|
2005-01-20, 15:05
|
#12 |
Aug 2002
Termonfeckin, IE
22×691 Posts |
Yes, there is a small chance. But with any luck we will be able to factor the number completely and know what is the smallest factor for sure.
|
|
|
|
2005-01-20, 15:18
|
#13 |
Jan 2005
Transdniestr
50310 Posts |
Hi Xyxxy,
Great. I'll kick off the 40 digit search soon. I have a Pentium M 1.6Mhz with 512M of RAM that won't be running anything else. Please advise if you want to split that search and specifically the ranges we should try. Thanks |
|
|
|
|
2005-01-20, 15:22
|
#14 | |
"Mike"
Aug 2002
2·23·179 Posts |
Quote:
Just do a batch of curves with B1=3e6 and report them... If you are using GMP-ECM, we need 2900 curves, so if we report in every 100 we should be okay... If you are using Prime95 or Mprime we need to do a little bit of extra work to calculate the conversion ratio, since 1 curve in Prime95/Mprime is worth less than one curve with GMP-ECM... If you need help setting up GMP-ECM or Prime95/Mprime let us know... PS - Should this thread be moved to the factoring forum or should we leave it here? |
|
|
|
|
|
2005-01-20, 15:49
|
#15 |
|
Jan 2005
Transdniestr
503 Posts |
Ah understood, the curves are random. Yes, this really is turning into a factoring thread.
|
|
|
|
|
2005-01-22, 03:49
|
#16 |
|
"Mike"
Aug 2002
2·23·179 Posts |
1000 curves done using B1=3e6 and B2=4016636514...
|
|
|
|
|
2005-01-22, 04:58
|
#17 |
|
Jan 2005
Transdniestr
50310 Posts |
Xyxxy,
This is great. I have only done about 80 so far. How long does the average curve take for you? Right now, it's about 4 minutes for me at 3e6. |
|
|
|
|
2005-01-22, 05:09
|
#18 | |
|
"Mike"
Aug 2002
823410 Posts |
Quote:
I have a pile of work queued up, so I'm not sure how long it will take me to finish another set of curves... Keep us informed about your progress and hopefully other people will kick in some work too... I'll try to knock out another 500 tonight if I can... |
|
|
|
|
|
2005-01-22, 15:53
|
#19 |
|
Jan 2005
Transdniestr
503 Posts |
Step1 and Step2 times
Xyzzy,
I noticed something rather odd. The sum of the step1 and step2 times consistently averages just under 4 minutes for me. However, the actual time per curve is much longer. For instance, I kicked off three different sessions last night with the same command line. The sum of the step1 and step2 times consistently averages just under 4 minutes for the three sessions. One sessions has processed 33 curves, one 32, and the last 59. This is after 9 hours or so. It's quite puzzling. Have you seen this sort of issue before? Thanks, Grandpa |
|
|
|
|
2005-01-22, 16:15
|
#20 |
|
"Mike"
Aug 2002
2·23·179 Posts |
Yes, the time is not wall time, it is computed time...
For example, say you are running a process at 95% total CPU power and run ECM for the remaining 5%... It still takes 4 minutes to run your curve, but since the ECM process is only getting 5% of the CPU, it takes it 20 times longer, from a wall clock point of view... Running more than one instance just slows everything down... I run several instances but I do so at different "nice" levels, and I do this for a specific reason... |
|
|
|
|
2005-01-22, 16:17
|
#21 |
|
"Mike"
Aug 2002
202A16 Posts |
500 curves done using B1=3e6 and B2=4016636514...
|
|
|
|
|
2005-01-24, 01:22
|
#22 | |
|
Jan 2005
Transdniestr
503 Posts |
Quote:
|
|
|
|
|
 |
| Thread Tools | |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Carl Pomerance himself about 210 | YuL | Math | 3 | 2017-06-02 10:51 |
| Exercise 1.23 in Crandall & Pomerance | sean | Factoring | 2 | 2006-10-23 21:08 |
| The original paper on the Crandall/Fagin DWT | Barry Fagin | Math | 2 | 2006-01-04 19:46 |
| Crandall & Pomerance | Numbers | Math | 16 | 2005-10-16 00:53 |
| Question about a power series | Orgasmic Troll | Math | 1 | 2004-09-13 19:01 |
