I was so happy when I tested this and it worked. First 2, then 3, skipped 4 and added 5…

I just need someone to confirm the code and conclude that it is a working Prime Number Generator.
https://scratch.mit.edu/projects/99036385/
P.S.
Ok, I know people have made things like this before. But I have yet to have seen one on Scratch. So I decided to make one, and after hours of debugging, I'm done.

There certainly are plenty of prime number generators on Scratch - a quick search shows that. Here's mine for example https://scratch.mit.edu/projects/36613800/

Yours does work but can be improved.

Since 2 is the only even prime it makes sense to treat it as a special case and only test odd numbers after that.

Unless the number is a square any factor greater than the square root must have a partner less than the square root so you only need to check up to the square root - this makes a big difference as the numbers get bigger.

If a number isn't divisible by any of the primes you've found so far it is a prime - skip the composite numbers when checking for factors.

Combine these and for a number like 401 you'd only check these numbers as factors - 3, 5, 7, 11, 13, 17, 19 whereas you check all the numbers from 2 to 400.

deck26 wrote:

There certainly are plenty of prime number generators on Scratch - a quick search shows that. Here's mine for example https://scratch.mit.edu/projects/36613800/

Yours does work but can be improved.

Since 2 is the only even prime it makes sense to treat it as a special case and only test odd numbers after that.

Unless the number is a square any factor greater than the square root must have a partner less than the square root so you only need to check up to the square root - this makes a big difference as the numbers get bigger.

If a number isn't divisible by any of the primes you've found so far it is a prime - skip the composite numbers when checking for factors.

Combine these and for a number like 401 you'd only check these numbers as factors - 3, 5, 7, 11, 13, 17, 19 whereas you check all the numbers from 2 to 400.

Thank you for the tips, though in the case of 401, I only check up to 200.

LastAttack1257 wrote:

Thank you for the tips, though in the case of 401, I only check up to 200.

OK, I didn't look that closely. You're still checking 199 potential factors instead of my 7 though.

deck26 wrote:

LastAttack1257 wrote:

Thank you for the tips, though in the case of 401, I only check up to 200.

OK, I didn't look that closely. You're still checking 199 potential factors instead of my 7 though.

Yes, you are quite right. I will get to work on that. If it took 5 hours to get my slow generator to work, it's going to take a couple days to refine it to your calibre.

Feel free to copy any of my coding.

deck26 wrote:

Feel free to copy any of my coding.

Why, thank you. I won't directly copy, but I will try to use the concept of square root factors. Also, could you go into depth about what I should do when the number DOES happen to be square?

LastAttack1257 wrote:

deck26 wrote:

Feel free to copy any of my coding.

Why, thank you. I won't directly copy, but I will try to use the concept of square root factors. Also, could you go into depth about what I should do when the number DOES happen to be square?

As long as you check up to and including the square root there's no problem.

deck26 wrote:

LastAttack1257 wrote:

deck26 wrote:

Feel free to copy any of my coding.

Why, thank you. I won't directly copy, but I will try to use the concept of square root factors. Also, could you go into depth about what I should do when the number DOES happen to be square?

As long as you check up to and including the square root there's no problem.

So, just to sum it up, I round the square root of the number I'm checking, then check all the prime numbers up to, and including, that number to see if it is a composite. If not, I add it to the list, and change the number by 1, and restart the process.

LastAttack1257 wrote:

deck26 wrote:

LastAttack1257 wrote:

deck26 wrote:

Feel free to copy any of my coding.

Why, thank you. I won't directly copy, but I will try to use the concept of square root factors. Also, could you go into depth about what I should do when the number DOES happen to be square?

As long as you check up to and including the square root there's no problem.

So, just to sum it up, I round the square root of the number I'm checking, then check all the prime numbers up to, and including, that number to see if it is a composite. If not, I add it to the list, and change the number by 1, and restart the process.

Pretty much it but no need to round the square number. If it's not a whole number it doesn't affect anything. I'd just put even numbers above 2 straight into the composites though and change the number by 2.

So store 2 in primes, then repeat the following
1. increase number by 1 and check for prime - store in appropriate list
2. increase number by 1 and store as composite

deck26 wrote:

LastAttack1257 wrote:

deck26 wrote:

LastAttack1257 wrote:

deck26 wrote:

Feel free to copy any of my coding.

Why, thank you. I won't directly copy, but I will try to use the concept of square root factors. Also, could you go into depth about what I should do when the number DOES happen to be square?

As long as you check up to and including the square root there's no problem.

So, just to sum it up, I round the square root of the number I'm checking, then check all the prime numbers up to, and including, that number to see if it is a composite. If not, I add it to the list, and change the number by 1, and restart the process.

Pretty much it but no need to round the square number. If it's not a whole number it doesn't affect anything. I'd just put even numbers above 2 straight into the composites though and change the number by 2.

So store 2 in primes, then repeat the following
1. increase number by 1 and check for prime - store in appropriate list
2. increase number by 1 and store as composite

So, I have taken what you have told me, and some snippets of your code, and now I have created this:
https://scratch.mit.edu/projects/99195997/#player

I'm not quite sure if this is the fastest possible, so I was hoping you could check.

Here is a faster remix > https://scratch.mit.edu/projects/99197535/

P444 wrote:

Here is a faster remix > https://scratch.mit.edu/projects/99197535/

I literally added a hide list function in mine the moment I got a notification about your faster method

Mine is still faster cause I am using the custom block without screen refresh xd

P444 wrote:

Mine is still faster cause I am using the custom block without screen refresh xd

Also because you're not showing a timer. (edit - sorry I thought you were comparing yours with mine)

Stripping out the variable updates etc can make it at least twice as fast as yours. So it's a case of priorities really - maximising the prime calculation or showing what's happening.

deck26 wrote:

P444 wrote:

Mine is still faster cause I am using the custom block without screen refresh xd

Also because you're not showing a timer.

Stripping out the variable updates etc can make it at least twice as fast as yours. So it's a case of priorities really - maximising the prime calculation or showing what's happening.

So, I could make the generator faster by completely removing the Composite List and all associated blocks?
Also, hiding the lists make it faster, so maybe hiding the variables might make it slightly faster?

EDIT: After two hours, my new generator has calculated over 200,000 prime numbers, which is much faster than my previous, which would only calculate 2 or 3 thousand in that time.

If you want to see who can produce the quickest version I suggest setting a time limit - eg 60 seconds. The green flag script should hide any variables that need hiding following a previous run and then reset the timer and you should use a

when [timer v] > (60)
stop [other scripts in sprite v]
then show any variables you want to show

I see no need to store composite numbers and displaying lists that are changing is always slow.

Obviously it depends on the speed of the computer etc but I've just generated around 30,400 primes in 60 seconds by not showing anything changing on the screen..

deck26 wrote:

Obviously it depends on the speed of the computer etc but I've just generated around 30,400 primes in 60 seconds by not showing anything changing on the screen..

I am keeping my generator on for as long as possible to see the highest number I can get to,
but I'll test in another tab later.

LastAttack1257 wrote:

deck26 wrote:

Obviously it depends on the speed of the computer etc but I've just generated around 30,400 primes in 60 seconds by not showing anything changing on the screen..

I am keeping my generator on for as long as possible to see the highest number I can get to,
but I'll test in another tab later.

You may hit problems with large numbers after a while though. Off the top of my head I think Scratch will only cope with around 14 digits before it can't store integer values accurately.