diff --git a/data/examples/17.txt b/data/examples/17.txt
new file mode 100644
index 000000000..e69de29
diff --git a/src/bin/17.rs b/src/bin/17.rs
new file mode 100644
index 000000000..375644a
--- /dev/null
+++ b/src/bin/17.rs
@@ -0,0 +1,72 @@
+use hashbrown::HashSet;
+use itertools::Itertools;
+
+fn parse_input(input: &str) -> (usize, Vec<usize>) {
+    let mut input_iter = input.lines();
+
+    let ua = input_iter.next().unwrap();
+    let uprogram = input_iter.last().unwrap();
+
+    let a = ua.strip_prefix("Register A: ").unwrap().parse().unwrap();
+    let program = uprogram
+        .strip_prefix("Program: ")
+        .unwrap()
+        .split(',')
+        .filter_map(|x| x.parse().ok())
+        .collect();
+
+    (a, program)
+}
+
+fn quick_exec(a: usize) -> usize {
+    // hardcoded for my input
+    // below operation was produced by reducing below steps
+    //
+    // B = A % 8
+    // B = B ^ 7
+    // C = A >> B
+    // B = B ^ 7
+    // A = A >> 3
+    // B = B ^ C
+    // out(B % 8)
+    // jmp(0)
+    //
+    // remaining mutation after returning is A >> 3
+    ((a % 8) ^ (a >> ((a % 8) ^ 7))) % 8
+}
+
+pub fn part_one(input: &str) -> Option<String> {
+    let (mut a, _) = parse_input(input);
+    let mut v = vec![];
+
+    while a != 0 {
+        v.push(quick_exec(a));
+        a >>= 3;
+    }
+
+    v.iter().join(",").into()
+}
+
+pub fn part_two(input: &str) -> Option<usize> {
+    let (_, program) = parse_input(input);
+
+    let mut n_candidates = HashSet::new();
+    let mut candidates = HashSet::new();
+    candidates.insert(0);
+
+    for n in program.iter().copied().rev() {
+        for candidate in candidates.iter() {
+            n_candidates.extend(
+                (0..8)
+                    .map(|shift| (candidate << 3) + shift)
+                    .filter(|&m| quick_exec(m) == n),
+            );
+        }
+        candidates.clear();
+        candidates.extend(n_candidates.drain())
+    }
+
+    candidates.iter().min().copied()
+}
+
+aoc::solution!(17);