exp()
Posted: Sat Nov 10, 2018 12:26 amBy documentation there are quite few functions offered (excl. what Ruby offers).
Would there be any point to write some approximations using DSP Code component or Assembler? Would it be much slower compared to build-in function(s)?
I was thinking some of these I've used in C++ coding (as SSE versions) :
http://www.chokkan.org/software/dist/fastexp.c.html
https://root.cern.ch/doc/v606/exp_8h_source.html
or just some simple exp approx like:
Would there be any point to write some approximations using DSP Code component or Assembler? Would it be much slower compared to build-in function(s)?
I was thinking some of these I've used in C++ coding (as SSE versions) :
http://www.chokkan.org/software/dist/fastexp.c.html
https://root.cern.ch/doc/v606/exp_8h_source.html
or just some simple exp approx like:
- Code: Select all
return (40 * x * (x * x + 42))/(x * (x * ((x - 20) * x + 180) - 840) + 1680) + 1;