Dear All, I have the following matrix implementation using Ublas. One is plain old matrix ('m') and the other is sparse matrix ('sm') My question is how can I check the byte size of 'sm' and 'm'? If my implementation is correct 'sm' byte size should give much smaller byte size than 'm'. Since both 'm' and 'sm' print the same output here. //____ BEGIN__ matrix<double> m(matDim[0], matDim[1], 0); coordinate_matrix<double> sm(matDim[0],matDim[1], 0); for (size_t i = 0; i < realValue.size(); ++i){ m(rowIndex[i]-1,colIndex[i]-1) = realValue[i]; sm(rowIndex[i]-1,colIndex[i]-1) = realValue[i]; } // End The full and compilable code can be found here: http://cpp.codepad.org/MbPm3OuK - Gundala Viswanath Jakarta - Indonesia