first implementation but far to be optimal

src/obikseq/benches/superkmer.rs

@@ -0,0 +1,95 @@
+use criterion::{BatchSize, BenchmarkId, Criterion, Throughput, criterion_group, criterion_main};
+use obikseq::superkmer::SuperKmer;
+
+const LENGTHS: &[usize] = &[1, 4, 8, 16, 40, 64, 128, 255, 256];
+
+fn make_ascii(len: usize) -> Vec<u8> {
+    (0..len).map(|i| b"ACGT"[i % 4]).collect()
+}
+
+fn bench_from_ascii(c: &mut Criterion) {
+    let mut group = c.benchmark_group("from_ascii");
+    for &len in LENGTHS {
+        let ascii = make_ascii(len);
+        group.throughput(Throughput::Bytes(len as u64));
+        group.bench_with_input(BenchmarkId::from_parameter(len), &ascii, |b, ascii| {
+            b.iter(|| SuperKmer::from_ascii(std::hint::black_box(ascii)));
+        });
+    }
+    group.finish();
+}
+
+fn bench_to_ascii(c: &mut Criterion) {
+    let mut group = c.benchmark_group("to_ascii");
+    for &len in LENGTHS {
+        let sk = SuperKmer::from_ascii(&make_ascii(len));
+        group.throughput(Throughput::Bytes(len as u64));
+        group.bench_with_input(BenchmarkId::from_parameter(len), &sk, |b, sk| {
+            b.iter(|| std::hint::black_box(sk).to_ascii());
+        });
+    }
+    group.finish();
+}
+
+fn bench_write_ascii(c: &mut Criterion) {
+    let mut group = c.benchmark_group("write_ascii");
+    for &len in LENGTHS {
+        let sk = SuperKmer::from_ascii(&make_ascii(len));
+        group.throughput(Throughput::Bytes(len as u64));
+        group.bench_with_input(BenchmarkId::from_parameter(len), &sk, |b, sk| {
+            let mut buf = Vec::with_capacity(len);
+            b.iter(|| {
+                buf.clear();
+                std::hint::black_box(sk).write_ascii(&mut buf);
+            });
+        });
+    }
+    group.finish();
+}
+
+fn bench_revcomp(c: &mut Criterion) {
+    let mut group = c.benchmark_group("revcomp");
+    for &len in LENGTHS {
+        let sk = SuperKmer::from_ascii(&make_ascii(len));
+        group.throughput(Throughput::Bytes(len as u64));
+        group.bench_with_input(BenchmarkId::from_parameter(len), &sk, |b, sk| {
+            b.iter_batched(
+                || sk.clone(),
+                |mut s| { std::hint::black_box(&mut s).revcomp(); s },
+                BatchSize::SmallInput,
+            );
+        });
+    }
+    group.finish();
+}
+
+fn bench_canonical(c: &mut Criterion) {
+    let mut group = c.benchmark_group("canonical");
+    for &len in LENGTHS {
+        let sk_rc  = SuperKmer::from_ascii(&vec![b'T'; len]);
+        let sk_fwd = SuperKmer::from_ascii(&vec![b'A'; len]);
+        group.throughput(Throughput::Bytes(len as u64));
+        group.bench_with_input(
+            BenchmarkId::new("needs_revcomp", len), &sk_rc, |b, sk| {
+                b.iter_batched(
+                    || sk.clone(),
+                    |s| std::hint::black_box(s).canonical(),
+                    BatchSize::SmallInput,
+                );
+            }
+        );
+        group.bench_with_input(
+            BenchmarkId::new("already_canonical", len), &sk_fwd, |b, sk| {
+                b.iter_batched(
+                    || sk.clone(),
+                    |s| std::hint::black_box(s).canonical(),
+                    BatchSize::SmallInput,
+                );
+            }
+        );
+    }
+    group.finish();
+}
+
+criterion_group!(benches, bench_from_ascii, bench_to_ascii, bench_write_ascii, bench_revcomp, bench_canonical);
+criterion_main!(benches);