Srsieve Application failed

[PawnProver44]

I was installing the executable file for Srsieve and when I tried to run it a command window came up for just a split second and left? I also used a Srsieve format file and tried to run it using Pfgw but did not recognize it. Does anyone know what went wrong? (or if that was normal)

[paulunderwood]

You need to open a terminal. I am not familiar with Windoze. I think you need to open the run dialog box and issue the command cmd. Alternatively, search for a terminal launch in the menus. Once you have a terminal open you will need to change directory with the cd command entered into the terminal like this cd \blah\blah\srsieve_directory and then run srsieve as per the manual.

As regards to the PFGW file format problem: we can cross that bridge when we get to it.

[PawnProver44]

Originally Posted by paulunderwood

You need to open a terminal. I am not familiar with Windoze. I think you need to open the run dialog box and issue the command cmd. Alternatively, search for a terminal launch in the menus. Once you have a terminal open you will need to change directory with the cd command entered into the terminal like this cd \blah\blah\srsieve_directory and then run srsieve as per the manual.

As regards to the PFGW file format problem: we can cross that bridge when we get to it.

Well see, I extracted the zip folder to a different folder and does not specify the path found, and after that I need an input file and I don't know how to make srsieve sieve it just yet.

[LaurV]

- unpack the exe somewhere
- for the purpose of this tutorial, we will assume it is called "srsieve_xx.exe", this is because it usually has a complicate name which says the version and the OS (operating system) supposed to work on, 32 or 64 bits, etc. I usually rename it "srsieve.exe" or simpler "sr.exe", which helps me later when I play with it. You can do the same (or not) if you like, this will be used in the last steps below.
- use notepad (or your text editor at choice) to create a file, say "mysieve.txt", with the content on a single line "60*79^n+19", without quotes, and type an enter at the end (carriage return), srsieve won't like lines which don't end with CR or LF.
- put this file in the same folder/directory as the exe file you just unpacked from srsieve package.
- if you don't have a "normal" file browser (like "total commander" from ghisler, or so) which offers a command line, then open windows explorer and navigate to the folder where you have unpacked the exe file and you have created the "mysieve.txt" file.
- click on the explorer bar on top (that where you see the path to the current folder) and delete everything there, and type "cmd" instead (the path is already selected when you click on the empty space of the path line, so you don't need to press delete, only need to type "cmd", without quotes)
- press Enter - this will open a command prompt window with the path already in the folder, so you don't need to type any "cd" (change directory) commands.
- type "dir" (without quotes) and enter, in the command prompt window to make sure you have the two files there. If not, you are doing it wrong.
- assuming the name of you exe is still "srsieve_xx.exe", type at the command prompt (in the black window) "srsieve -h" (without quotes) to get a help about the commands srsieve will accept. This is optional, but as other people said before, you want to be proficient, then "read the fucking manual!".
- type (without quotes) "srsieve_xx -P 1e9 -n 1e3 -N 1e5 mysieve.txt" (you can copy paste from here, if you right click on the top of the command window and say edit/paste - this should work in all versions of windoze, or effectively type it, it will not hurt you! Be careful to the spaces and case sensitivity, srsieve is case sensitive)
- press enter
- enjoy

Explanation: This command will sieve your 60*79^n+19 from n=1000 to n=100000, with primes from 2 to 1000000000, it only takes few minutes, and it will eliminate ~91000 candidates, letting only about 8000 (from the 99000 sieved). It will create an output file which is called "srsieve.out" or something like that, which you can rename later, and use it as input to sieve higher.

- to sieve higher, say to 2000000000 prime (assuming you renamed the output file to "mysv2.txt") you use "srsieve_xx -P 2e9 mysv2.txt", without quotes.

- you should use some primality test - pfgw can do prp test and eliminate smaller candidates very fast, so you can manually edit your input txt files and eliminate them, etc, which will speed up sieving. You always sieve a bit, prp test a bit, sieve a bit, prp test a bit, etc, do the one which eliminate candidates faster. At the end, you have a (very short) list with PRPs which you can prove prime if you like. Much faster.

edit: and read those instructions/manuals, the srxxxxx suite can do much better things, and pfgw is also very complex, you even can write your own scripts in it. Learn a bit about ABCx formats, etc.

[PawnProver44]

Originally Posted by LaurV

- unpack the exe somewhere
- for the purpose of this tutorial, we will assume it is called "srsieve_xx.exe", this is because it usually has a complicate name which says the version and the OS (operating system) supposed to work on, 32 or 64 bits, etc. I usually rename it "srsieve.exe" or simpler "sr.exe", which helps me later when I play with it. You can do the same (or not) if you like, this will be used in the last steps below.
- use notepad (or your text editor at choice) to create a file, say "mysieve.txt", with the content on a single line "60*79^n+19", without quotes, and type an enter at the end (carriage return), srsieve won't like lines which don't end with CR or LF.
- put this file in the same folder/directory as the exe file you just unpacked from srsieve package.
- if you don't have a "normal" file browser (like "total commander" from ghisler, or so) which offers a command line, then open windows explorer and navigate to the folder where you have unpacked the exe file and you have created the "mysieve.txt" file.
- click on the explorer bar on top (that where you see the path to the current folder) and delete everything there, and type "cmd" instead (the path is already selected when you click on the empty space of the path line, so you don't need to press delete, only need to type "cmd", without quotes)
- press Enter - this will open a command prompt window with the path already in the folder, so you don't need to type any "cd" (change directory) commands.
- type "dir" (without quotes) and enter, in the command prompt window to make sure you have the two files there. If not, you are doing it wrong.
- assuming the name of you exe is still "srsieve_xx.exe", type at the command prompt (in the black window) "srsieve -h" (without quotes) to get a help about the commands srsieve will accept. This is optional, but as other people said before, you want to be proficient, then "read the fucking manual!".
- type (without quotes) "srsieve_xx -P 1e9 -n 1e3 -N 1e5 mysieve.txt" (you can copy paste from here, if you right click on the top of the command window and say edit/paste - this should work in all versions of windoze, or effectively type it, it will not hurt you! Be careful to the spaces and case sensitivity, srsieve is case sensitive)
- press enter
- enjoy

Explanation: This command will sieve your 60*79^n+19 from n=1000 to n=100000, with primes from 2 to 1000000000, it only takes few minutes, and it will eliminate ~91000 candidates, letting only about 8000 (from the 99000 sieved). It will create an output file which is called "srsieve.out" or something like that, which you can rename later, and use it as input to sieve higher.

- to sieve higher, say to 2000000000 prime (assuming you renamed the output file to "mysv2.txt") you use "srsieve_xx -P 2e9 mysv2.txt", without quotes.

- you should use some primality test - pfgw can do prp test and eliminate smaller candidates very fast, so you can manually edit your input txt files and eliminate them, etc, which will speed up sieving. You always sieve a bit, prp test a bit, sieve a bit, prp test a bit, etc, do the one which eliminate candidates faster. At the end, you have a (very short) list with PRPs which you can prove prime if you like. Much faster.

edit: and read those instructions/manuals, the srxxxxx suite can do much better things, and pfgw is also very complex, you even can write your own scripts in it. Learn a bit about ABCx formats, etc.

Thanks! You are very helpful! Everything you said worked, however gave me this Error:

ERROR: Line 2: Malformed Line in candidate sequence format file 'mysieve.txt'

How is that fixable?

[LaurV]

Ignore it, if it was sieving properly.

Your file should have a single line, ended with a CR, like this.

60*79^n+19

Some text editors add a ctrl+Z or other "end of the file" character, which you have in the second line, but that won't influence the sieving, each line is sieved separately. You can add more lines into the file, with different k and c, or learn the ABC format.

[PawnProver44]

Originally Posted by LaurV

Ignore it, if it was sieving properly.

Your file should have a single line, ended with a CR, like this.

60*79^n+19

Some text editors add a ctrl+Z or other "end of the file" character, which you have in the second line, but that won't influence the sieving, each line is sieved separately. You can add more lines into the file, with different k and c, or learn the ABC format.

If I am supposed to ignore that, then where are my results going to?

[LaurV]

Did you see any numbers running on screen?
Did a file called srsieve.out was created? There are your results. It looks like:

pmin=1000000000
60*79^n+19
1011
1027
1033
1035
1041
1048
1053
1071
1086
1089
1122
1125
1126
1128
1176
1179
1195
1201
1216
1221
1242
1245
1267
etc...

If not, then is there a file called srsieve.log? (errors and history/log messages)
What does it say?

Did you use -h switch and maybe found out about -o switch and renamed the output?

How about using a GOOD text editor and create a file without garbages in the second line? (just open notepad, copy/paste the line from my previous post, save with your favourite name?) Then run again the sieving?

[PawnProver44]

There is an .out file and has the list of exponents in that range. So I am just supposed to manually PRP test those ones (not on the file) from there, right?

[paulunderwood]

Copy the ".out" file to input1.txt (or something else) and, in the copy, precede the list of numbers with ABC 60*79^$a+19 instead. Then you can run with the new file directly in PFGW.

[PawnProver44]

I am still unsure weather the numbers listed in the .out file are eliminated candidates or not.