one more bench improvement

vortex-fastlanes/benches/bitpacking.rs

@@ -1,4 +1,5 @@
 use criterion::{black_box, criterion_group, criterion_main, Criterion};
+use fastlanez::TryBitPack;
 use rand::distributions::Uniform;
 use rand::{thread_rng, Rng};
 use vortex_fastlanes::{bitpack_primitive, unpack_primitive, unpack_single_primitive};
@@ -23,21 +24,32 @@ fn pack_unpack(c: &mut Criterion) {
     let bits: usize = 8;
     let values = values(1_000_000, bits);
 
-    c.bench_function("bitpack_primitive", |b| {
+    c.bench_function("bitpack_1M", |b| {
         b.iter(|| black_box(bitpack_primitive(&values, bits)));
     });
 
     let packed = bitpack_primitive(&values, bits);
-    c.bench_function("unpack_primitive", |b| {
+    c.bench_function("unpack_1M", |b| {
         b.iter(|| black_box(unpack_primitive::<u32>(&packed, bits, values.len())));
     });
 
-    c.bench_function("unpack_all_singles", |b| {
+    c.bench_function("unpack_1M_singles", |b| {
         b.iter(|| black_box(unpack_singles(&packed, 8, values.len())));
     });
 
-    c.bench_function("unpack_single_primitive", |b| {
-        b.iter(|| black_box(unsafe { unpack_single_primitive::<u32>(&packed, 8, 0) }));
+    // 1024 elements pack into `128 * bits` bytes
+    let packed_1024 = &packed[0..128 * bits];
+    let mut output: Vec<u32> = Vec::with_capacity(1024);
+    c.bench_function("unpack_1024", |b| {
+        b.iter(|| {
+            output.clear();
+            TryBitPack::try_unpack_into(packed_1024, bits, &mut output).unwrap();
+            black_box(output[0])
+        })
+    });
+
+    c.bench_function("unpack_single", |b| {
+        b.iter(|| black_box(unsafe { unpack_single_primitive::<u32>(packed_1024, 8, 0) }));
     });
 }