[Trial]Contest 1 - VB (.NET & Legacy) - Prime Numbers

Please notice that I have added an extra part to the rules which was overlooked at first (remember this is the first contest and we haven't fine tuned the process straight off).
If your entry doesn't meet this requirement then you will be able to resubmit the entry.

See here for the rules, the chnages have been highlighted like this.
http://www.vbforums.com/showthread.p...postid=1782180


BTW work on the next contest is well underway and shouldn't need such changes like this as we plan to get the rules right before it goes live (expected date should be same day this one ends :)) ;).



Tec-Nico: his current one, the green blob :lol: :D.

Well guys, I've submitted my entry, have fun with it :D

Quote:

---

Originally posted by Electroman
Please notice that I have added an extra part to the rules which was overlooked at first (remember this is the first contest and we haven't fine tuned the process straight off).
If your entry doesn't meet this requirement then you will be able to resubmit the entry.

See here for the rules, the chnages have been highlighted like this.
http://www.vbforums.com/showthread.p...postid=1782180


BTW work on the next contest is well underway and shouldn't need such chnages like this as we plan to get the rules right before it goes live (expected date should be same day this one ends :)) ;).



Tec-Nico: his current one, the green blob :lol: :D.

---

:p Damn! I knew I shouldn't have sent my code in to the last minute......

But why change it, surely it's just another refinement to the coding. I managed to get my whole routine (including display) down the 0.6ms. I honestly think that was a fair enough way to optomise the code, and spent a fair bit of time working on it. :)

Isn't it a bit unfair to penalise my ingenuity :p

Quote:

---

Posted by Ex-FB
:p Damn! I knew I shouldn't have sent my code in to the last minute......

But why change it, surely it's just another refinement to the coding. I managed to get my whole routine (including display) down the 0.6ms. I honestly think that was a fair enough way to optomise the code, and spent a fair bit of time working on it. :)

Isn't it a bit unfair to penalise my ingenuity :p

---

I do think it was rather clever of you to think of that loop-hole. However it was kinda implied that we wanted one prime per entry and it was a mistake on our behalve to have not made it clear. Mind there is still 10 days left so that should be plenty, if this had only been noticed later on then we'd have let you off with it cos it wouldn't be fair on you to change the rules at the end of the contest :).

:) I don't really see it as a loop-hole as it is a valid way to tune an application. it's not just adding multiple entries per line, it's the way I build up the strings and the implement the scrolling etc.

When 90% of an applications speed is taken up by one bit of code, that's the major bit that should be tuned. That's what I did.

It's going to be difficult to judge the speed of the algo's when they are going to be so heavily outweighed by the listbox and you won't let us tune it.

I don't really mind, it is after all just a competition, but it does seem somewhat unfair and pointless, I'm not sure what your reasoning is (if it's just to make it easier to check the numbers, then it's easy for me to add an attachment that will spool them to a file... besides, I would have thought that rather than check all of them you would just check the winning entry and save yourself some time).

As an outsider joining this competition it just smacks a bit of changing the goal posts to suit the entries you have. I realise this is a test competition, and your still refining things, but isn't that all the more reason to be not to bothered about changing the rules to exclude someone's work.

Am I going to spend time working out another way of doing this only to find that you change the rules again to exclude me from winning?

I understand that really this problem shouldn't have been an issue in the first case if we'd put it in the rules from the start (I actually thought it was in the rules till a couple of days ago when I realised, I then querried others, competition runners, to see if we should add it) but as it was missed out we've had to make the decission.

It wasn't your entry that triggered this, as I stated in the other thread, but another member asked me whether it would be allowed. My first reaction was that the rules already said one per line and I then checked and saw it wasn't in. I then said well if it is allowed then it would loose points for appearance unless it was done very well with proper indenting or something like that. And as you can see after checking with others the change was made.

I suppose seen as though you feel so strongly and give good reasons then I do kind of regret it now, but taking it away now just means everyone will have the idea and then everyone will do it. Which I suppose would have been better just the odd one or two doing it.

Quote:

---

Originally posted by Electroman
I understand that really this problem shouldn't have been an issue in the first case if we'd put it in the rules from the start (I actually thought it was in the rules till a couple of days ago when I realised, I then querried others, competition runners, to see if we should add it) but as it was missed out we've had to make the decission.

It wasn't your entry that triggered this, as I stated in the other thread, but another member asked me whether it would be allowed. My first reaction was that the rules already said one per line and I then checked and saw it wasn't in. I then said well if it is allowed then it would loose points for appearance unless it was done very well with proper indenting or something like that. And as you can see after checking with others the change was made.

I suppose seen as though you feel so strongly and give good reasons then I do kind of regret it now, but taking it away now just means everyone will have the idea and then everyone will do it. Which I suppose would have been better just the odd one or two doing it.

---

:p Don't worry, It doesn't really bother me too much. I'm guessing that maybe the person who asked you was one of the people who tested my executable :) Which is fine. It would make for a closer competition though if everyone did it (i.e. put more emphasis on the actual algorithm itself).

Whatever you decide is fine by me. :thumb:

Thanks for running the comp.

This is also one of the major reasons I'd like atleast adding to listbox to be excluded from the timing. Without such hacks, there is no way making listbox any faster, and I myself preferred the individual selection of each item (for a reason, which you see when you happen to get your hands on my project). Instead of just speeding up I concentrated thoughts on the other areas too, I just didn't want to make a dull list of 1000 prime numbers.


Algorithm only: 0.02 ms
Algorithm and preparing memory: 0.03 ms
Algorithm and all initialization: 0.08 ms

All listbox stuff: 6.45 ms

Quote:

---

Posted by Ex-FB
:p Don't worry, It doesn't really bother me too much. I'm guessing that maybe the person who asked you was one of the people who tested my executable :) Which is fine.

---

It was yes but it wasn't in a spiteful way at all, they were actually checking to make sure you wouldn't get disqualified when it came to judging as that would have been rather bad.

I do agree that the listbox is a really big bottleneck and we should have either made them calculate more primes or just required you to fill in a array of ints/longs. The listbox was added to add an extra edge which would make you think about something that wasn't the algo. Yes I agree you did come up with a good thing but choice has been made on that. Also people shouldn't be just thinking about the speed there is other factors that will play a part in this too and as I see Ex-FB from your screen shot in the other thread you haven't ignored the appearance :).

For the next contest we are planning it in a way that means you will have the full scoring system from the beginning of the contest. Its not clear yet if the next one will be timed but if it is then we are going to avoid such bottlenecks or make the weighting of the time a low priority ;).

Quote:

---

Originally posted by Electroman
It was yes but it wasn't in a spiteful way at all, they were actually checking to make sure you wouldn't get disqualified when it came to judging as that would have been rather bad.

---

Don't worry, I didn't for a second think it would have been spiteful :p This is a game, a fun mental challenge. I'd already mentioned it to a few people when they asked how i managed to get such an over-all fast time.

I also appreciate that it's very hard to know before you set the challenge as to how the solutions are going to work out, which makes your job of creating the marking method all the harder. I don't think there's going to be a measurable difference between all the top runners in this comp (we seem to be getting =<0.02ms calc times), but that would have been difficult for you to know before the comp.

But of course, that what this test comp was probably for, to iron out all these wrinkles, so in many ways you could say it's already been a resounding success :thumb:

Here's a humble suggestion.... instead of using timings as a measure of winning, how about having a competition between different programs in a sort of playoff.

For example. The problem could be to guide a robot around a randomly generated maze shooting at other player robots, or a game of Othello/Reversi, where programs play each other in a leauge.

Of course both of these would require some intermediary code (provided as part of the challenge) that would provide as an interface, and maybe a standard file format (which should be easy to set up) that this intermediary code could read and a standard output file that it would write back, ready for the next turn.

But I'm sure given the resources available here you'd have no end of volunteers for such a project.

Just a thought.. :)

Quote:

---

Posted by Ex-FB
Here's a humble suggestion.... instead of using timings as a measure of winning, how about having a competition between different programs in a sort of playoff.

For example. The problem could be to guide a robot around a randomly generated maze shooting at other player robots, or a game of Othello/Reversi, where programs play each other in a leauge.

Of course both of these would require some intermediary code (provided as part of the challenge) that would provide as an interface, and maybe a standard file format (which should be easy to set up) that this intermediary code could read and a standard output file that it would write back, ready for the next turn.

But I'm sure given the resources available here you'd have no end of volunteers for such a project.

Just a thought.. :)

---

That actually plays on what we have for the next one :eek:. Mind the pitting the entries against each other isn't in it, would be a good idea for a future one though, will make a note :D.

Quote:

---

Originally posted by Ex-FB
:p Don't worry, It doesn't really bother me too much. I'm guessing that maybe the person who asked you was one of the people who tested my executable :)

---

Well I must admitt that I have tested most of the guys code...and I know that you are not the only one that was on this thought....One guy actulay told me that he thought it was my idea beacause one of the first posts I had in this thread....the one with all the primes lined up in 4 coloums......;)

Quote:

---

Originally posted by NoteMe
Well I must admitt that I have tested most of the guys code...and I know that you are not the only one that was on this thought....One guy actulay told me that he thought it was my idea beacause one of the first posts I had in this thread....the one with all the primes lined up in 4 coloums......;)

---

:p Not to worry. I still have plenty more ace's up my sleeve ;)

Quote:

---

Originally posted by NoteMe
Well I must admitt that I have tested most of the guys code...and I know that you are not the only one that was on this thought....One guy actulay told me that he thought it was my idea beacause one of the first posts I had in this thread....the one with all the primes lined up in 4 coloums......;)

---

P.S. Did anyone get faster than mine?

Quote:

---

Originally posted by Ex-FB
P.S. Did anyone get faster than mine?

---

Well I didn't look at all of them...but you where the fastest of the ones I tested...but you where also the one that I tested that added most primes to one line...;)...

I feel like the timeline for this contest was a little too long...

Yeah, a week less would have been fine for this simple project. Of course there is a chance for people who haven't yet noticed this thing and would like to join, but still... quite long. If one wanted to, he could really perfect the prime number finding. And of course, even the most brain injured is able to find out a fast algorithm :D

Quote:

---

Originally posted by Merri
And of course, even the most brain injured is able to find out a fast algorithm :D

---

Yeah, Bodwad has already submitted his entry. :lol:




j/k pal.

Quote:

---

Originally posted by Merri
And of course, even the most brain injured is able to find out a fast algorithm :D

---

Stop it :cry: :cry: :cry:

I need to resubmit :D

Aw, darn, still five days to go :(

Have everyone submitted their work so the contest could be ended earlier? :D

yep!
Darn Listbox has sapped all my motivation.
:(

not yet:)

Not submitted mine yet - still tweaking

Gotta find time to go back and change it in line with the new rules.

Don't know if I will have the time to do it though.... busy busy busy.

Well, had a quick hack at it, don't have time to waste on it making it any faster, wasted to much time already. It's done and dusted.

Cheers.

the time is now 24 September, GMT - 5, 7:30 PM.
tickticktick..

:wave:

The time is now 25th September, GMT +2, 2:35 AM

Tick to you too :D

Two hours to go until the deadline! :D


And no, I haven't been staying awake all night because of waiting the results, I've had a week off and I messed my sleep rhythm so bad I have to stay awake long enough to fix it enough!

I'm going to wait until 0.2 milliseconds before the deadline to submit my final prog :D :lol:

Ok everybody the contest is now over :)

Hope you all got your entries in on time!

We will get the marks to you as soon as possible

Will the "time" for the app be published all at the same time, or will you add one and one and one...after you have tested it...that would be soooo much more fun...then maybe add the score for sexy ness afterwards...;)

Ohhh and DON'T let HeMan test my code....his computer ****s ups my algo way beyound the infinity...

When can we expect the result(s) to be posted?

Will you be posting the config of the "Test" machine used too?

Final revised project, VB.Net 2002. :wave:

Cheers,
NTG

Quote:

---

Originally posted by ntg
Final revised project, VB.Net 2002. :wave:

Cheers,
NTG

---

read the other threads on how to post the code

Quote:

---

Originally posted by Acidic
read the other threads on how to post the code

---

I've posted for the contest already, just sharing some refinements.

Cheers,
NTG

Quote:

---

Originally posted by ntg
I've posted for the contest already, just sharing some refinements.

Cheers,
NTG

---

Did I do something wrong?

Quote:

---

according to you
' From Nth prime formula we know that the 1000 primes are below 7920.

---

Assuming you are reffering to "The Prime Number Theorem: approximating pi(x)" as seen on this page:

http://www.utm.edu/research/primes/howmany.shtml

The formula:

Consequence One: You can Approximate pi(x) with x/(log x - 1)

{Where log is actually the natural log, or ln}


seems to return 992.836 approximate primes below 7920.

So It would seem you can't know for a fact that there are 1000 primes below 7920, if you are using this formula.


:wave:
-Lou

Quote:

---

Originally posted by Electroman
That actually plays on what we have for the next one :eek:. Mind the pitting the entries against each other isn't in it, would be a good idea for a future one though, will make a note :D.

---

How would they interact?? :confused:

Oh, I know, you could have us all design a module which has a "main" function that you (with a small bit of coding) call with all the input data form the main loop. The output would be "move, shoot" etc.

I WANT IN ON THAT ONE!

In fact you should make movies of them fighting... muahaha

Quote:

---

Originally posted by alkatran
How would they interact?? :confused:

Oh, I know, you could have us all design a module which has a "main" function that you (with a small bit of coding) call with all the input data form the main loop. The output would be "move, shoot" etc.

I WANT IN ON THAT ONE!

In fact you should make movies of them fighting... muahaha

---

Exactly.... That would be a blast!

Even if we were to start in on that idea for the next contest, it
would take too long for it to be ready. I dont want to wait
months for the next contest. Maybe somewhere down the road.

Any other ideas for the next contest?

Maybe the fastest method of showing numbers 1 to 1000 in the listbox so that each item would be individual.

Nah, that'd be completely uncool. Anything related to timing controls you can't really change is uncool :D

Anyways, I'd recommend something simple, yet challenging. Pretty much like this prime numbers contest (a little harder though). Too bad these are rather hard to come up with... I didn't even have knowledge on prime numbers before this contest, I had heard something about them sometime before, but completely forgotten about them. And I can't come up with anything similar. The only related thing would be "count the first 1000 decimals of PI". Actually, that might be even hard to come up with... I have no idea at the moment how I'd start coding that!

There are something that could be done, more useful (as said by someone else already). Like, make a control that does something. Then this control would be rated by its useability, the quality of the code and overall feeling.

Atleast came up with some suggestions :)

Quote:

---

Originally posted by RobDog888
Even if we were to start in on that idea for the next contest, it
would take too long for it to be ready. I dont want to wait
months for the next contest. Maybe somewhere down the road.

Any other ideas for the next contest?

---

its almost ready for you all to start ;)

Think you can give us an ETA on the results?

:wave:

I think that the next contest should be something that is
not too involved and something that requires the contestant to
think and be unique. Something that is not a search google and
copy and paste some code and then modify it a bit and call it
done. Also it should be a shorter deadline.

Yeah, the whole timing thing is very difficult when you are measuring in milliseconds. if it was another timed one, I think it would have to be in the magnitude of minutes to calculate to make it measurable.

After all, if you time the prime generator individually you get wildly random results, if you stick the calc in a loop 1,000 times and average it out, you always get a much smaller number than all of the other individual timings because of the on-board processor cache which stores it after the first loop, so you can't even use that method....

As for the previous idea about programs interacting, you could do it in stages. Have a maze solving algorithm as the first comp, which avoids other players, then release the best ones for the next comp so people all start from the same point, then add guns, or multiple bots/tanks/whatever you want in the maze etc.
You could keep progressing it until you almost had an AI game engine.

As for othewr challenges, how about the best remake of an 80's arcade game (people can choose what they want to remake (Pacman, Space invaders, Asteroids etc.)) This would involve a number of different programming skills and would be quite fun for everyone after the comp to play around with.

Or how about a competition to make the most fun card game or one armed bandit.

Just some more random thoughts.

Hey, this contest did take time to think. Atleast for me. I didn't take a look to any previous algorithms until I already had about done my contest entry. I now know I ended up in using a method invented by someone else a looong time ago - in a slightly optimized manner.

Quote:

---

Originally posted by Merri
Maybe the fastest method of showing numbers 1 to 1000 in the listbox so that each item would be individual.

Nah, that'd be completely uncool. Anything related to timing controls you can't really change is uncool :D

Anyways, I'd recommend something simple, yet challenging. Pretty much like this prime numbers contest (a little harder though). Too bad these are rather hard to come up with... I didn't even have knowledge on prime numbers before this contest, I had heard something about them sometime before, but completely forgotten about them. And I can't come up with anything similar. The only related thing would be "count the first 1000 decimals of PI". Actually, that might be even hard to come up with... I have no idea at the moment how I'd start coding that!

There are something that could be done, more useful (as said by someone else already). Like, make a control that does something. Then this control would be rated by its useability, the quality of the code and overall feeling.

Atleast came up with some suggestions :)

---

What about the "fibernatch numbers" (SP?) or the 100000 first desimales of the natural number e...;)

Those say completely nothing to me at the moment, so they sound real good :D

I'm not saying that. I'm trying to say that if we didn't have the
alogorithums out there what would people do. Yes, it still takes
talent to do whatever you did to make it so fast, but I was
focusing on the general consensous that everyone searched for
the best alogorithum and used that. What if everyone had to
come up with their own? Then the competition would not have
been to the millisecond and would be a little more spread out
based on creative concept and code.

Well there is an algorithm for everything...

Actually, I did include some other creative parts to my prime numbers project too. I made it show some "nice" animation generated by the prime numbers, added some minor extra useability stuff for using pleasure :)

Did you already forget it wasn't only about the speed of the code? ;)

No I didn't forget. I didn't even use a published concept on generating
primes. I used common sense approach even though it was a
little slower than yours. So I am counting on originality and
simplicity. I only used a few lines of code. I only used the number
2 and that was all. At first I had a progressbar and some stuff in
it but that was WAY to slow.

I had too much work to do so I couldnt spend allot of time on it
and thats why I almost didnt make the deadline. Now that I have
the time to think about it, I could have added code to add to the
listbox in the main loop instead of looping twice. :(

Quote:

---

Originally posted by RobDog888
I'm not saying that. I'm trying to say that if we didn't have the
alogorithums out there what would people do. Yes, it still takes
talent to do whatever you did to make it so fast, but I was
focusing on the general consensous that everyone searched for
the best alogorithum and used that. What if everyone had to
come up with their own? Then the competition would not have
been to the millisecond and would be a little more spread out
based on creative concept and code.

---

Er, actually the only bit of knowledge I used was that if a number has a divisor, then one of them must be prime. I didn't find an algorithm that matched mine until after I had worked it out myself.... :p

Quote:

---

I didn't even use a published concept on generating
primes.

---

Neither did I. There is no exact concept of what I did. The way I did it was that I used the extended common sense I have ;)

Btw. I just got one more idea for optimization! Aw darn! Why did I get it NOW?

Quote:

---

Originally posted by NotLKH
Did I do something wrong?



Assuming you are reffering to "The Prime Number Theorem: approximating pi(x)" as seen on this page:

http://www.utm.edu/research/primes/howmany.shtml

The formula:

Consequence One: You can Approximate pi(x) with x/(log x - 1)

{Where log is actually the natural log, or ln}


seems to return 992.836 approximate primes below 7920.

So It would seem you can't know for a fact that there are 1000 primes below 7920, if you are using this formula.


:wave:
-Lou

---

Right. I found something here initially and clicking away I came up to a web that provided another formula, can't seem to find it now though :rolleyes: so I stand corrected. My initial source could be ammended to use a different number, but that may not be necessary considering that the values of pi(x) are pre-calculated and widely published on the net.

Cheers,
NTG

Yes, extended common sense. I have it to. This is what the contest
should be based on. Not published mathematical alogorithums. :)

Guess we should see what alogs everyone used before we (I)
make any more judgements. :D

I used no division/multiplication or anything of a higher
mathematical operator.

Can't wait to see the results.

Well, it makes sence to not just have contests on just mathematical algos that everyone can already find cause they are all over the net... Just keep in mind that it's the begining and this was kind of a "test" contest to see how everything goes :)

I have that in mind. I hope for the next contest we keep on the
same lines of something that will appeal to a wide range of
members.

Any word on the progress of evaluating the entries is going?

:)

One of the things that needs working out. I hope/believe it will be sorted tomorrow (or later, however it is for everyone) :)