The New Sudoku Players' Forum

[champagne]

I am running that test using exclusively public tools described in the project pointed 2 posts above

I see nothing "public" in the project you cite.

you have the sources, you have the projects definitions, what more is needed to be public ???

How about *.cpp, *.h, etc. files in a *.zip or other compressed format so that people don't have to be familiar with the "svn" tool? [I think we had a very similar discussion for skfr.] Otherwise, how about hints on how to obtain and use our own copies of "svn"?

BTW how about makefiles for the compiler(s) supported?

I answer to that post here where the previous discussion took place;

The overall conditions are the same : a project in development located in what specialists (I am not) consider as the appropriate structure, a subversion manager.

This is the code on which I am working and the project is somehow the safety copy of my code.

The right way to be in line with the code is to get locally a subversion manager. I am using "Tortoise svn", a free program and this is also the subversion manager chosen by other contributors to the skfr project.

As in skfr project, it is possible, from time to time, to build a zip file collecting the necessary files for compilation, the doc and the .exe files.
We did it twice so far for skfr.

A similar frequency can be expected here. I intend to do it for the first time as soon as the basic documentation will be ready if potential users ask for it, but I strongly recommend to those who are willing to follow that project to use the svn support.

Regarding the makefile, it is implicit in the project definition. May be in other platforms an explicit makefile has to be defined.

Some discussions took place earlier on portability in that thread, The same problems should come with that project using another platform that Microsoft Visual C++

[ronk]

I am using "Tortoise svn", a free program and this is also the subversion manager chosen by other contributors to the skfr project.

That part was surprisingly easy. I was able to install "Tortoise svn" and checkout the "sudoku-multi-purpose-program" source code. AFAIK this is the first time your use of "Tortoise svn" has been posted publicly.

As in skfr project, it is possible, from time to time, to build a zip file collecting the necessary files for compilation, the doc and the .exe files.
...
Some discussions took place earlier on portability in that thread, The same problems should come with that project using another platform that Microsoft Visual C++

However, since I don't have Visual C++, I'll have to think about how to proceed from here. May just wait to see if a *.exe file (for Windows) becomes available.

[champagne]

However, since I don't have Visual C++, I'll have to think about how to proceed from here. May just wait to see if a *.exe file (for Windows) becomes available.

The answer is much depending on what you are expecting and what is your current use of C++.

I switched from Borland Ide to C++ to have a development on a free platform.

Visual C++ is as easy to install as Tortoise SVN. I am using the 2010 version. Mladen noticed some small problems using an older version.

The main advantage installing the Visual C++ Ide is that the updated code will compile without any problem.

Another advantage to have the local copy is that the doc files are easily loaded in the net browser.
I think a quick overview of the doc can help to see how you can have some benefit from that code.

I will rearrange my priorities to be in a position to have the first .zip file ready by the end of next week.

champagne

PS If you decide to use Visual C++, I can prepare some comments to have a good start with the project

[dobrichev]

Hello,

I updated the skfr project (rating engine only).

Source code is ported to GNU C++ compiler for x86 processors.
There are no changes in the rating process since last modifications made by Champagne in April 2012.
There are major changes in architecture, performance, and rating results from the previous binary release from November 2011.
There could be last-minute bugs introduced while I merged the code from different machines and platforms so any feedback is welcome.

Windows 32 and 64-bit executables are available for downloading.
All source files + executable for 64-bit Linux are also available in a single zip file.

There are no plans for further major changes in the rating algorithms.
Maybe some restructuring will be done to ease access from script languages.
The progress in the rating algorithms made by Champagne in his multi-purpose program (the more recent project discussed in this thread) eventually could be merged at some point.

[champagne]

Hi all,

I made preliminary tests on the 64 bits version uploaded by Mladen
(I can not generate 64 bits applications with the free version of microsoft visual C++)

Some warnings to potential users:

The first version has been replaced after a bug has been fixed
This is a multi thread version. Any user would be wise to limit the number of threads authorised.
To improve the overall performance, Mladen has organised a process by blocs of 1920 puzzles. The print is done at the end of each sub lot.

If the lot is made of puzzles with high ratings, the delay to get the first print can be of several minutes (tenth of minutes in my test)

I have no yet made tests on the performances compared to my 32 bits version

[coloin]

ok
sorry if im out of line here.......

ive downloaded the executable - realised i need to use it in c++

ive made a virtual turbo "dosbox"

now i need to "compile" a cpp and run it

ive downloded visual C++ express [unfortunately its in french - but hey i dont mind]

i am probably stuck now and dont want to give up just yet

Thanks in anticipation

C

[champagne]

Hi coloin,

dobritchev has prepared the three zip files, so he will better answer to your point.

There are several points to clarify to understand the problem.

a) there are 3 zip files.
skfr_win32
skfr_win34
skfr_GNU_LINU
which one do you have

b) on which computer do you try to run it and what is your Operation System

c) you write "I need to compile it"

the three pack are executables.

If you simulate windows, the process would work in any of the following conditions
c.1) you put the .exe in the directory window\system32
c.2) you put the .exe in the directory where you want to run it

If you rename the .exe you have downloaded in "skfr.exe"
in the appropriate directory, you start a console application and enter the command

skfr -imyfile

regards

champagne


EDIT just 2 more pieces of information

the three downloads prepared by mladen are designed for a multi thread execution.
to control the number of threads, the safer way for me is to use 2 lines of command (here for 2 threads)

set OMP_NUM_THREADS=2
skfr -imyfile


The design is made to optimise the run time for big lots.
The first print can come in block after several minutes with lots exceeding 2000 puzzles

[coloin]

Thanks
i started of by putting the extracted win32 exe file in a directory and renamed it to skfr32, running windows 7.
from the dos prompt in the directory - typing skfr32 -ifile.txt as i do with other programs but it didnt work

c:\Suxx>set OMP_NUM_THREADS=2

c:\Suxx>skfr32 -idobhard0.txt
Version: V2.0.1 dated October 16th 2012
summary of parsing
problem in open input

my file was a series of valid 81-string puzzles

EDIT
It works
i had to lose the ".txt"
WOW

C

[dobrichev]

Hi coloin,

I appreciate your interest in this tool.

Just want to add that limiting the concurrent threads to 2 is campagne's personal choice. You can read some directions on controlling the parallelism and overheating in this post.

An open issue is whether 64-bit version works faster than the 32-bit one. If you are using 64-bit windows, you may compare the speeds.

[ronk]

EDIT
It works
i had to lose the ".txt"
WOW

IMO the user should have complete freedom in naming files. Requiring a file name at all implies that stdin is not being used, which prevents piping of results from another program.

[dobrichev]

IMO the user should have complete freedom in naming files. Requiring a file name at all implies that stdin is not being used, which prevents piping of results from another program.

Agree.
Such changes have been postponed during the active development phase and maybe now these defects should be fixed.

[coloin]

Yeah i feel stupid now .... but i didnt take the instructions literally enough

except in red in the first post in this thread

The executable file will not work if the runtime microsoft visual C++ express has not been dowloaded before.
Unhappily, I have no tool to produce directly an executable including the runtime components.

this statement has precluded me from trying the program for a long time.....
so i didnt need to do this then !!!!

i have downloaded the -64 but it said i didnt have the right processor etc

i have a dual core so suppose the setting to 2 is correct.
How do you type another command line without entering the "new line/enter" key and starting the command ? !

C

[dobrichev]

The latest windows executables don't depend of external libraries (except windows kernel).

[champagne]

i have a dual core so suppose the setting to 2 is correct.
How do you type another command line without entering the "new line/enter" key and starting the command ? !

C

Hi coloin,

1) I am no a professional in EDP and at the moment I wrote the red comment, I had not the right linking options.
As said just above, mladen cleaned the situation .

2) Mladen gave the link to the place where a detailed explanation is given for the thread issue.

I wanted just to send a warning because I locked my computer with a use of skfr with other batches in competition.
limiting the number of threads to 2 is equivalent to doing nothing with a dual core.

3) I don't catch your last point

champagne

[coloin]

I know that now.
should have listened to

SKFR is written in C++ and doesn't use any specific API from Windows.

except i believed you !

It is fact that the current code is not portable. The compiler/OS specifics are not so much, they are isolated, and it is matter of few hours to port the code to other modern compiler.

of course that was some time ago and i really didnt appreciate the efforts you both made ! Thank-you now.
anyhow what matters is .....
the program works in window dos box .......just type skfr -ifile and press enter

C