Please help i am new

[maniacken]

like the title says i am new to the project. i have done one test for the sob project, and then started reading the forums. Sieving sounded kind of interesting. I wanted to try this once to see which i liked doing better.


i downloaded
Sob text document
Sob probe document
SobSieve_1_34
sobsieveconsole_1_34

i then click on the SobSieve_1_34 icon and i get a box that says

sieve from p= ???????? billion

to p= ???????? billion


i was wondering what numbers i should put in for this, aka do i need to reserve a range. also how long should this take. i would like a range that should take an hour or so to get my feet wet. i have an amd running at 1875Mhz.

any help is appreciated

kenny

[maniacken]

just wondering if i should reserve the range 152700-152800. since that was the lowest open number i saw on the top post.

[MikeH]

Please take a look here , this should answer most of your questions.

Just re-read one of your comments. With a PC of that spec you'll sieve 100G in about 3 days. We have now sieved so deep that if you take a range that will run for an hour you will find zero factors.

[maniacken]

thanks for your quick reply. i think i will try the (hopefully 100G segment i said earlier) or is their a lower range to do? 3 days isnt that long anyway. what all do i have to do? just reserve the range on the 1M - 20M post.

and how long do we have to return the results??

[Nuri]

to make sure:

this is the location of the dat file: http://www.aooq73.dsl.pipex.com/SobDat_n1M-20M.zip

This is the location of the latest client: http://n137.ryd.student.liu.se/proth_sieve.php

reserve the range on 1m-20m post: http://www.free-dc.org/forum/showthr...&threadid=3501

A post like

Reserving
152700-152800 maniacken

will be enough. Please check if the range you want to reserve is not reserved by anyone else on the list and/or on one of the remaining posts on that page.

run the client for the range you reserved (if you reserved 152700-152800, then)

sieve from p= 152700

to p= 152800



submit the factors you find to the factor submission page when finished (or if you reserved a larger range, from time to time. once every couple of days will be enough) (http://www.seventeenorbust.com/sieve/)

post once more on 1m-20m post: http://www.free-dc.org/forum/showthr...&threadid=3501

Something like:

152700-152800 maniacken [complete]


At that point, if you liked sieving, reserve your second range......

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


As far as the time you have to return the results, I do not think that there will be a problem as long as we hear from you that you're working on the range (either through your posts or through the factors you submit). But if you want to feel comfortable at it, try not to fall too much behind (i.e. stay at the "Lowest active ranges" for a couple of weeks).

[maniacken]

thanks nuri

i just tried 152700-152701 for kicks. i dont think i got any results.

pmax=152701000000000
pmin=152700000000000
pmin=152700337782791
pmin=152700686696489
Done

is this what i return


i will probably reserve a range and try this tomorrow after i finish the sob Factoring test.

[hc_grove]

It sounds very realistic that you haven't found any factors in a range that small.
If you got any results they would be in the fact.txt file, so take a look there (if you haven't found any factors yet, the file probably doesn't exist).

As you found no factors nothing should be submitted through the 'Sieve Result Submission' page (http://www.seventeenorbust.com/sieve/), so all you have to do is post

152700-152701 maniacken [complete]

to the coordination thread.

As there is a seperate P-1 factoring subproject, you shouldn't call the main project's test for 'sob Factoring test', the normal term is prp test.

EDIT:
Now I can see in the coordination thread that you reserved the 152700-152800 range. That means that you shall not post to the coordination thread until that range is done (except if it takes a lot longer than expected).

[maniacken]

So far 152700 - 1527220 complete 1 factor found so far.

just wondering what i return.


pmax=152800000000000
pmin=152700000000000
pmin=152700008185062
pmin=152700148838116
pmin=152700476719121
pmin=152700804661277
pmin=152701042571382
pmin=152701053397746
pmin=152701381902341
pmin=152701709844487
pmin=152702037524489
pmin=152702141000946
pmin=152702469013519
pmin=152702796431429
pmin=152703124897853
pmin=152703184806816
pmin=152703230599648
pmin=152703557697541
pmin=152703884853253
pmin=152704212009023
pmin=152704539951127
pmin=152704867106867
pmin=152705194524697
pmin=152705521942531
pmin=152705849884681
pmin=152706177302557
pmin=152706505244677
pmin=152706832924673
pmin=152707160080399
pmin=152707487236163
pmin=152707815178253
pmin=152708143120453
pmin=152708471062577
pmin=152708798742529
pmin=152709126422551
pmin=152709453840443
pmin=152709780996121
pmin=152710107889711
pmin=152710435569673
pmin=152710763249669
pmin=152711090929711
pmin=152711418085493
pmin=152711744978953
pmin=152712073183261
pmin=152712401387561
pmin=152712729329687
pmin=152713057271809
pmin=152713385476117
pmin=152713713418271
pmin=152714041098251
pmin=152714368778249
pmin=152714696720423
pmin=152715023876167
pmin=152715351818269
pmin=152715679236103
pmin=152716006916129
pmin=152716334333963
152716345532477 | 21181*2^18351404+1
pmin=152716673024053
pmin=152717000441881
pmin=152717327859749
pmin=152717655015451
pmin=152717981909021
pmin=152718308278297
pmin=152718637006889
pmin=152718965211139
pmin=152719292891147
pmin=152719620833363
pmin=152719948251149
pmin=152720275931147
pmin=152720603611201
pmin=152720931291161
pmin=152721258971159
pmin=152721585602561
pmin=152721899651119

do i just return the factor line, or do i wait till the whole test is complete and return all the factors at once? when the test is over is there a new file created to return?

[Nuri]

All you have to submit is the factor line(s). pmin lines are just for the client not to forget where it is (in case you close and restart it etc).

As far as I can see from format of the lines you posted above, you are currently using "sobsieve_1_34" client. I would recommend you to download and use the proth sieve client (version 39. ). You can find it from the link I gave above. please also note, PIV performs better with the SSE version and other machines i.e. AMD, PIII etc. perform better with the cmov version. I guess you download all three and try them. Proth sieve v39 is roughly 30% to 40% faster than sobsieve_1_34.

Ok, where to find factors easiest.

If you prefer to use sobsieve, just doubleclick the sobsieve icon on the taskbar (@ bottom right corner of the screen - I am not sure if it's called taskbar, so it's better to explain where it is ). There, you'll see your factors lined up (no pmin lines. )

If you prefer proth sieve, you'll see that your factors are accumulated in a different file named fact.txt under the folder you are running the client (you will also see factexcl.txt and factrange.txt files. you do not need to do anything with them. Just leave these two as is). If you prefer to use proth sieve, I would also recommend you download http://n137.ryd.student.liu.se/sobistrator_118.zip . It will make things much easier.

And a last note on when to submit the factors. Generally speaking, you do not need to hurry much. Once a week or so is just fine. But in case you see a factor that is very close to the next k/n pair that will be assigned for testing, you might prefer to submit your factors so that you may save a test for the project that is normally very close to being assigned (and gain more sieve points immediately). You can see what k/n pair will be assigned next from the link: http://www.seventeenorbust.com/sieve/next.txt . Generally speaking, if the n of the factor you found is within "next to next+200,000" range, then you can consider it close.

[maniacken]

ok i downloaded the 039 cmov version and the sobistrator. hopefully this will run faster than 364,000 p/s.

i found one factor. does this mean that 11 less prp tests need to be done in the future?

thanks for the help

kenny

[Nuri]

Unfortunately not.

152716345532477 | 21181*2^18351404+1 means you found a factor for the k/n pair where k= 21181 and n=18351404. You can read it as 152716345532477 divides 21181*2^18351404+1


It might save be 0, 1 or 2 tests, based on a couple of parameters like; is it unique or duplicate, is the n value of your factor above or below the n of next k/n pair that will be tested, and/or will we find a prime for that particular k before or after that n range is prp tested, etc.

[maniacken]

i found a factor that is out of range

152725749022051 | 33661*2^22512792+1

do we discard really large factors. also do we discard factexcl factors.

i really like the speed boost, usually 480000-490000 p/sec.

[Keroberts1]

out of range factors will be saved in a different folder. Although these factors have no use to us at the moment at some time in the future the sieve effort will be altered to sieve values above 20,000,000 because we will eventually pass that point. Once that day comes these factors will be much more useful. At the same time however you need to remember that when this day comes most of the original sieveing will be redone by newer and possibly much faster computers. So for that reason these factors are not incredibly valuable. Still they are worth keeping.

[maniacken]

thanks keroberts.

just an update about 40G done so far and i found 2 factors, about 30 duplicate factors and 2 out of range factors. this sieving seems fun, and ill have to reserve another range in the near future.

just wondering how long does it take to show up in the stats. i am wondering if i am returning factors correctly, since i returned the two i found and cant find my name in the stats.

[Mystwalker]

The stats get updated every 6 hours.
Possibly MikeH has to map your SoB name to a name in his stats...

[maniacken]

ok mystwalker

60% complete 4 factors.

i got a wierd file that is showing up. it is named ~ and is 173kb

ill post a little bit of it

œËË�uÒ‘X ÀOyV¤  _  ¤
Є D D
Є � Є ä!
Є ´¦
Є „+
ô
 ´
¤    À F €
3 2 8 5 4
€
3 2 8 5 5 €
3 2 8 5 6 €
3 2 8 5 7 €
3 2 8 1 2 then the number starts increasing by one.

next line

M a i n I d e n t i t y ' s C o n t a c t s 
x

ô

hopefully someone has had this happen before.

[maniacken]

just did some digging on the forum and help site. the stats i could previously get to were dated dec 16th.

to MikeH

three of the four factors for the anonymous accout were in my range (sorry i didnt log in). just giving you a heads up. the three lowest scoring ones.

[maniacken]

do i do anything with stats.txt


Statistics:
pmin : 152725225216072
pmax : 152800000000000
# Tested p : 2289480483
# Tested k : 25184285313
# Whacked k : 20504050870
# Whacked p : 704606539
Total time: 154471297 ms
484kp/s.

also what does wacked and tested mean.

[mklasson]

Nah, stats.txt is just there for kicks, statistics, and debugging.

"# tested" is the number of k and p values that were processed. For SOB the k count is 11 times the p count. "#tested p : 10" means that 10 primes were checked and the k*2^n+1 values that they divide were found.

"#whacked k" is the total number of k values that were eliminated in some way or another prior to the baby-steps-giant-steps phase.

"#whacked p" is the total number of p for which all k were whacked and for which consequently no b-s-g-s were done at all.

High whack counts are good.

[maniacken]

thanks for the info.