Begin New Subject | Flat View | Search | List | List Messageboards | Help |

square_sub_x |
---|

Author: | Date: 2015-05-15 07:29 |

I have implemented a vectorized double-double (https://en.wikipedia.org/wiki/Quadruple-precision_floating-point_format#Double-double_arithmetic) type using your Vector Class Library. Your mul_sub_x function is very useful. For example to multiply two doubles to get a double-double the function is simply template<typename TYPEV> where TYVEV is e.g. Vec2d in which case DOUBLEFLOATN is something like struct {Vec2d hi; Vec2d lo;} However, in my application I need x*x more often than I need x*y. You already provide a square() function but in order to implement a square() function for double-double I need (for efficiency) a square_sub_x function so I though you might want to consider this. Below you can find an example of a square_sub_x function I created for Vec2d based on your mul_sub_x function. I understand of course that you cannot implement everything for every special case but I just thought it was something you might want to consider. Thank you for your very useful Vector Class Library. static inline Vec2d square_sub_x(Vec2d const & a, Vec2d const & c) { |

Reply To This Message |

square_sub_x - zboson - 2015-05-15 |

Begin New Subject | Flat View | Search | List | List Messageboards | Help |