Skip to content
This repository has been archived by the owner on Jul 27, 2022. It is now read-only.

google/fpc

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

62 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Fixed-point Calculator

This program takes a range [min, max] and precision and calculates constants and information to be used in code performing calculations in this range using fixed-point integer math.

This is not an official Google product.

Try it!

$ git clone ssh://git@stash.nestlabs.com:7999/user/fpc.git
$ make
$ ./fpc -g 30 1800 0.1
[PARAMETERS]
  min: 30 (30 requested)
  max: 1800 (1800 requested)
  precision: 0.0625 (0.1 requested)

[CODE]
  code density: 62.5%
  offset: 0
  code range: [480, 28800]

[ENCODING]
  machine bit width: 16 (15 used)
    fractional bits: 4
    integer bits: 11
  use signed: no
  machine integer type: uint16_t
  Q notation: Qu12.4

[CONVERSION]
#include <math.h>
#include <stdint.h>
// can lose precision, for display only
double convert_to_double(uint16_t x) {
  if(x < UINT16_C(480) &&
     x > UINT16_C(28800)) {
    return NAN;
  }
  return round(ldexp(x, -4) * 10) * 0.1;
}
$ make convert
$ ./convert 480 600 10000 28800
480 -> 30
600 -> 37.5
10000 -> 625
28800 -> 1800

Also, check this out:

$ ./fpc -2^7 -l-p 2^-8
[PARAMETERS]
  min: -128 (-128 requested)
  max: 127.99609375 (127.99609375 requested)
  precision: 0.00390625 (0.00390625 requested)
...

fpc has an expression evaluator built in to allow simple expressions!

  • supports (in order of precedence)
    • parenthesis: (x)
    • exponentiation: x ^ y
    • multiply: x * y, divide: x / y
    • addition: x + y, subtraction: x - y
  • some variables
    • min: l
    • max: h
    • precision: p

About

FPC calculates information to be used in fixed-point integer calculations

Resources

License

Code of conduct

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published