用QR分解来求矩阵的全部特征值 包括：QR分解，矩阵转置，；矩阵求逆，矩阵相乘，最后迭代得出特征值-With QR decomposition to the full matrix eigenvalue include: QR decomposition, matrix transpose, matrix inversion, matrix multiplication, and finally reached eigenvalue iteration

cuda平台开发的矩阵变换并行操作（Matrix transpose），属于cuda sdk中的内容-Small Matrix transpose with Cuda (Example for a 16x16 matrix)

用三元组表示稀疏矩阵,可以转置 加法，减法，乘法，求逆,Triples representing sparse matrix transpose addition, subtraction, multiplication, and inversion