View Single Post
Old 2009-05-11, 12:24   #27
gd_barnes
 
gd_barnes's Avatar
 
May 2007
Kansas; USA

2·5,087 Posts
Default

With the sieving drive now complete, I have now sent a file sieved to P=45T to David for n=630001-650000 for reloading into port 8000. This will involve stopping the server temporarily, deleting the current k/n pairs in that range at the end of the file, pasting the new pairs in at the end, and restarting the server.

I debated whether to go through with this hassle but ultimately felt it was worth it. It means that 4000+ composite k/n pairs will NOT have to be tested meaning that we should find primes around 2% faster than before when we get to that range, which should be in < 2 days.

Thanks to everyone for the excellent and fast job of sieving the range. Now, n=630K-1M will all test 2% faster, which amounts to nearly 75,000 k/n pairs that will not have to be tested! This will especially save a huge amount of testing time for n>900K, where tests will take, on average, more than twice as long as the current n=~625K range.


Gary
gd_barnes is offline   Reply With Quote