|
2013-08-17, 09:48
|
#12 |
|
Tribal Bullet
Oct 2004
3,541 Posts |
Well, the next 'level' of ECM would have taken several minutes. Would you have been happier if the whole job took 20+several minutes in return for a chance at needing only several?
This is the dilemma in scheduling ECM+QS. How big would the value of 'several' have to be before you consider it 'paying too much for insurance' that QS not be wasted effort? |
|
|
|
2013-08-17, 16:09
|
#13 |
Sep 2009
2·1,039 Posts |
The rule of thumb is to run ECM for about 1/3 of the time needed for QS or GNFS. So 5-6 minutes ECM would be about right.
The script I'm using now would have run part of T30, varying the number of curves depending how big the number is. Chris |
|
|
|
|
2013-08-17, 16:47
|
#14 | |
|
Bamboozled!
"πΊππ·π·π"
May 2003
Down not across
10,753 Posts |
Quote:
I only ever fire up msieve when I know beyond all reasonable doubt that no factors under 30 digits are to be found. In this situation, spending even a few seconds on ECM is wasted effort. Paul |
|
|
|
|
 |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Two questions: | Dubslow | GPU Computing | 1 | 2011-08-05 18:22 |
| Questions about the QS | Carmichael | Factoring | 8 | 2007-04-10 11:30 |
| Questions | OmbooHankvald | Prime Sierpinski Project | 2 | 2005-08-01 20:18 |
| LLR questions | OmbooHankvald | Math | 6 | 2005-06-23 11:42 |
| A few questions :) | xtreme2k | Lounge | 59 | 2002-10-31 06:20 |
