2009-08-28, 15:09 | #1 |
Account Deleted
"Tim Sorbera"
Aug 2006
San Antonio, TX USA
2×3×23×31 Posts |
TF factor credit
How is the credit for factors found using TF calculated? On M455847149 I found a factor of 13426491253603038089 (63.54 bits) and got 0.0064 GHz-days credit. Then on M455848213 I found a factor of 17700866372653227383 (63.94 bits) and got 0.0044 GHz-days credit. Why would the smaller factor on two similarly-sized candidates get a significant amount more credit?
http://mersenne-aries.sili.net/credit.php says it should be 0.003107 and 0.001061 GHz-days, respectively. I'm more confused now. It doesn't make any sense, and nothing agrees. (I should note that in this area, 59 to 64 bit level always gives 0.0076 GHz-days, and 64 to 65, 0.0082. This agrees with the site linked above, which calculates 0.007640 and 0.008197.) Last fiddled with by Mini-Geek on 2009-08-28 at 15:15 |
2009-08-28, 16:01 | #2 |
Apr 2007
Spessart/Germany
2×83 Posts |
Hello,
as far as I understood the TF it is not linear within a bitlevel. Thus the factors were not found after 54% and 94% of bitlevel 64 but random. The CPU-credit you got should be linear to the time you needed to find the factor. TF stops after finding a factor, thus the numbers could still have smaller factors with 64 bit (but no factors with less bits). |
2009-08-28, 18:37 | #3 |
"Jacob"
Sep 2006
Brussels, Belgium
73F_{16} Posts |
The calculation is explained in the following post http://www.mersenneforum.org/showpos...&postcount=204. In short factors are tested in classes, if you find a factor in low mod120 class you will earn less credit than in a high class. It aims to be proportional to the time spent.
Jacob |
Thread Tools | |
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
why DC credit can be so different | rudi_m | PrimeNet | 14 | 2017-10-27 09:53 |
TF Factor Found CPU Credit | TheMawn | GPU Computing | 3 | 2013-06-17 06:21 |
mfaktc TF credit 2x higher if factor found? | S34960zz | PrimeNet | 10 | 2011-10-13 07:00 |
v4 Credit on v5 | precius1 | Information & Answers | 3 | 2008-11-03 22:23 |
Shortest time to complete a 2^67 trial factor (no factor) | dsouza123 | Software | 12 | 2003-08-21 18:38 |