Add benchmark for neuquant

There were no benchmarks for neuquant, and it's worthwhile to figure out what consumes time when
quantizing the colors. This creates two benchmarks, one for the creation of the quantization,
and one for the searching of the nearest color. It tests on a static image which is mostly green
and red.

Author: JulianKnodt

Cargo.toml

@@ -6,3 +6,8 @@ authors = ["nwin <[email protected]>"]
 readme = "README.md"
 description = "Color quantization library to reduce n colors to 256 colors."
 repository = "https://github.com/image-rs/color_quant.git"
+
+[[bench]]
+name = "neuquant"
+path = "benches/neuquant.rs"
+test = false

benches/neuquant.rs

@@ -0,0 +1,55 @@
+#![feature(test)]
+
+extern crate test;
+
+use color_quant::NeuQuant;
+use test::{black_box, Bencher};
+
+fn make_pixels() -> Vec<u8> {
+    let mut pixels = vec![0; 4 * 256 * 256];
+    let mut r: u8 = 0;
+    let mut g: u8 = 0;
+    let mut b: u8 = 0;
+
+    for (i, p) in pixels.iter_mut().enumerate() {
+        match i % 4 {
+            0 => *p = r,
+            1 => *p = g,
+            2 => {
+                *p = b;
+                if let Some(next_r) = r.checked_add(1) {
+                    r = next_r;
+                    continue;
+                }
+                r = 0;
+                if let Some(next_g) = g.checked_add(1) {
+                    g = next_g;
+                    continue;
+                }
+                g = 0;
+                b += 1
+            }
+            3 => *p = 255,
+            _ => unreachable!(),
+        }
+    }
+    pixels
+}
+
+#[bench]
+fn bench_map_pixel(b: &mut Bencher) {
+    let pixels = make_pixels();
+    let nq = NeuQuant::new(10, 256, &pixels);
+    let mut i = 0u8;
+    b.iter(|| {
+        i = i.wrapping_add(1);
+        let mut color = black_box([123, i, 4, 255]);
+        nq.map_pixel(&mut color[..]);
+    })
+}
+
+#[bench]
+fn bench_new_nq(b: &mut Bencher) {
+    let pixels = make_pixels();
+    b.iter(|| NeuQuant::new(10, 128, &pixels));
+}