Register FAQ Search Today's Posts Mark Forums Read

 2020-12-04, 18:49 #804 YuL     Feb 2012 Paris, France 7×23 Posts Let n be a prime number with 500 < n < 1000. I'm planning on doing some ECM on the numbers 17n-1. For n=503, I did 7100@3e8 back in 2014/2015. For n>503, I'll begin with a t35. I found three factors: P33 for 17^521-1 P32 for 17^523-1 P29 for 17^929-1 Note: exponents 563, 577 and 683 are excluded from this effort for the moment because Ryan did some effort on them (maybe a t50 since a 49 digits factor has been found for 17^577-1) since they are in mwrb2100.txt. Last fiddled with by YuL on 2020-12-04 at 18:57 Reason: Added a note and the factor found while I was writting the post
2020-12-05, 17:39   #805
lavalamp

Oct 2007
Manchester, UK

53D16 Posts

Quote:
 Originally Posted by YuL Let n be a prime number with 500 < n < 1000. I'm planning on doing some ECM on the numbers 17n-1.
It might be faster for these numbers to run stage 1 ECM with prime95, and only run stage 2 with GMP-ECM. It was definitely faster for me when I was running 10^999+13.

2020-12-05, 23:10   #806
YuL

Feb 2012
Paris, France

7×23 Posts

Quote:
 Originally Posted by lavalamp It might be faster for these numbers to run stage 1 ECM with prime95, and only run stage 2 with GMP-ECM. It was definitely faster for me when I was running 10^999+13.

Thanks, I used to do that... Now I just run a GMP-ECM linked with GWNUM
and it does the job :)

Code:
GMP-ECM 7.0.4 [configured with GMP 6.1.2, GWNUM 29.8, --enable-asm-redc] [ECM]
Tuned for x86_64/k8/params.h
Due to incompatible licenses, this binary file must not be distributed.
Input number is (17^509-1)/(2^4*1019*10181*27487*91621) (609 digits)
Found number: 1*17^509 + -1
[Sat Dec  5 22:00:48 2020]
Using mpz_mod
Using B1=3000000, B2=4592487916, polynomial Dickson(6), sigma=0:16358333240206248675
dF=14400, k=2, d=150150, d2=17, i0=3
Expected number of curves to find a factor of n digits:
35    40    45    50    55    60    65    70    75    80
336    2462    21289    210320    2370354    3e+07    4.2e+08    6.4e+09    1.2e+11    9.6e+15
Using gwnum_ecmStage1(1, 17, 509, -1, 3000000, 1)
Step 1 took 23068ms
Estimated memory usage: 194.24MB
Initializing tables of differences for F took 68ms
Computing roots of F took 1876ms
Building F from its roots took 2316ms
Computing 1/F took 1048ms
Initializing table of differences for G took 116ms
Computing roots of G took 1640ms
Building G from its roots took 2380ms
Computing roots of G took 1844ms
Building G from its roots took 2340ms
Computing G * H took 692ms
Reducing  G * H mod F took 912ms
Computing polyeval(F,G) took 4648ms
Computing product of all F(g_i) took 36ms
Step 2 took 19972ms
********** Factor found in step 2: 10880130373003275175614033001151500697
Found prime factor of 38 digits: 10880130373003275175614033001151500697
Composite cofactor ((17^509-1)/(2^4*1019*10181*27487*91621))/10880130373003275175614033001151500697 has 572 digits

2020-12-06, 02:50   #807
lavalamp

Oct 2007
Manchester, UK

32·149 Posts

Quote:
 Originally Posted by YuL Thanks, I used to do that... Now I just run a GMP-ECM linked with GWNUM and it does the job :)
Well now I'm jealous. Wish I knew how to compile things for Windows, let alone with fancy alternate libraries.

I'm not sure why it says this:
Quote:
 Originally Posted by YuL Due to incompatible licenses, this binary file must not be distributed.
Because the only info I can find for the license of GWNUM is this:
Quote:
 Copyright (c) 1996-2017, Mersenne Research, Inc. All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: (1) Redistributing source code must contain this copyright notice, limitations, and disclaimer. (2) If this software is used to find Mersenne Prime numbers, then GIMPS will be considered the discoverer of any prime numbers found and the prize rules at http://mersenne.org/prize.htm will apply.
I don't know too much about licenses but it doesn't see like this would be incompatible with GPL to me.

This code exists in ECM, so presumably it CAN be GPL compliant?
Code:
#ifdef HAVE_GWNUM
#ifdef gwnum_is_gpl
if (! gwnum_is_gpl ())
#endif
printf ("Due to incompatible licenses, this binary file must not "
"be distributed.\n");

2020-12-06, 15:14   #808
ryanp

Jun 2012
Boulder, CO

2·7·19 Posts

Quote:
 Originally Posted by lavalamp Well now I'm jealous. Wish I knew how to compile things for Windows, let alone with fancy alternate libraries.
Don't know a thing about compiling on Windows, but something like this worked for me on Linux. First get prime95 source and build gwnum:

Code:
$wget http://www.mersenne.org/ftp_root/gimps/p95v303b6.source.zip$ unzip p95v303b6.source.zip
$make -f make64 -j 16 After that, you should have a "gwnum.a". Then configure and build GMP-ECM with it: Code: $ ./configure --with-gwnum=$HOME/p95/gwnum$ make -j 16
The gcc command failed for me at first, so I manually had to redo it and add "-ldl" at the end.

Last fiddled with by ryanp on 2020-12-06 at 15:14 Reason: typo

2020-12-06, 15:28   #809
henryzz
Just call me Henry

"David"
Sep 2007
Cambridge (GMT/BST)

10110101111102 Posts

Quote:
 Originally Posted by lavalamp Well now I'm jealous. Wish I knew how to compile things for Windows, let alone with fancy alternate libraries. I'm not sure why it says this:Because the only info I can find for the license of GWNUM is this:I don't know too much about licenses but it doesn't see like this would be incompatible with GPL to me. This code exists in ECM, so presumably it CAN be GPL compliant? Code: #ifdef HAVE_GWNUM #ifdef gwnum_is_gpl if (! gwnum_is_gpl ()) #endif printf ("Due to incompatible licenses, this binary file must not " "be distributed.\n");
The easy answer on Windows 10 is usually to compile on WSL.

2020-12-06, 16:03   #810
YuL

Feb 2012
Paris, France

16110 Posts

Quote:
 Originally Posted by henryzz The easy answer on Windows 10 is usually to compile on WSL.

Yes, I remember making it compile/work on Windows 10 back in early 2019
but I can't remember whether I did it using WSL or whether I was able to
make it compile using cygwin + mingw-gcc.

 2020-12-08, 15:35 #811 ryanp     Jun 2012 Boulder, CO 2×7×19 Posts FWIW, updated list of factors found so far: https://cs.stanford.edu/~rpropper/opn.txt
2020-12-09, 11:04   #812
RichD

Sep 2008
Kansas

2·31·53 Posts

Quote:
 Originally Posted by ryanp FWIW, updated list of factors found so far: https://cs.stanford.edu/~rpropper/opn.txt
I get expired certificate.

 2020-12-09, 11:08 #813 kruoli     "Oliver" Sep 2017 Porta Westfalica, DE 439 Posts Working fine yesterday and currently for me. Chrome Canary 89.0.4349.2. Windows 10, v2004. Accessed from Germany.
2020-12-09, 16:03   #814
YuL

Feb 2012
Paris, France

16110 Posts

Quote:
 Originally Posted by YuL Let n be a prime number with 500 < n < 1000. I'm planning on doing some ECM on the numbers 17n-1. [...]

Factors found so far:
17 509 10880130373003275175614033001151500697
17 521 507732028088571194591510571526693
17 523 89163041550752855557680359884199
17 541 210851698687146481442232546078496897
17 541 902304286550029874166103436203835437
17 911 1732550124080428168413011275808287
17 929 24635452104352854844270989359
17 967 34714203016693178317175265507286452661453
17 997 247157806053731011193258357444587433

 Similar Threads Thread Thread Starter Forum Replies Last Post Xyzzy GPU Computing 1 2017-05-17 20:22 Mark Rose GPU Computing 52 2016-07-02 12:11 firejuggler GPU Computing 12 2016-02-23 06:55 Elhueno Homework Help 5 2008-06-12 16:37 jchein1 Factoring 30 2005-05-30 14:43

All times are UTC. The time now is 13:12.

Mon Mar 8 13:12:19 UTC 2021 up 95 days, 9:23, 0 users, load averages: 1.81, 2.30, 2.39