Afficient way to create "sum matrix"

Question

Orr Streicher 2021-2-1

编辑： Orr Streicher 2021-2-1

Hi,

So i would like to implement a function that get a vector as input lets :

g=[g1, g2, g3]

and return a matrix that contain a sum of pairs: so for g , the output matrix G contain 9 elements the G(i,j)=g(i)+g(j)

The problem is that g will be very large vector.

Does anyone know an efficient way to do that with sparse matrices

Thanks

Answer 1

Stephen23 2021-2-1

g = 1:3;
m = g+g(:)
m = 3×3
     2     3     4
     3     4     5
     4     5     6

Orr Streicher 2021-2-1

Awesome!

thanks

Answer 2

Orr Streicher 2021-2-1

Hi,

So its looks like i was happy too soon =P

Since my vector g is very large ( vector in size of 65536)

I got memrory error ("Error using + Requested 65536x65536 (64.0GB)...)

Maybe there is other solution?

Stephen23 2021-2-1

编辑：Stephen23 2021-2-1

You could create a function, e.g.:

g = 1:3
g = 1×3
     1     2     3
f = @(r,c)sum(g([r,c]));
f(1,2)
ans = 3

Your question mentions sparse matrices: are most of the values in g zero?

Orr Streicher 2021-2-1

编辑：Orr Streicher 2021-2-1

Hi,

In my case g doent have to be sparse (most of the values are not zero)

In addition i need the matrix form of G (for follwing calculations)