1.请问像这样的比较工整的矩阵的相乘怎么样改写为汇编?
2.再问改写为汇编后代码的大小会减少吗?
t32BITS term1, term2, term3, termSum;
term1=srcM1[0]*srcM2[0];
term2=srcM1[1]*srcM2[3];
term3=srcM1[2]*srcM2[6];
termSum = term1 + term2 + term3;
term1=srcM1[0]*srcM2[1];
term2=srcM1[1]*srcM2[4];
term3=srcM1[2]*srcM2[7];
termSum = term1 + term2 + term3;
term1=srcM1[0]*srcM2[2];
term2=srcM1[1]*srcM2[5];
term3=srcM1[2]*srcM2[8];
termSum = term1 + term2 + term3;
term1=srcM1[3]*srcM2[0];
term2=srcM1[4]*srcM2[3];
term3=srcM1[5]*srcM2[6];
termSum = term1 + term2 + term3;
term1=srcM1[3]*srcM2[1];
term2=srcM1[4]*srcM2[4];
term3=srcM1[5]*srcM2[7];
termSum = term1 + term2 + term3;
term1=srcM1[3]*srcM2[2];
term2=srcM1[4]*srcM2[5];
term3=srcM1[5]*srcM2[8];
termSum = term1 + term2 + term3;
term1=srcM1[6]*srcM2[0];
term2=srcM1[7]*srcM2[3];
term3=srcM1[8]*srcM2[6];
termSum = term1 + term2 + term3;
term1=srcM1[6]*srcM2[1];
term2=srcM1[7]*srcM2[4];
term3=srcM1[8]*srcM2[7];
termSum = term1 + term2 + term3;
term1=srcM1[6]*srcM2[2];
term2=srcM1[7]*srcM2[5];
term3=srcM1[8]*srcM2[8];
termSum = term1 + term2 + term3; |