// SPDX-License-Identifier: AGPL-3.0-or-later
// Pearson correlation coefficient for batch vectors — f64 precision
// r = Σ(x-μx)(y-μy) / (σx·σy·n)
//
// Input: two vectors x and y of same length (f64)
// Output: single correlation coefficient per pair
//
// Applications: signal correlation, feature selection, portfolio analysis
// Reference: Standard statistical formula
@group(0) @binding(0) var<storage, read> x: array<f64>;
@group(0) @binding(1) var<storage, read> y: array<f64>;
@group(0) @binding(2) var<storage, read_write> output: array<f64>;
@group(0) @binding(3) var<uniform> params: Params;
struct Params {
size: u32, // Length of each vector
num_pairs: u32, // Number of (x,y) vector pairs
stride: u32, // Stride between vectors (typically == size)
}
@compute @workgroup_size(256)
fn main(@builtin(global_invocation_id) global_id: vec3<u32>) {
let idx = global_id.x;
if (idx >= params.num_pairs) {
return;
}
let x_offset = idx * params.stride;
let y_offset = idx * params.stride;
let size = params.size;
// Compute means
var sum_x: f64 = f64(0.0);
var sum_y: f64 = f64(0.0);
for (var i = 0u; i < size; i = i + 1u) {
sum_x = sum_x + x[x_offset + i];
sum_y = sum_y + y[y_offset + i];
}
let mean_x = sum_x / f64(size);
let mean_y = sum_y / f64(size);
// Compute covariance and standard deviations
var cov: f64 = f64(0.0);
var var_x: f64 = f64(0.0);
var var_y: f64 = f64(0.0);
for (var i = 0u; i < size; i = i + 1u) {
let dx = x[x_offset + i] - mean_x;
let dy = y[y_offset + i] - mean_y;
cov = cov + dx * dy;
var_x = var_x + dx * dx;
var_y = var_y + dy * dy;
}
let denom = sqrt(var_x * var_y);
if (denom < f64(1e-15)) {
output[idx] = f64(0.0); // Avoid division by zero
return;
}
let result = cov / denom;
output[idx] = result;
}