二次剩余伪随机数发生器

hbghlyj

readme中说

The math behind this is pure magic. Whoever figured this out, please explain lol.

我也想弄明白这个代码是什么原理
E、A、O必须是小于U的素数，其中E最大，A次之，O相对较小。
E 应该是小于 U 的最大素数。
U 可以是任何 bigint。
所有数字都必须是 bigint。
I 是一个递增的 bigint

1. const permute = require('@lancejpollard/quadratic-residue-prng.js')
3. const E = 3132343537383103113163n
4. const A = 975319753197531975319n
5. const O = 541613713n
6. const U = 32 ** 15
8. let i = 0
9. while (i < U) {
10.   const x = permute(i, E, A, O, U)
11.   console.log(x)
12.   i++
13. }

复制代码

hbghlyj

So for example, you want to generate 3 sets of pseudo random IDs, you're going to need 3 different sequences (output), but can use the same range of input sequences (0 to max U). Pick a unique A for each sequence, but fix E, O, and U. Then when you input i, you will get different values for each sequence at index i. Sometimes you might want to change U to be a different max, to limit the range (and size) of possible values, as well. But if U is constant, to generate different sequences, vary A on a per-sequence basis.