benchmark PySet::new

benches/bench_set.rs

@@ -4,6 +4,20 @@ use pyo3::prelude::*;
 use pyo3::types::PySet;
 use std::collections::{BTreeSet, HashSet};
 
+fn set_new(b: &mut Bencher<'_>) {
+    Python::with_gil(|py| {
+        const LEN: usize = 100_000;
+        // Create Python objects up-front, so that the benchmark doesn't need to include
+        // the cost of allocating LEN Python integers
+        let elements: Vec<PyObject> = (0..LEN).into_iter().map(|i| i.into_py(py)).collect();
+        b.iter(|| {
+            let pool = unsafe { py.new_pool() };
+            PySet::new(py, &elements).unwrap();
+            drop(pool);
+        });
+    });
+}
+
 fn iter_set(b: &mut Bencher<'_>) {
     Python::with_gil(|py| {
         const LEN: usize = 100_000;
@@ -44,6 +58,7 @@ fn extract_hashbrown_set(b: &mut Bencher<'_>) {
 }
 
 fn criterion_benchmark(c: &mut Criterion) {
+    c.bench_function("set_new", set_new);
     c.bench_function("iter_set", iter_set);
     c.bench_function("extract_hashset", extract_hashset);
     c.bench_function("extract_btreeset", extract_btreeset);