup the chain
that you build must merge with a chain in the table for the lookup to be
successful. The merge has to happen in the very first step of the
computation or else you will not find a chain in the table that leads
up to your data point. So you have to find a merge but it can be anywhere
in the table, for the 2 chains to merge and produce the same end value.

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

Reply via email to