high dimensional matrix takes too long to generate

I have two matrices. A(3x226981), B(3x40800501).

I want to do the following process. but it takes too long.

I used parfor but it's still too long. is there a simple way?

for x=1:length(B)

for y=1:length(A)

P= [B(:,x);A(:,y)]

TOP=[TOP P]

end

end

[Bruno Luong]

it takes more than 404 Tb to store your result, let alone your code is badly written.

[savas yilmaz]

How would you suggest I write?

[Steve Eddins]

Before trying to make specific suggestions, I'd like to clarify what you're trying to do. If I understand that code and the matrix sizes correctly, you are trying to construct a matrix that has 6 rows and approximately 9 trillion columns. Such a matrix would need approximately 400 terabytes of memory.

Is that your intent?

[savas yilmaz]

Yes

[Bruno Luong]

@savas yilmaz "How would you suggest I write?"

I suggest you NOT to wrte it at all, since you likely have no place to store it.

For toy example

A=randi(10,3,4)

A = 3×4

9 2 8 8 6 9 4 9 9 4 8 5

B=randi(10,3,5)

B = 3×5

6 4 5 9 9 9 8 8 8 2 9 1 4 7 1

TOP = [repelem(B,1,size(A,2));repmat(A,1,size(B,2))]

TOP = 6×20

6 6 6 6 4 4 4 4 5 5 5 5 9 9 9 9 9 9 9 9 9 9 9 9 8 8 8 8 8 8 8 8 8 8 8 8 2 2 2 2 9 9 9 9 1 1 1 1 4 4 4 4 7 7 7 7 1 1 1 1 9 2 8 8 9 2 8 8 9 2 8 8 9 2 8 8 9 2 8 8 6 9 4 9 6 9 4 9 6 9 4 9 6 9 4 9 6 9 4 9 9 4 8 5 9 4 8 5 9 4 8 5 9 4 8 5 9 4 8 5

[Bruno Luong]

@savas yilmaz "Yes"

Do you know what is a Tb?

[savas yilmaz]

of course. I realized that creating this matrix is a bit difficult. however, although it is small in size, I learned to do this with a different command. thanks

[Steven Lord]

Your machine almost certainly doesn't have enough memory to store that data, and even if you could processing the combinations brute-force would take a very, very long time.

numColumns = 226981*40800501

numColumns = 9.2609e+12

Let's say you could process a thousand columns per second. How long would it take to process that data?

theTime = years(seconds(numColumns/1000))

theTime = 293.4675

I don't think you want to wait for almost 300 years for your answer.

If you explain what problem you're hoping to use this enormous matrix to solve, we may be able to offer a solution that won't make you wait until sometime in 2316 for the answer.

datetime('now') + seconds(numColumns/1000)

ans = datetime
   10-Feb-2316 09:07:03