Work to do
Message boards :
Questions/Problems/Bugs :
Work to do
Message board moderation
Author | Message |
---|---|
Send message Joined: 26 Jun 08 Posts: 646 Credit: 475,830,698 RAC: 237,992 |
Cunningham numbers have always been the first priority. Unfortunately the 15e sieve becomes inefficient at SNFS difficulty above 275 digits, and we have finished all Cunningham numbers below that. The primary focus of the 15e sieve will soon shift to Fibonacci/Lucas numbers. The 16e sieve will continue to focus on Cunningham numbers. |
Send message Joined: 26 Jun 08 Posts: 646 Credit: 475,830,698 RAC: 237,992 |
With 16e, we can do anything between 275-320 digits, but only one number at a time. It's a little easier to get supercomputer support for 320 digits. :-) |
Send message Joined: 5 Sep 09 Posts: 18 Credit: 63,574 RAC: 0 |
How about add a new type of 16e to factor numbers between diff 275-300, and another 16e to tackle difficulty above 300. that would fill the gap. |
Send message Joined: 26 Jun 08 Posts: 646 Credit: 475,830,698 RAC: 237,992 |
That's a possibility, but not during this grant year. I proposed a fairly aggressive schedule, so I don't want to take computers away from those numbers. Following 2,1049+, I've proposed three GNFS factorizations of 202, 207, and 212 digits, presuming they survive ECM pretesting. |
Send message Joined: 26 Sep 09 Posts: 218 Credit: 22,841,893 RAC: 1 |
I may not live long enough to see a new grant year......... Yes you will. |
Send message Joined: 2 Oct 09 Posts: 50 Credit: 111,128,218 RAC: 0 |
That's a possibility, but not during this grant year. I proposed a fairly aggressive schedule, so I don't want to take computers away from those numbers. Following 2,1049+, I've proposed three GNFS factorizations of 202, 207, and 212 digits, presuming they survive ECM pretesting. As I was remarking over on mersenneforum Dec 28 09:43 dec27-cu26987-p60up-5t55-10M770 Jan 1 11:21 dec31-cu25200-p60up-5t55-10M770 Jan 8 15:46 jan07-cu26986-p60up-5t55-10M770 corresponding to c. 3t60 ran on the C212. (That's 79K curves with B1 = 400M, gmp-ecm default B2.) I also had t60's on the C202 and C207, which are from the recent base-3 extension, where Sam was saying that he ran t60's before adding them to the regular list. I'd say that these have already survived ECM testing. -Bruce (Incidently, 7p373 C303 has also survived a t60 (with B1 = 600M), and I'll run a bit further.) |
Send message Joined: 26 Sep 09 Posts: 218 Credit: 22,841,893 RAC: 1 |
I've proposed three GNFS factorizations of 202, 207, and 212 digits, presuming they survive ECM pretesting. First one on the list already sieving is 3,745+. What are the others two? Carlos |