Distribute computing of reactors

[LordXaosa]

Hello.
Currently I made a small program, that calculates different random reactor sets, and count best power efficiency. I think it is possible to use it in distribute calculations, combine it with web server, which would manage best results, and everyone could find best reactor sets for his structure.
I'm planning to make it on C# for Windows, and may be later on Java for Linux and MacOS, on demand.

Of course all would be open source.

I could do it by my own, any help appreciated, but I'd like to know, will it be useful or not?

UPD: First version of program here.
Sources: https://github.com/LordXaosa/ReactorBreeder

UPD 3/19:
- Fixed crash if continuos chosen
- Fixed fonts for Windows 7
- Fixed error in formula, now it count groups right.
- Added energy per block info.
- Fixed bug, when you change parameters, but program still writes old results, if it was worse then with new set.

UPD 3/23:
- Added performance info (cycles per second)

UPD 3/28:
- Added displaying of free blocks and possible shield capacity or recharge if replace free blocks with SC or SR.

 

[Saxopi]

For people who are very meticulous and specific about the power outputs of their reactors, yes.

It could also be useful for designing small reactors for smaller ships, given how very limited the space is based on design.

 

[LordXaosa]

For people who are very meticulous and specific about the power outputs of their reactors, yes.

It could also be useful for designing small reactors for smaller ships, given how very limited the space is based on design.

Yes, I know this, but I'd like to know, are there any of these people
Coding for nobody is so sad...

 

[Megacrafter127]

may be later on Java for Linux and MacOS, on demand.

demand++;

combine it with web server, which would manage best results, and everyone could find best reactor sets for his structure.

I'm all for it, if there is an option to not use the webserver at all.

You also forgot another option in the poll: "Yes, I will use it and will compute offline, without using the computation-distribution feature"

 

[LordXaosa]

You also forgot another option in the poll: "Yes, I will use it and will compute offline, without using the computation-distribution feature"

Ok, I'll do offline version, but what's the point? With online you can get result without computing, just visit the web page, but in offline you won't help the community, but spend CPU time for nothing, 'cause you will not get best set of reactors.

 

[jayman38]

There are people out there who can use it, but don't know they need it until they try it.

 

[Megacrafter127]

Ok, I'll do offline version, but what's the point? With online you can get result without computing, just visit the web page, but in offline you won't help the community, but spend CPU time for nothing, 'cause you will not get best set of reactors.

If the internet will be cut off regardless, one does not have a choice.

 

[LordXaosa]

I'm sad, that feedback was too low, so I'd like to try to fix it.
Here is the first version of my "calculator". It supports local storing of best results, without sending it to server, and online mode, with best results sharing.
https://github.com/LordXaosa/ReactorBreeder
maxX,maxY and maxZ - it's bounding box for array. Not recommended to use more than 30x30x30. On fast computers you can calculate even 100x100x100, but it would be slow.
Cycles - how many different random sets will be tested on EACH CORE. Not recommended set values, more then 5 minutes of computing. If any crash - all data per launch will be lost.
Send best results to server - if checked - will send results to my web server
Continuos - if checked, will compute untill it's unchecked or app closed.

Compiled version for Windows attached.
.Net Framework 4.5 required.

For now I need help with webpage that could nice visualize best results.
Also I need help with random when arrays fill. Now random is not so random as I wish, and in 1M elements will set near 500k true's, it's not good...

UPD 3/19/15: Fixed crash if continuos chosen, fixed fonts for Windows 7, fixed error in formula, now it count groups right. Added energy per block info.

See attach at first post.

 

[LordXaosa]

Any feedback?

 

[Snk]

This seems like a really great project. I'd love to see the results, but I'm still a newbie at code.

 

[Megacrafter127]

Any feedback?

I downloaded it, and will give a feedback, once I decide to boot my old windows install.

 

[Jay_do]

I can go through and process some things up if you want and send them in. What exactly are cycles?
EDIT: Nevermind about the cycles, I'm stupid and didn't read.
[DOUBLEPOST=1426780043,1426779455][/DOUBLEPOST]ERROR: Its creating text files with different names but with the same info as my first test.
EDIT: I restarted the program and it seems to be working fine now.


EDIT: Maybe Im doing this wrong but wouldnt a 3Y 3X 50Z be a 3x3x50 rectangle area with a volume of 450 blocks? The program is writing it as a 25x25x50 with 8548 blocks. Ive tried restarting it several times and it still creates the same info.

 

[LordXaosa]

I can go through and process some things up if you want and send them in. What exactly are cycles?
EDIT: Nevermind about the cycles, I'm stupid and didn't read.
[DOUBLEPOST=1426780043,1426779455][/DOUBLEPOST]ERROR: Its creating text files with different names but with the same info as my first test.
EDIT: I restarted the program and it seems to be working fine now.


EDIT: Maybe Im doing this wrong but wouldnt a 3Y 3X 50Z be a 3x3x50 rectangle area with a volume of 450 blocks? The program is writing it as a 25x25x50 with 8548 blocks. Ive tried restarting it several times and it still creates the same info.

Thanks for your reply. About error - I fixed it, redownload pls.
About block fill - it will fill random amount of blocks and will show only LAST BEST result. So after some cycles it may not changed, because program didn't found better set.

 

[LordXaosa]

I need MOAR feedback! Or may be bugs, feature-requests... Anything?

 

[Jay_do]

How exactly is the power generation generated? Is it the volume of the space divided by the total power gen? Or is it the total amount of power generators divided by power generation?

EDIT:I wrote those formulas backwards.

 

[LordXaosa]

How exactly is the power generation generated? Is it the volume of the space divided by the total power gen? Or is it the total amount of power generators divided by power generation?

Full info you can find here: http://starmadepedia.net/wiki/Power_Systems

 

[Jay_do]

What I mean is what is your programs definition of best power generation. Is it just simply the highest or is it the best ratio of power generation to a given volume.

 

[LordXaosa]

What I mean is what is your programs definition of best power generation. Is it just simply the highest or is it the best ratio of power generation to a given volume.

Program tries to find most powerfull set at the same volume. It doesn't calculate efficiency like "ea per block" or something.

 

[Jay_do]

Hmm alright. I might bug test it some more later. I am currently not near mt computer.

 

[LordXaosa]

Hmm alright. I might bug test it some more later. I am currently not near mt computer.

Any feedback?