Complexity of spconvert in matlab

Which is the complexity of spconvert function in matlab? I am trying to create the sparse representation of a sparse matrix. My matrix is the adjacency matrix of a graph. I ve got two adj_matrices. Ones with 7.000 nodes and ones with 400.000 nodes. I am using the following code:

f_adj = textread('mat.txt'); % a nx3 matrix represent the adjacency matrix nxn
adj = spconvert(f_adj);
[a b] = eigs(adj);
spy(adj);

The text file mat.txt is a nX3 files which contain the edges of the graph. What I want to ask is the complexity of the sponvert. In my 4gb ram pc I am able to create the first graph but I couldnt create the second graph.