Vectorize += for Symbols

This makes += ~8x faster, and encoding/decoding benchmarks ~15% faster

src/symbol.rs

@@ -52,10 +52,44 @@ impl Symbol {
 impl<'a> AddAssign<&'a Symbol> for Symbol {
     fn add_assign(&mut self, other: &'a Symbol) {
         assert_eq!(self.value.len(), other.value.len());
-        for i in 0..self.value.len() {
+        let self_ptr = self.value.as_mut_ptr() as *mut u64;
+        let other_ptr = other.value.as_ptr() as *const u64;
+        for i in 0..(self.value.len() / 8) {
+            unsafe {
+                *self_ptr.add(i) ^= *other_ptr.add(i);
+            }
+        }
+        let remainder = self.value.len() % 8;
+        for i in (self.value.len() - remainder)..self.value.len() {
             unsafe {
                 *self.value.get_unchecked_mut(i) += other.value.get_unchecked(i);
             }
         }
     }
 }
+
+#[cfg(test)]
+mod tests {
+    extern crate rand;
+
+    use symbol::tests::rand::Rng;
+    use symbol::Symbol;
+
+    #[test]
+    fn add_assign() {
+        let symbol_size = 17;
+        let mut data1: Vec<u8> = vec![0; symbol_size];
+        let mut data2: Vec<u8> = vec![0; symbol_size];
+        let mut result: Vec<u8> = vec![0; symbol_size];
+        for i in 0..symbol_size {
+            data1[i] = rand::thread_rng().gen();
+            data2[i] = rand::thread_rng().gen();
+            result[i] = data1[i] ^ data2[i];
+        }
+        let mut symbol1 = Symbol::new(data1);
+        let symbol2 = Symbol::new(data2);
+
+        symbol1 += &symbol2;
+        assert_eq!(result, symbol1.bytes());
+    }
+}