// SPDX-License-Identifier: AGPL-3.0-or-later
// Normal distribution CDF Φ(x) = P(X ≤ x) where X ~ N(μ, σ²) (f64 canonical)
// Standard normal: Φ(x) = 0.5 * (1 + erf(x / √2))
// General normal: Φ(x; μ, σ) = 0.5 * (1 + erf((x - μ) / (σ√2)))
//
// Also provides PDF φ(x) = exp(-x²/2) / √(2π)
//
// Applications: Black-Scholes, probit models, hypothesis testing, Bayesian inference
// Reference: Abramowitz & Stegun §26.2
@group(0) @binding(0) var<storage, read> input: array<f64>;
@group(0) @binding(1) var<storage, read_write> output: array<f64>;
@group(0) @binding(2) var<uniform> params: Params;
struct Params {
size: u32,
mu: f64, // Mean (0.0 for standard normal)
sigma: f64, // Standard deviation (1.0 for standard normal)
mode: u32, // 0 = CDF, 1 = PDF
}
const PI: f64 = 3.14159265358979323846;
const SQRT_2: f64 = 1.41421356237309504880;
const INV_SQRT_2PI: f64 = 0.3989422804014327; // 1/√(2π)
// Abramowitz and Stegun approximation for erf (embedded for module independence)
fn erf_f64(x: f64) -> f64 {
let a1 = 0.254829592;
let a2 = -0.284496736;
let a3 = 1.421413741;
let a4 = -1.453152027;
let a5 = 1.061405429;
let p = 0.3275911;
let sign = select(-1.0, 1.0, x >= 0.0);
let abs_x = abs(x);
let t = 1.0 / (1.0 + p * abs_x);
let y = 1.0 - (((((a5 * t + a4) * t) + a3) * t + a2) * t + a1) * t * exp_f64(-abs_x * abs_x);
return sign * y;
}
// Standard normal CDF: Φ(z) = 0.5 * (1 + erf(z / √2))
fn std_norm_cdf(z: f64) -> f64 {
return 0.5 * (1.0 + erf_f64(z / SQRT_2));
}
// Standard normal PDF: φ(z) = exp(-z²/2) / √(2π)
fn std_norm_pdf(z: f64) -> f64 {
return INV_SQRT_2PI * exp_f64(-0.5 * z * z);
}
// General normal CDF: Φ(x; μ, σ)
fn norm_cdf(x: f64, mu: f64, sigma: f64) -> f64 {
let z = (x - mu) / sigma;
return std_norm_cdf(z);
}
// General normal PDF: φ(x; μ, σ)
fn norm_pdf(x: f64, mu: f64, sigma: f64) -> f64 {
let z = (x - mu) / sigma;
return std_norm_pdf(z) / sigma;
}
@compute @workgroup_size(256)
fn main(@builtin(global_invocation_id) global_id: vec3<u32>) {
let idx = global_id.x;
if (idx >= params.size) {
return;
}
let x = input[idx];
if (params.mode == 0u) {
// CDF mode
output[idx] = norm_cdf(x, params.mu, params.sigma);
} else {
// PDF mode
output[idx] = norm_pdf(x, params.mu, params.sigma);
}
}