running/cumulative sum along all columns of a matrix

[Frank Erickson]

I know this can be done in a loop, but am curious if (i) I missed some built-in way of doing this or (ii) am doing it inefficiently somehow. Here are two loop approaches that give the desired output:

x = (150, 20, 10 \ 10, 20, 30 \ 30, 10, 10 )

nr = rows(x)
nc = cols(x)

xcol_rsum = J(nr,nc,0)

rsumi = J(1,nc,0)
for (i=1;i<=nr;i++){
    rsumi = rsumi + x[i,]
    xcol_rsum[i,] = rsumi
}

// alternately

xcol_rsum2 = J(nr,nc,0)
for (j=1;j<=nc;j++) xcol_rsum2[,j] = runningsum(x[,j])