Implement improvements: rayon parallelism, safer CLI, output file option, tests, CI workflow, tuned release profile

.github/workflows/ci.yml

@@ -0,0 +1,27 @@
+name: CI
+
+on:
+  push:
+    branches: [ "**" ]
+  pull_request:
+    branches: [ "**" ]
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Install Rust toolchain
+        uses: actions-rs/toolchain@v1
+        with:
+          toolchain: stable
+          profile: minimal
+          override: true
+      - name: Run cargo fmt check
+        run: cargo fmt -- --check
+      - name: Run clippy
+        run: cargo clippy --all-targets --all-features -- -D warnings
+      - name: Run tests
+        run: cargo test --verbose
+      - name: Build release
+        run: cargo build --release --verbose

Cargo.toml

@@ -1,8 +1,17 @@
 [package]
 name = "pi"
 version = "0.1.0"
-edition = "2024"
+edition = "2021"
 
 [dependencies]
 clap = { version = "4.5.4", features = ["derive"] }
 rug = "1.24.1"
+rayon = "1.7"
+
+[dev-dependencies]
+criterion = "0.4"
+
+[profile.release]
+opt-level = 3
+lto = true
+codegen-units = 1

README.md

@@ -1,20 +1,18 @@
 # Pi Calculator
 
-This is a multi-threaded Rust program that calculates the first n digits of Pi using the Bailey–Borwein–Plouffe (BBP) formula. It uses arbitrary-precision arithmetic to ensure the accuracy of the calculated digits.
+This is a multi-threaded Rust program that calculates the first n digits of Pi using the Bailey–Borwein–Plouffe (BBP) formula. It uses arbitrary-precision arithmetic (rug) and parallelism (rayon).
 
-## Features
+## Improvements in this branch
 
-*   Calculates the first n digits of Pi.
-*   Multi-threaded to speed up the calculation.
-*   Configurable number of threads.
-*   Uses the BBP algorithm.
-*   High-precision calculation using the `rug` crate.
+* Parallelized BBP summation with rayon for better thread control and load balancing.
+* Safer argument validation and error handling (avoids unwraps on runtime errors).
+* Optional output-to-file support.
+* Added CI workflow to run formatting, clippy, tests and build on push/PR.
+* Release profile tuned for better optimized builds (LTO, opt-level=3).
 
 ## Building
 
-To build the program, you need to have Rust and Cargo installed. You can install them from [https://rustup.rs/](https://rustup.rs/).
-
-Once you have Rust and Cargo installed, you can build the program with the following command:
+Requires Rust and Cargo. Build with:
 
 ```bash
 cargo build --release
@@ -22,26 +20,28 @@ cargo build --release
 
 ## Usage
 
-To run the program, you can use the following command:
-
 ```bash
 ./target/release/pi <N> [OPTIONS]
 ```
 
-### Arguments
+Arguments
 
-*   `<N>`: The number of digits of Pi to calculate.
+* `<N>`: Number of digits after the decimal point to calculate.
 
-### Options
+Options
 
-*   `-t`, `--threads <THREADS>`: The number of threads to use. Defaults to 4.
-*   `-h`, `--help`: Print help information.
-*   `-V`, `--version`: Print version information.
+* `-t`, `--threads <THREADS>`: Number of threads to use (default 4).
+* `-o`, `--output <FILE>`: Write output to FILE instead of stdout.
+* `-h`, `--help`: Print help.
 
-### Example
+Example
 
-To calculate the first 1000 digits of Pi using 8 threads, you can run the following command:
+Calculate 1000 digits using 8 threads and write to a file:
 
 ```bash
-./target/release/pi 1000 -t 8
+./target/release/pi 1000 -t 8 -o pi1000.txt
 ```
+
+Notes
+
+For very large numbers of digits, using a decimal-friendly algorithm such as Chudnovsky (with binary splitting) will be far faster and more memory-efficient than BBP; consider switching to Chudnovsky for production-grade large computations.

src/main.rs

@@ -1,19 +1,28 @@
 use clap::Parser;
-use rug::{Float, ops::Pow};
-use std::thread;
+use rug::Float;
+use rug::ops::Pow;
+use rayon::prelude::*;
+use std::fs::File;
+use std::io::Write;
+use std::path::PathBuf;
 
 #[derive(Parser, Debug)]
 #[command(author, version, about, long_about = None)]
 struct Args {
-    /// Number of digits of Pi to calculate
+    /// Number of digits of Pi to calculate (digits after the decimal point)
     n: u32,
 
     /// Number of threads to use
     #[arg(short, long, default_value_t = 4)]
     threads: usize,
+
+    /// Optional output file (writes result there if provided)
+    #[arg(short, long)]
+    output: Option<PathBuf>,
 }
 
 fn bbp_term(k: u32, prec: u32) -> Float {
+    // Compute one BBP term at precision `prec`.
     let mut term = Float::with_val(prec, 4);
     term /= Float::with_val(prec, 8 * k + 1);
 
@@ -30,52 +39,95 @@ fn bbp_term(k: u32, prec: u32) -> Float {
     term -= term4;
 
     let sixteen = Float::with_val(prec, 16);
-    term /= sixteen.pow(k);
+    term /= sixteen.pow(k as i32);
 
     term
 }
 
-fn main() {
-    let args = Args::parse();
-    let n = args.n;
-    let num_threads = args.threads;
-
-    // Precision for rug::Float. We need a bit more than n decimal digits.
-    // log2(10) is approx 3.32. So, we need n * 3.32 bits.
-    let prec = (n as f64 * 3.33).ceil() as u32 + 10;
-
-    let num_terms = n + 5; // Use more terms for better accuracy
-    let terms_per_thread = (num_terms + num_threads as u32 - 1) / num_threads as u32;
-
-    let mut handles = vec![];
-
-    for i in 0..num_threads {
-        let start = i as u32 * terms_per_thread;
-        let end = ((i + 1) as u32 * terms_per_thread).min(num_terms);
-        let handle = thread::spawn(move || {
-            let mut partial_sum = Float::with_val(prec, 0);
-            for k in start..end {
-                partial_sum += bbp_term(k, prec);
+/// Calculate Pi to `n` decimal digits using a parallelized BBP summation.
+/// Returns a decimal string containing Pi truncated to `n` digits after the decimal point.
+pub fn calculate_pi(n: u32, num_threads: usize) -> Result<String, String> {
+    if n == 0 {
+        return Err("n must be > 0".into());
     }
-            partial_sum
+    if num_threads == 0 {
+        return Err("threads must be > 0".into());
+    }
+
+    // Bits of precision: log2(10) ~= 3.321928. Add some guard bits.
+    let prec = (n as f64 * 3.3219280948873626).ceil() as u32 + 20;
+
+    // BBP converges in base-16; use a modest overestimate for term count.
+    let num_terms = (n as usize / 1) + 20; // conservative
+
+    // Use rayon thread pool to control threads for parallel work.
+    let pool = rayon::ThreadPoolBuilder::new()
+        .num_threads(num_threads)
+        .build()
+        .map_err(|e| format!("Failed to build thread pool: {}", e))?;
+
+    let pi = pool.install(|| {
+        // Parallel iterator over term indices.
+        (0..num_terms as u32)
+            .into_par_iter()
+            .map(|k| bbp_term(k, prec))
+            .reduce(|| Float::with_val(prec, 0), |a, b| a + b)
     });
-        handles.push(handle);
-    }
 
-    let mut pi = Float::with_val(prec, 0);
-    for handle in handles {
-        pi += handle.join().unwrap();
-    }
+    // Convert to decimal string with a few extra digits for safe truncation.
+    let extra = 10usize;
+    let pi_string = pi.to_string_radix(10, Some(n as usize + extra));
 
-    // The user wants n digits after the decimal, and the output to be truncated.
-    // We can achieve this by getting a string with more precision and then truncating it.
-    let pi_string = pi.to_string_radix(10, Some(n as usize + 5)); // Get extra digits for accurate truncation
-    let dot_pos = pi_string.find('.').unwrap_or(1);
+    // Find dot safely and truncate or pad as needed.
+    let dot_pos = pi_string.find('.').unwrap_or(pi_string.len());
     let end_pos = dot_pos + 1 + n as usize;
 
-    if pi_string.len() > end_pos {
-        println!("Pi: {}", &pi_string[..end_pos]);
+    let out = if pi_string.len() >= end_pos {
+        pi_string[..end_pos].to_string()
     } else {
-        println!("Pi: {}", pi_string);
+        // If not enough digits were produced, pad with zeros.
+        let mut s = pi_string;
+        if !s.contains('.') {
+            s.push('.');
+        }
+        while s.len() < end_pos {
+            s.push('0');
+        }
+        s
+    };
+
+    Ok(out)
+}
+
+fn main() {
+    let args = Args::parse();
+
+    match calculate_pi(args.n, args.threads) {
+        Ok(pi_str) => {
+            if let Some(path) = args.output {
+                match File::create(&path) {
+                    Ok(mut f) => {
+                        if let Err(e) = writeln!(f, "{}", pi_str) {
+                            eprintln!("Failed to write to {}: {}", path.display(), e);
+                        }
+                    }
+                    Err(e) => eprintln!("Failed to create {}: {}", path.display(), e),
+                }
+            } else {
+                println!("Pi: {}", pi_str);
+            }
+        }
+        Err(e) => eprintln!("Error: {}", e),
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::calculate_pi;
+
+    #[test]
+    fn pi_10_digits() {
+        let pi = calculate_pi(10, 2).expect("calculation failed");
+        assert_eq!(pi, "3.1415926535");
     }
 }