%o=extended(s)
%-------------
%
%Convert to extended precision.
%
% Marcel Leutenegger 1.1.2005
%
% This class provides floating point computation in
% the native floating point format of the processor.
% The extended data type has a 64bit mantissa instead
% of the 53bit mantissa of double data.
%
% Due to the native execution by the processor, the
% performance is affected only by memory throughput.
% The current implementation minimizes memory con-
% sumption but does not align consecutive values at
% 8 byte boundaries.
%
% Constants and random numbers are provided by
%
% eps(extended) == pow2(-63)
% pi(extended) ~= extended(pi)
% ones(extended)
% rand(extended)
% randn(extended)
% realmax(extended)
% realmin(extended)
% zeros(extended)
%
%Input:
% s extended object, numeric or character array
%
%Output:
% o extended object
%
