Wacky residuals thread

[mdettweiler]

Dug back through a PRPnet client log to get the following:

71301*6^123107+1 is not prime. Residue cecaaec90ac80ebe
68195*6^123738+1 is not prime. Residue bdaaeddeea51b031
59506*6^123939+1 is not prime. Residue babe8b8cdcbe5aac (an actual valid English word in this one!)
139413*6^122112+1 is not prime. Residue aaaa20b029981508
14505*6^120696+1 is not prime. Residue 5beaf226b0fc0f87 (not quite an actual English word, but still somewhat phonetical)
14505*6^120636+1 is not prime. Residue 4beeb88b09f0dabc (similar to the previous one, and was in fact right before it in the results file)
14505*6^117824+1 is not prime. Residue fceedfd85f5a9512 (do misspellings of real words count? lol)
14505*6^104684+1 is not prime. Residue daaaf3369a2c806c
95407196*3^75574-1 is not prime. Residue a3f1d7b4decada93
85837064*3^75436-1 is not prime. Residue cbc9deeccbf1f2d8
95412274*3^75276-1 is not prime. Residue ebdbe9f610251c18
97221374*3^75063-1 is not prime. Residue cb95795d264ccaaa
10107*6^133607+1 is not prime. Residue 5bfacbd4a3e96f22 (sooooo close!)
33706*6^122254+1 is not prime. Residue 604bdcddbbac16a3
81976618*3^75942-1 is not prime. Residue 1680529000367096 (not a single letter in this one!)
68443118*3^75997-1 is not prime. Residue e1eadd3662c72dcf (this one makes a valid English word, if you count "1" as an L)
59077924*3^75177-1 is not prime. Residue b2aaeecdd75ab12d
10463066*3^75742-1 is not prime. Residue 873beedcd9272fc5 (redux of "do misspellings count"? )
6878756*3^75655-1 is not prime. Residue aa215403bbbbdbc5

Granted, most of the above do not actually contain real English words, but many do contain "words" that can be phonetically pronounced, as well as a number of interesting letter combinations.

[MooooMoo]

Originally Posted by gd_barnes

Unless you're already a vegetarian, that would be quite a life change. (Meee's thinks he may already be a vegetarian so wouldn't be too hard; lol)

I'm not a vegetarian and have never been one, but keep in mind that I'll only give up beef if you find that residue. As Mini-Geek posted earlier, chicken, pork, lamb, fish, dairy products, etc are all fair game, so it wouldn't be that much of a lifestyle change for me.

I may start mulling through my gobs of results in the next few days and see what I come up with

Good luck if you're trying to find that "BEEF15BAD" residue, you'll need it

[kar_bon]

i've searched all results from NPLB LLRnet-servers so far.

i'v found the lowest and highest residue:
user=gd_barnes
[11/10/08 23:44:04]
543*2^544923-1 is not prime. Res64: FFFFFBAC20D7DA76 Time : 477.0 sec.

user=Anonymous
[06/14/08 07:40:15]
423*2^212242-1 is not prime. Res64: 000006AEE57490C4 Time : 1160977.0 sec.

there're some others like for smaller n:
746841*2^9995-1 is not prime. LLR Res64: 0000030DDC1BBD95 Time : 239.956 ms.

the smallest residue (for a low n) i found:
790717071 > 2^14, so we can only do a PRP test for 790717071*2^14+1.
790717071*2^14+1 is not prime. RES64: 0000000000000051. OLD64: 00000000000000F0 Time : 12.089 ms.

and here's another thing i have to discuss:
4259877765 > 2^4, so we can only do a PRP test for 4259877765*2^4+1.
4259877765*2^4+1 is not prime. RES64: 00000000000002D9. OLD64: 0000000000000888 Time : 93.317 ms.
4259877765 > 2^10, so we can only do a PRP test for 4259877765*2^10+1.
4259877765*2^10+1 is not prime. RES64: 00000000000002D9. OLD64: 0000000000000888 Time : 124.524 ms.
4259877765 > 2^16, so we can only do a PRP test for 4259877765*2^16+1.
4259877765*2^16+1 is not prime. RES64: 00000000000002D9. OLD64: 0000000000000888 Time : 111.452 ms.

all 3 residues for one k (n=4, 10 and 16) are the same! why?

and many other same residues found for other k's but no others (except the one above), always $2D9! why?

[kar_bon]

Originally Posted by Mini-Geek

The problem is that I have many (dozens, maybe a hundred) lresults files with only four different names in many folders. Unless the *.txt could also be told to do all *.txt in all subfolders, this doesn't help very much.

try this:
make a batch-file (say cp.bat) with the following line in it:

for /f "delims=" %%i in ('dir lresults*.txt /b /s') do @type "%%i" >> all.txt

explanation:

you're in DOS-box in the directory with all subdirs in it with all your result-files, say

[dir1]
lresults.txt lresults1.txt lresults2.txt lresults3.txt
[dir2]
lresults.txt lresults1.txt lresults2.txt lresults3.txt
[dir3]
lresults.txt lresults1.txt lresults2.txt lresults3.txt

the "(...)" list all files called 'lresults*.txt' in all subdirs (-> dir1,dir2,dir3).
the for-statement goes through this list of files and type all in a new file called 'all.txt'.

so all your results are in one big file with only the new batch-command 'cp'!

hope this helps!

[henryzz]

Originally Posted by kar_bon

i've searched all results from NPLB LLRnet-servers so far.

i'v found the lowest and highest residue:
user=gd_barnes
[11/10/08 23:44:04]
543*2^544923-1 is not prime. Res64: FFFFFBAC20D7DA76 Time : 477.0 sec.

user=Anonymous
[06/14/08 07:40:15]
423*2^212242-1 is not prime. Res64: 000006AEE57490C4 Time : 1160977.0 sec.

there're some others like for smaller n:
746841*2^9995-1 is not prime. LLR Res64: 0000030DDC1BBD95 Time : 239.956 ms.

the smallest residue (for a low n) i found:
790717071 > 2^14, so we can only do a PRP test for 790717071*2^14+1.
790717071*2^14+1 is not prime. RES64: 0000000000000051. OLD64: 00000000000000F0 Time : 12.089 ms.

and here's another thing i have to discuss:
4259877765 > 2^4, so we can only do a PRP test for 4259877765*2^4+1.
4259877765*2^4+1 is not prime. RES64: 00000000000002D9. OLD64: 0000000000000888 Time : 93.317 ms.
4259877765 > 2^10, so we can only do a PRP test for 4259877765*2^10+1.
4259877765*2^10+1 is not prime. RES64: 00000000000002D9. OLD64: 0000000000000888 Time : 124.524 ms.
4259877765 > 2^16, so we can only do a PRP test for 4259877765*2^16+1.
4259877765*2^16+1 is not prime. RES64: 00000000000002D9. OLD64: 0000000000000888 Time : 111.452 ms.

all 3 residues for one k (n=4, 10 and 16) are the same! why?

and many other same residues found for other k's but no others (except the one above), always $2D9! why?

a list up to n=1000 made using llr's predecessor prp

4259877765*2^4+1
4259877765*2^10+1
4259877765*2^16+1
4259877765*2^139+1
4259877765*2^142+1
4259877765*2^328+1
4259877765*2^415+1
4259877765*2^583+1
4259877765*2^868+1
4259877765*2^883+1

these numbers are all prime:

(4259877765*2^4+1)/7
(4259877765*2^10+1)/7
(4259877765*2^16+1)/7
(4259877765*2^139+1)/7
(4259877765*2^142+1)/7
(4259877765*2^328+1)/7
(4259877765*2^415+1)/7
(4259877765*2^583+1)/7
(4259877765*2^868+1)/7
(4259877765*2^883+1)/7

edit:
it works the other way as well i searched for primes of the form "(4259877765*2^n+1)/7" up to n=10k and then checked the residue for "(4259877765*2^n+1)" and it was always the same
edit2: just realised that the numbers all have the same residue to each other in other prp bases as well though different to 3-prp

[thommy]

Originally Posted by kar_bon

all 3 residues for one k (n=4, 10 and 16) are the same! why?

Short answer: Well, the difference is k*2^6 between those numbers, which is k*64. As that residue is "something mod 64" a multiple of 64 doesn't make a difference.

[Mini-Geek]

Originally Posted by kar_bon

try this:
make a batch-file (say cp.bat) with the following line in it:

for /f "delims=" %%i in ('dir lresults*.txt /b /s') do @type "%%i" >> all.txt

explanation:

you're in DOS-box in the directory with all subdirs in it with all your result-files, say

[dir1]
lresults.txt lresults1.txt lresults2.txt lresults3.txt
[dir2]
lresults.txt lresults1.txt lresults2.txt lresults3.txt
[dir3]
lresults.txt lresults1.txt lresults2.txt lresults3.txt

the "(...)" list all files called 'lresults*.txt' in all subdirs (-> dir1,dir2,dir3).
the for-statement goes through this list of files and type all in a new file called 'all.txt'.

so all your results are in one big file with only the new batch-command 'cp'!

hope this helps!

It worked! I had actually looked up the for command, assuming it'd help me somehow, but I didn't know enough about it to use it, and its syntax is very confusing and hard to learn.
By the way, I have no BEEF15BAD residues. I'll look for other interesting things...
My highest:
513*2^542406-1 is not prime. LLR Res64: FFFFC30BBD0FC9A9 Time : 732.068 sec.
My lowest:
741*2^418549-1 is not prime. LLR Res64: 0000B409715D6A57 Time : 373.817 sec.

[Mini-Geek]

My longest string of a single character:
579*2^530349-1 is not prime. LLR Res64: 3A739D22C444444E Time : 744.985 sec.
Words:
639*2^530595-1 is not prime. LLR Res64: A5636380BE1005ED Time : 712.950 sec. (loosed)
897*2^540364-1 is not prime. LLR Res64: 395E64EEBA13375B Time : 714.415 sec. ("baleet"...okay that's not really a word, but it sounds close to delete)
5*2^179648-1 is not prime. LLR Res64: DEE7BBBBA1E7DFD3 Time : 50.201 sec. ("baletd"...okay that's not really a word either, but it sounds close to "deleted")

More later.

[henryzz]

Originally Posted by thommy

Short answer: Well, the difference is k*2^6 between those numbers, which is k*64. As that residue is "something mod 64" a multiple of 64 doesn't make a difference.

that would make sense for those small ks but what about ns like 883
883 is odd not even so it cant be the same mod 6
also why do they all divide by 7 and then the remaining cofactor is prime
it seems like a good way to find some primes of a odd form checking which residues match

[kar_bon]

Originally Posted by Mini-Geek

It worked! I had actually looked up the for command, assuming it'd help me somehow, but I didn't know enough about it to use it, and its syntax is very confusing and hard to learn.
By the way, I have no BEEF15BAD residues. I'll look for other interesting things...
My highest:
513*2^542406-1 is not prime. LLR Res64: FFFFC30BBD0FC9A9 Time : 732.068 sec.
My lowest:
741*2^418549-1 is not prime. LLR Res64: 0000B409715D6A57 Time : 373.817 sec.

nice to hear!

here another option to choose:

put in the batch-file the following line:

for /f "delims=" %%i in ('dir LLRnet*.txt /b /s') do find /i "not prime" %%i>>all.txt

now all lines with only the string "not prime" will be listed (and the name of the file first).
it's useful for many LLRnet result-files with 3 lines per k/n-pair in it!

[Mini-Geek]

Originally Posted by kar_bon

nice to hear!

here another option to choose:

put in the batch-file the following line:

for /f "delims=" %%i in ('dir LLRnet*.txt /b /s') do find /i "not prime" %%i>>all.txt

now all lines with only the string "not prime" will be listed (and the name of the file first).
it's useful for many LLRnet result-files with 3 lines per k/n-pair in it!

I had forgotten about my LLRnet results. Thanks again, I've got those too now.