From 2ed2157ce347bec8468d05fe9c73417ced1340f8 Mon Sep 17 00:00:00 2001
From: Sylvestre Ledru <sylvestre@debian.org>
Date: Tue, 23 Sep 2025 23:00:02 +0200
Subject: [PATCH] numfmt: add benchmark

---
 Cargo.lock                            |   2 +
 src/uu/numfmt/Cargo.toml              |   9 ++
 src/uu/numfmt/benches/numfmt_bench.rs | 120 ++++++++++++++++++++++++++
 3 files changed, 131 insertions(+)
 create mode 100644 src/uu/numfmt/benches/numfmt_bench.rs

diff --git a/Cargo.lock b/Cargo.lock
index e5f2b501c..d2cf4d7af 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -3731,7 +3731,9 @@ name = "uu_numfmt"
 version = "0.2.2"
 dependencies = [
  "clap",
+ "codspeed-divan-compat",
  "fluent",
+ "tempfile",
  "thiserror 2.0.16",
  "uucore",
 ]
diff --git a/src/uu/numfmt/Cargo.toml b/src/uu/numfmt/Cargo.toml
index 5cfd2dc46..177f2e3b8 100644
--- a/src/uu/numfmt/Cargo.toml
+++ b/src/uu/numfmt/Cargo.toml
@@ -23,6 +23,15 @@ uucore = { workspace = true, features = ["parser", "ranges"] }
 thiserror = { workspace = true }
 fluent = { workspace = true }
 
+[dev-dependencies]
+divan = { workspace = true }
+tempfile = { workspace = true }
+uucore = { workspace = true, features = ["benchmark"] }
+
 [[bin]]
 name = "numfmt"
 path = "src/main.rs"
+
+[[bench]]
+name = "numfmt_bench"
+harness = false
diff --git a/src/uu/numfmt/benches/numfmt_bench.rs b/src/uu/numfmt/benches/numfmt_bench.rs
new file mode 100644
index 000000000..ee09b7d0f
--- /dev/null
+++ b/src/uu/numfmt/benches/numfmt_bench.rs
@@ -0,0 +1,120 @@
+// This file is part of the uutils coreutils package.
+//
+// For the full copyright and license information, please view the LICENSE
+// file that was distributed with this source code.
+
+use divan::{Bencher, black_box};
+use tempfile::TempDir;
+use uu_numfmt::uumain;
+use uucore::benchmark::{create_test_file, run_util_function};
+
+/// Generate numeric data for benchmarking
+fn generate_numbers(count: usize) -> String {
+    (1..=count)
+        .map(|n| n.to_string())
+        .collect::<Vec<_>>()
+        .join("\n")
+}
+
+/// Setup benchmark environment with test data
+fn setup_benchmark(data: String) -> (TempDir, String) {
+    let temp_dir = tempfile::tempdir().unwrap();
+    let file_path = create_test_file(data.as_bytes(), temp_dir.path());
+    let file_path_str = file_path.to_str().unwrap().to_string();
+    (temp_dir, file_path_str)
+}
+
+/// Benchmark SI formatting with different number counts
+#[divan::bench(args = [1_000_000])]
+fn numfmt_to_si(bencher: Bencher, count: usize) {
+    let (_temp_dir, file_path_str) = setup_benchmark(generate_numbers(count));
+
+    bencher.bench(|| {
+        black_box(run_util_function(uumain, &["--to=si", &file_path_str]));
+    });
+}
+
+/// Benchmark SI formatting with precision format
+#[divan::bench(args = [1_000_000])]
+fn numfmt_to_si_precision(bencher: Bencher, count: usize) {
+    let (_temp_dir, file_path_str) = setup_benchmark(generate_numbers(count));
+
+    bencher.bench(|| {
+        black_box(run_util_function(
+            uumain,
+            &["--to=si", "--format=%.6f", &file_path_str],
+        ));
+    });
+}
+
+/// Benchmark IEC (binary) formatting
+#[divan::bench(args = [1_000_000])]
+fn numfmt_to_iec(bencher: Bencher, count: usize) {
+    let (_temp_dir, file_path_str) = setup_benchmark(generate_numbers(count));
+
+    bencher.bench(|| {
+        black_box(run_util_function(uumain, &["--to=iec", &file_path_str]));
+    });
+}
+
+/// Benchmark parsing from SI format back to raw numbers
+#[divan::bench(args = [1_000_000])]
+fn numfmt_from_si(bencher: Bencher, count: usize) {
+    // Generate SI formatted data (e.g., "1.0K", "2.0K", etc.)
+    let data = (1..=count)
+        .map(|n| format!("{:.1}K", n as f64 / 1000.0))
+        .collect::<Vec<_>>()
+        .join("\n");
+    let (_temp_dir, file_path_str) = setup_benchmark(data);
+
+    bencher.bench(|| {
+        black_box(run_util_function(uumain, &["--from=si", &file_path_str]));
+    });
+}
+
+/// Benchmark large numbers with SI formatting
+#[divan::bench(args = [1_000_000])]
+fn numfmt_large_numbers_si(bencher: Bencher, count: usize) {
+    // Generate larger numbers (millions to billions range)
+    let data = (1..=count)
+        .map(|n| (n * 1_000_000).to_string())
+        .collect::<Vec<_>>()
+        .join("\n");
+    let (_temp_dir, file_path_str) = setup_benchmark(data);
+
+    bencher.bench(|| {
+        black_box(run_util_function(uumain, &["--to=si", &file_path_str]));
+    });
+}
+
+/// Benchmark different padding widths
+#[divan::bench(args = [(1_000_000, 5), (1_000_000, 50)])]
+fn numfmt_padding(bencher: Bencher, (count, padding): (usize, usize)) {
+    let (_temp_dir, file_path_str) = setup_benchmark(generate_numbers(count));
+    let padding_arg = format!("--padding={padding}");
+
+    bencher.bench(|| {
+        black_box(run_util_function(
+            uumain,
+            &["--to=si", &padding_arg, &file_path_str],
+        ));
+    });
+}
+
+/// Benchmark round modes with SI formatting
+#[divan::bench(args = [("up", 100_000), ("down", 1_000_000), ("towards-zero", 1_000_000)])]
+fn numfmt_round_modes(bencher: Bencher, (round_mode, count): (&str, usize)) {
+    let (_temp_dir, file_path_str) = setup_benchmark(generate_numbers(count));
+    let round_arg = format!("--round={round_mode}");
+
+    bencher.bench(|| {
+        black_box(run_util_function(
+            uumain,
+            &["--to=si", &round_arg, &file_path_str],
+        ));
+    });
+}
+
+fn main() {
+    divan::main();
+}