[RC5] LIFO? Is this true?

gindrup at okway.okstate.edu gindrup at okway.okstate.edu
Mon May 11 11:45:35 EDT 1998


     Such a file would have to have room for the following list of 
     blocks:
     1) one complete fetch,
     2) one complete flush, and
     3) one block per client cracking thread.
     
     This size would have to be updated any time a configuration file was 
     changed.  If the configuration files differed, then all sorts of 
     hell would break loose as various fetchers/flushers would not agree 
     on the expected size of the file and would be trying to reorganize 
     it while other cracking threads were locking parts of the files 
     down.
     
     This is part of what several of these rewriting schemes are 
     overlooking: pieces of the buff-* files are locked down at any given 
     instant by the clients that are sharing the buffers.  Breaking the 
     buffer sharing is not a good idea.
            -- Eric Gindrup ! gindrup at Okway.okstate.edu


______________________________ Reply Separator _________________________________
Subject: Re: [RC5] LIFO?  Is this true? 
Author:  <rc5 at llamas.net> at SMTP
Date:    5/8/98 1:24 PM


On Thu, 7 May 1998 23:09:04 -0400 (EDT), Phil Gregory wrote:
     
>...  FIFO rewriting the file
>every time also adds them at the end of the file, but the constant 
>rewriting is bad anyway.  The other scenario for a FIFO either adds at 
>the end or the beginning.  If at the end, then the file keeps growing and 
>growing.  If at the beginning, what do you do if you have more blocks than 
>room for them?
     
Or you can keep a separate 1 bits per block map to keep track of which 
blocks have been loaded or checked plus each client keeps a pointer to 
the current block and then you would run through the blocks in a 
circular fashion.  This would have neither of your disadvantages and be 
mostly FIFO.  
     
Main difference in code would be the config file itself could not hold 
the number of blocks to load, but this information could be easily 
determined from the file itself.  If the filesize is zero then it 
should autmatically run the initialization/resize routine would set up 
the file.
     
0 - empty and/or already checked and results placed in buff-out.rc5 
1 - loaded but not completely checked
     
     
--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net 
rc5-digest subscribers replace rc5 with rc5-digest
     
     

--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list