Does MATLAB support quadruple precision - 128-bit floating point arithmetics?

Question

MathWorks Support Team 2012-11-15

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/98215-does-matlab-support-quadruple-precision-128-bit-floating-point-arithmetics

评论： Walter Roberson 2024-8-16，22:53

I need more precision than offered by the standard double data type in MATLAB. I am looking for a quadruple precision in MATLAB.

请先登录，再回答此问题。

Answer 1

MathWorks Support Team 2024-7-25

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/98215-does-matlab-support-quadruple-precision-128-bit-floating-point-arithmetics#answer_107565

编辑：MathWorks Support Team 2024-8-16，21:42

The floating point arithmetic format that occupies 128 bits of storage is known as binary128 or quadruple precision.

The following blog post by MathWorks' Chief Mathematician Cleve Moler describes an implementation of quadruple precision programmed entirely in the MATLAB language:

https://blogs.mathworks.com/cleve/2017/05/22/quadruple-precision-128-bit-floating-point-arithmetic/

Please note that quadruple precision is not available as of MATLAB 2024b. As a workaround, it is possible to obtain more precision using Variable Precision Arithmetic from the Symbolic Math Toolbox. Another possibility is to use the Fixed-Point Toolbox to obtain more precision. Note that in either case, the numbers will not be of the floating point format.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Walter Roberson 2024-8-16，22:53

Note: the internal representation of variable precision numbers is not documented. It is believed that it has something to do with 30 bit chunks held within 32 bit numbers.

请先登录，再进行评论。