Draft proposed standard for Mersenne hunting software neutral exchange format

[M344587487]

Originally Posted by chris2be8

Because sooner or later they will find themselves on a box with vi as the only editor that's been installed. If they are really unlucky the only editor will be ed (or something even less friendly). BDDTGTTS.

Chris

I'd argue if you're going to become a sysadmin (or whatever) and haven't taken the initiative to familiarise yourself with the core tools you're likely to use you deserve to learn the hard way. I don't really care and shouldn't be arguing the point but I guess this shows I have too much free time today :P

[Dylan14]

I have used the recordtest.txt provided in post 34 to test the other popular text editors in Linux. Attached are the results from emacs (25.2.2), vim (8.0.1453) and nano (2.9.3). These were tested on Ubuntu 18.04 "Bionic". Note with nano the output is the result after soft wrapping is enabled as by default it is not and the first line goes well beyond the limits of the terminal window.

Also attached is the test file after resaving in nano, which shows a change in file size (560 vs. 563). Not sure why this is happening.

[kriesel]

Originally Posted by Dylan14

Also attached is the test file after resaving in nano, which shows a change in file size (560 vs. 563). Not sure why this is happening.

Records 8 and 9 don't say so, but as I posted the file, they had the same CR LF termination as record 7. Nano is probably quietly deleting those CR's at save time.

Originally Posted by Dylan14

I have used the recordtest.txt provided in post 34 to test the other popular text editors in Linux. Attached are the results from emacs (25.2.2), vim (8.0.1453) and nano (2.9.3). These were tested on Ubuntu 18.04 "Bionic". Note with nano the output is the result after soft wrapping is enabled as by default it is not and the first line goes well beyond the limits of the terminal window.

Also attached is the test file after resaving in nano, which shows a change in file size (560 vs. 563). Not sure why this is happening.

In nano if the line goes beyond the terminal window this is indicated with a $ and you can scroll the line

[kriesel]

I find myself using export code more often to peek into the application interim file contents, than to do export.

A useful addition, not to the draft standard, but to export software behavior, is to take the least significant 8 bytes of the binary data and output to the console the properly human-readable ordered hexadecimal representation.

In the data, the bytes are 01234567 (least significant first).

The res64 hexadecimal output is MSB first, 0x7766554433221100, so that it is directly comparable to what shows up in application console output or logs (such as CUDALucas, CUDAPm1, etc).

[kriesel]

Revised the draft standard to draft 5. See especially a few numbered questions/comments in red text in the attached pdf file.
Constructive comments are encouraged.

Sample program run:

C:\Users\Ken\Documents\starfish>cudapm1export.pl

CUDAPm1export for Windows V0.1 2018-08-15

Input file name for export please (output will be to same name, type .mne.txt):
c312000049s1.
Input file name received as <c312000049s1.
Extracted exponent q 312000049 from filename
Opened file c312000049s1 for read
Length of input file data read is 39000108 bytes, 9750027 words.
Exported file's header will read as follows:

Format Mersenne Neutral Exchange V1.0
FileOrigin "CUDAPm1export for Windows" "V0.1 2018-08-15" c312000049s1 2018 Aug 15 23:12:33 UTC
Type P-1 stage 2
Exponent 312000049
Iteration 3598983
N 17920000
AccumulatedTime 156999
B1 2495000
Reserved6 0
Reserved7 0
Reserved8 0
Reserved9 0
B2 0
D 0
E 0
NRP 0
M 0
K 0
T 0
Midpasstransforms 0
Itran_done 0
PtrandonePlusNumtran 0
Itime 0
Ptime 0
Reserved22 0
Reserved23 0
Reserved24 0
DataFormat binary bytes
CRC32 0x5660f8fc
DataBinaryByteCount 39000008
EndOfHeader 0x73503f16

Exported file header has length 547 including record terminators.
Processing bulk binary data input without output to console, please wait until program indicates Done.

Res64 of binary form of data is: 0xd1f1908c8f75b4c0

Printing 39000008 bytes bulk binary data in binary form.
Done.