[quote=axn;211238]could someone get the 64 bit sievers working on windows, please?[/quote]
If someone did this you would have my support as well.

Yeah, we have never deployed 15e on RSALS, partly because of lack of time and partly because it is a trivial way for RSALS and NFS@Home not to stomp on each other's toes: smaller sievers => smaller integers for RSALS, larger sievers => larger integers for NFS@Home.

[quote=10metreh;211269]Firstly, there won't be any automated poly selection. It's an SNFS number so very simple polys can be created by hand.

The sieving can be done on many machines and then joined up: [URL="http://www.mersenneforum.org/showthread.php?t=13133"]here[/URL] is a c171 GNFS that is being done this way.

For the matrix, it can't be run on multiple machines, but the Lanczos can be done on multiple cores.
The square root can be done in parallel too.[/quote]

I am all up for it and can run 6 +/-4 cores. Others have also indicate the willingness to participate so....

[quote=chris2be8;211204] I'll reserve the 3 numbers in t380.txt next, they should take a few days each:
[/quote]

All but 1 of the remaining small fry I reserved before then:
sigma(3226958732172192441631621169547522681619241918873519411412391273186619468910348410681419832069292266213968151378059834461739328162010783997^1)
r1=465525519940082625907596150086760475032388129 (pp45)
r2=42968165062922145361961788964387120522532996073114093 (pp53)

sigma(460863084950616131042131723595911434951353164389087358811^2)
r1=5751303329470692560038811446195633628971 (pp40)
r2=2296925895656464871936007427483592486996513701666708486363 (pp58)

sigma(588761162341205094850414198053241239006032736020251463569174595598850005653^2)
r1=2262434573266804907682123340471 (pp31)
r2=31872640844190047162391916296732200245803068403503844111906488046907 (pp68)

sigma(644524024952624237131603886126792193057893908823223821^2)
r1=19145619219867824006485846433947251552070603 (pp44)
r2=2982184993285385578322213681055944247588487561927588159 (pp55)

The last one, sigma(77384683949386242687211^6), will get done after the numbers from t380.txt. The script I'm using to do a series of numbers does them in alphabetical order by name.

Chris K

William,

Can you please pm me with the correct email address.

I am running poly selection on:

544941-1 (c115)
159753-1(C117)
1138947856317-1 (C118)
204139028584023516868625337-1 (c122)
from [url]http://oddperfect.org/composites.html[/url]

---------------------------------

Also does anyone have a GPU to run poly selection? These are small but a dozen or more so digits and I'll be getting the good ol "300 hours".

[quote=alexhiggins732;211313]William,
I am running poly selection on:

544941-1 (c115)
159753-1(C117)
1138947856317-1 (C118)
204139028584023516868625337-1 (c122)
from [URL]http://oddperfect.org/composites.html[/URL]

[/quote]

Two are good SNFS targets: [code]# sigma(20413902858402351686862533^6) = 807801526811506851480570640529 * 89588361373634892767649400463706547384552114712257915193288177740813996213250575213996132178579872697247420427062220264667
# SNFS difficulty = 152 SNFS equivalent = 109 GNFS difficulty = 122
n: 89588361373634892767649400463706547384552114712257915193288177740813996213250575213996132178579872697247420427062220264667
type: snfs
name: o20413902858402351686862533_6
m: 20413902858402351686862533
c6: 1
c5: 1
c4: 1
c3: 1
c2: 1
c1: 1
c0: 1
lss: 0
[/code]
[code]# sigma(5449^40) = 33458959591589296975357269208000879 * 8476949437780651275227153036180806569126630331555376053776236116864298908968802646333237686565475496935868954971719
# SNFS difficulty = 154 SNFS equivalent = 110 GNFS difficulty = 115
n: 8476949437780651275227153036180806569126630331555376053776236116864298908968802646333237686565475496935868954971719
type: snfs
name: o5449_40
m: 777202990921821906273566126401
c5: 5449
c4: 0
c3: 0
c2: 0
c1: 0
c0: -1
[/code]
Chris K

Chris,

Thanks for the SNFS Poly's. I already had 544941-1 (c115) completed.

I can't seem to get the perl one liner for SNFS to run w/ windows (activestate per) and am not familiar with perl.


William:

Here's the results of the last four:

[code]
544941-1 (c115)
Number: examplemc
N = 8476949437780651275227153036180806569126630331555376053776236116864298908968802646333237686565475496935868954971719 (115 digits)
Divisors found:
r1=340306547805576549488705929127 (pp30)
r2=187468232170762155162645199510511267 (pp36)
r3=132874487773914460188536355132905821308086828194091 (pp51)
Version: Msieve v. 1.44
Total time: 6.17 hours.


c117
N = 101068805287372379029332711875254224566311402807305343094690662800237581293907623300693961452786451297620782963620357 (117 digits)
Divisors found:
r1=416464852277331796460224399711164643607261599826119237027 (pp57)
r2=242682677144790015093498107583993489535833508861347376116791 (pp60)
Version: Msieve v. 1.44
Total time: 4.27 hours.


1138947856317-1 (C118)
Number: examplemc
N = 1436635830832043906961424692814108158878195526044082876397796046837374947664040135133186983201490709489054510114179783 (118 digits)
Divisors found:
r1=2859659241524071401232089177950313102143089274316063361807 (pp58)
r2=502380077308225297006783394381023499242456594064082277639369 (pp60)
Version: Msieve v. 1.44
Total time: 5.86 hours.



204139028584023516868625337-1 (c122)
N = 89588361373634892767649400463706547384552114712257915193288177740813996213250575213996132178579872697247420427062220264667 (122 digits)
SNFS difficulty: 152 digits.
Divisors found:
r1=966582970058894534634559620790542047343347498149195511 (pp54)
r2=92685640186870057437011797158697731710628635857283786716330317659197 (pp68)
Version: Msieve v. 1.44
Total time: 5.95 hours.
[/code]Also would like to reserve the next four:

[code]C123 (SNFS:160) 1804634779026713-1
236344936570220549364295662356101235610584497450705636579818692335753578100886908970495172183139960743987696903434396435057

C124 (SNFS:176) 183139449719-1

5188170227572451550798990202941840322204260865118879201661111149031441059359110594325951949605566237563793704558566652448739

C125 (SNFS:166)131953-1

49154493097762728406011947175707093833964920192086520859278912024223077443667862301287798116953142499677163570683955801279443

C125 (SNFS:171) (P43b5-1)*

94048195763790522874384977572565531419527053787785668875884015700873625057228193421829358403524429435807186247732520308705451
[/code] Any SNFS polys for these?

[quote=alexhiggins732;211478] Also would like to reserve the next four:

Any SNFS polys for these?[/quote]

3 have:
[code]#18046347790267 12 236344936570220549364295662356101235610584497450705636579818692335753578100886908970495172183139960743987696903434396435057
# sigma(18046347790267^12) = 5048072363890008487079611630851021353 * 236344936570220549364295662356101235610584497450705636579818692335753578100886908970495172183139960743987696903434396435057
# SNFS difficulty = 160 SNFS equivalent = 114 GNFS difficulty = 123
# cofactor = 53 * 79 * 2939 * 410225851276733147852331182321 (may not be prime)
n: 236344936570220549364295662356101235610584497450705636579818692335753578100886908970495172183139960743987696903434396435057
type: snfs
name: o18046347790267_12
#m: 204332734336957005738006849753373781413811577160218556275556233566357901269024877027165853817477952629895219964126978801805
Y1: 18046347790267
Y0: -325670668567274633819931290
c6: 1
c5: 1
c4: -5
c3: -5
c2: 6
c1: 3
c0: -1
[/code]
[code]#1319 52 49154493097762728406011947175707093833964920192086520859278912024223077443667862301287798116953142499677163570683955801279443
# sigma(1319^52) = 36433050172558221742542213532500157867 * 49154493097762728406011947175707093833964920192086520859278912024223077443667862301287798116953142499677163570683955801279443
# SNFS difficulty = 169 SNFS equivalent = 121 GNFS difficulty = 125
# cofactor = 107 * 3181 * 107040489155994035093126576702501 (may not be prime)
n: 49154493097762728406011947175707093833964920192086520859278912024223077443667862301287798116953142499677163570683955801279443
type: snfs
name: o1319_52
m: 21022906124188737850408991008502519
c5: 1
c4: 0
c3: 0
c2: 0
c1: 0
c0:-1739761
[/code]
[code]#5108198389729678011768295429708510355680291 4 94048195763790522874384977572565531419527053787785668875884015700873625057228193421829358403524429435807186247732520308705451
# sigma(5108198389729678011768295429708510355680291^4) = 7239699746005931776142659369681125105187754155 * 94048195763790522874384977572565531419527053787785668875884015700873625057228193421829358403524429435807186247732520308705451
# SNFS difficulty = 171 SNFS equivalent = 122 GNFS difficulty = 125
# cofactor = 5 * 11 * 131630904472835123202593806721475001912504621 (may not be prime)
n: 94048195763790522874384977572565531419527053787785668875884015700873625057228193421829358403524429435807186247732520308705451
type: snfs
name: o5108198389729678011768295429708510355680291_4
m: 5108198389729678011768295429708510355680291
c4: 1
c3: 1
c2: 1
c1: 1
c0: 1
[/code]
The other is slightly better done via GNFS.

Chris K

There is a program for automated SNFS poly generation available from [url=http://www.loria.fr/~kruppaal/numbth.html]Alex Kruppa's site[/url].

[quote=10metreh;211507]There is a program for automated SNFS poly generation available from [URL="http://www.loria.fr/%7Ekruppaal/numbth.html"]Alex Kruppa's site[/URL].[/quote]

10metreh

Thanks for the link. I will try it out. I just got MPIR compiled and running not to long ago (I am used to using the ASP.NET wrapper of libgmp but have grown tired of the slow overhead of marshaling from managed to unmanaged types), so my question is (not being a c/c++ expert):

Can I add this in a new solution and link to the MPIR lib or does this actually require GMP?

Not that I want to modify it, yet ;), but I like writing programs that automate things...

chris2be8

Thanks for the SNFS Polys.

ECM Benchmarks
 

William,


Here are the ECM benchmarks I promised you a few hours ago. As I stated up to a certain point running t35 takes the same amount of time (@6 hours) that it takes to run the number through MSIEVE or YAFU (see times in factors again).

I mind waiting a long time if I know I will get a result at then end, so to me it doesn't make sense to run smaller numbers up to 35 digits on ECM.

But then again, for those factors above I was using 8-10 cores over 3-4 computers so on the same token, that could have been 10 numbers ran up to 35 digits on ECM. I guess, as you said you will need to compare to the Silverman and Wagstaff numbers to see where the cut off point is for ECM for the most efficiency. But I am most likely telling you things you already know ;).

However, even if it is more efficient, I am not sure if I would want to do them (I might take some), because I like results. In fact, even for ECM, I wrote a GUI complete with a progress bar and all that allows me to select a digit level and input N and constantly updates me on the time elapsed and the ETA to completion.

For something like a this project, I have my mtecmrunner.exe which reads the input from an XML file (stored on a web server) runs a runs curves for digit level specified by the command line argument. I then mark any primes found so other computers don't rerun the same n at a higher level digit level.

Perhaps, I may run a certain number of ECM to finish off the auto-sumbmission of factors found to the web server, this way I at least have the framework for future projects. Currently, I manually into any primes found into a web form hosted on the same server.

---------------------------------------

Benchmarks:

The benchmarks where ran against the first occurrence of a composite of each length from the text files on oddperfect.org (including the first four factors I sumbitted to you) ranging from c92-c159). Benchmarking was performed on three systems, a Xeon QuadCore, Pentium D 2 core, and a Core Duo laptop.

The results are attached. The were to long to post here

If you need more info (the full output of individual commands, etc) let me know.

Alex