byte search v1

.gitignore

@@ -0,0 +1 @@
+/target

Cargo.lock

@@ -0,0 +1,7 @@
+# This file is automatically @generated by Cargo.
+# It is not intended for manual editing.
+version = 4
+
+[[package]]
+name = "simd-fun"
+version = "0.1.0"

Cargo.toml

@@ -0,0 +1,6 @@
+[package]
+name = "simd-fun"
+version = "0.1.0"
+edition = "2024"
+
+[dependencies]

rust-toolchain.toml

@@ -0,0 +1,2 @@
+[toolchain]
+channel = "nightly"

src/main.rs

@@ -0,0 +1,65 @@
+#![feature(test, portable_simd)]
+
+extern crate test;
+
+use std::{
+    env,
+    io::{self, Read},
+    simd::prelude::*,
+};
+
+fn main() {
+    let needle = *env::args()
+        .nth(1)
+        .expect("needle argument is required")
+        .as_bytes()
+        .first()
+        .expect("needle should have at least one byte come on");
+    let mut haystack = Vec::new();
+    let _ = io::stdin().read_to_end(&mut haystack).unwrap();
+    println!(
+        "normal: {:?}\nsimd: {:?}",
+        find(needle, &haystack),
+        find_simd(needle, &haystack)
+    );
+}
+
+const CHUNK_SIZE: usize = 64;
+
+// three times slower than the naive implementation because
+// compilers eat this kind of autovectorization for breakfast
+fn find_simd(needle: u8, haystack: &[u8]) -> Option<usize> {
+    let simd_needle = Simd::<u8, CHUNK_SIZE>::splat(needle);
+    for (i, chunk) in haystack.chunks(CHUNK_SIZE).enumerate() {
+        let chunk = Simd::<u8, CHUNK_SIZE>::load_or_default(chunk);
+        let mask = chunk.simd_eq(simd_needle);
+        if let Some(found) = mask.first_set() {
+            return Some(i * CHUNK_SIZE + found);
+        }
+    }
+    None
+}
+
+fn find(needle: u8, haystack: &[u8]) -> Option<usize> {
+    haystack.iter().position(|&b| b == needle)
+}
+
+#[cfg(test)]
+mod bench {
+    use super::*;
+    use std::process::Termination;
+    use test::Bencher;
+
+    const HAYSTACK: &[u8] = include_bytes!("/home/mbk/Documents/shakespeare.txt");
+    const NEEDLE: u8 = b'%';
+
+    #[bench]
+    fn bench_simd(b: &mut Bencher) -> impl Termination {
+        b.iter(|| assert_eq!(Some(5_576_727), find_simd(NEEDLE, HAYSTACK)));
+    }
+
+    #[bench]
+    fn bench_normal(b: &mut Bencher) -> impl Termination {
+        b.iter(|| assert_eq!(Some(5_576_727), find(NEEDLE, HAYSTACK)));
+    }
+}