Crates.io | q-num |
lib.rs | q-num |
version | 0.1.2 |
source | src |
created_at | 2023-09-28 16:57:24.20387 |
updated_at | 2024-08-13 01:30:28.111663 |
description | Q notation for fixed-point numbers via a proc_macro. |
homepage | |
repository | https://github.com/xpe/q-num |
max_upload_size | |
id | 986107 |
size | 35,516 |
This library provides the define_q_num!
procedural macro (evaluated at compile
time) to define a signed/unsigned binary fixed-point number type. It uses
ARM-style Q notation: Qm.n
(signed) or UQm.n
(unsigned), where:
Internally, the macro chooses the narrowest primitive integer type that can hold
m + n bits, up to u64
(unsigned) and i64
(signed). More internal details
are discussed below.
A Q number's value is the ratio of the stored number (having n + m bits) and a fixed denominator (equal to 2 ^ n).
For example, using the UQ3.2 specification, the bit pattern 0b10111 represents the value 5.75. Keeping in mind the denominator is 2 ^ 2 = 4, there are two ways to see this:
Here is one example:
define_q_num!(MyQ, Q6.2);
let a = MyQ::from(13.75);
let b = MyQ::from(-2.25);
let c = a + b; // 11.5