Hi all,

A friend of mine recently dragged me into the Euler Project, anyone here is spending his nights doing it as well ?


By the way, LabVIEW is not in the default "language list"

I've been a fan of it for a while, and it's good to see someone else doing the problems with LabVIEW. Just out of curiosity, how have you been handling the questions that require (or encourage) using big integers? For example, try to find 1000! in LabVIEW.

Chris M

Yes, but it's been a while, though. I love where my head goes when I'm trying to work out one of their problems.

Here's how:
http://forums.lavag.org/index.php?s=&s...post&p=5194

You might also follow up with this link:
http://forums.lavag.org/index.php?s=&s...post&p=5190

Wow, that's some beautiful code!

LabVIEW is now in the language list.

Jaegen

Antoine,
Thank you for introducing me to this fascinating challenge. My wife would not thank you I would be interested in any suggestions to improve my code. I can get the job done, but I am sure that my LabVIEW coding skills can only go up. Attached is my solution to the Euler project chalenge #3. I would love to hear any suggestions as to how to improve it's efficiency. I am not looking for fancy mathematical insights (won't help future projects). I have a solution, it's just slow (~4.5 minutes). What I am interested to learn is how to improve the code using more efficient LabVIEW practices. Thanks for any insights!


BTW Jaegen, are you farmilliar with a company named ISE in Port Coquitlam?

This post has been edited by JiMM: Jun 3 2008, 11:57 PM

That's cool !




I hope I'm putting you into trouble .

I started the Euler projet with the same intent : improving my LabVIEW skills.
The first solution I found for problem 3 is quite slow as well, now that I saw the above linked code I will try to get inspired and improve my subVIs because it's easy to draw a solution that WILL find the solution but not untill the end of the universe !

Don't have time right now to have a look at your code, but I will.

EDIT :
I tested the prime factor VI that won the coding challenge on NI Forums, it takes 2ms to find the solution of prob 3

Cheers

I can't open you VI .
What version of LabVIEW did you use ?
Can you repost it please ?

The first posted version was in LV8. I now have two revised versions. The first (Euler3_Take2.vi) fixes blatent errors in my first attempt - like I eliminated testing of even numbers (DUH!). That dropped execution to ~2.3 minutes. The third version (Euler3_VJT.vi) uses a theory that to test N for primality you only need to check for divisibility by all of the primes up to N/2. Not the route I was really looking for because it relies on number theory rather than programming efficiency, but interesting to try. That one dropped execution time to 1.2 minutes. And used up my lunch hour rather than a night The attached VIs are in LV7.1.

This post has been edited by JiMM: Jun 4 2008, 09:48 PM

Math efficiency is going to be a large part of keeping the speed down. A slow mathematical algorithm will be slow no matter have programmatically efficient you make it.



I don't know if my wife would appreciate me starting up. It does look fun though.

It is interesting to both solve the problem, and then say "how can I do this better?" At work I to often get in the mode of solve the problem and move on. If it works, the boss doesn't care (or notice) if it's elegant, or efficient.

This post has been edited by JiMM: Jun 4 2008, 10:59 PM

Holy Mackerel! I am stunned by how fast the Coding Chalenge winner did this. Of course it was obvious that clever math tactics would be necessary to minimize time, but .... just WOW. I got less than 1ms with that VI on my PC. AMAZING! I am not worthy!!!!!!

This post has been edited by JiMM: Jun 6 2008, 01:52 AM

I had a quick look at your VI "IsPrime.vi", one thing you can improve is the number of time you run the loop to see if the tested number is prime ; it's no use to test a prime divisor if it is bigger than the square root of the tested number

EDIT :

My solution takes less than 2 secondes on my old laptop and it is not really different rom yours .
A few other advices to improve performance :
- avoid to use "build array" primitive,
- avoid cohersion dots,
- inline code.

One interesting hit I found, in my solution there is one subVI, if I inline the code to have only one VI, it runs faster ~10% ; I guess this is due to the way LabVIEW passes inputs to callees and outputs to callers.

This is fun stuff. Thanks for posting the link. I'm up to number 9! WooHoo!