方法一：通过命令denseToSp可以产生一个sparse matrix. However, gauss is unable to store this sparse matrix on hard disk. The only way is writing your own program and produce a n*3 matrix P with P[i,1] = value, P[i,2] = row_index and P[i,3] = column_index. The problem with this solution is it often takes a long time to run the looping.

方法二：build up a structure with sparse matrix as its element and save an instance of this structure to your hard disk (like "class" concept in C++).

不晓得有没有人用gauss 12 新版本有直接存储稀疏矩阵到硬盘上的功能么？

求解答~