Performance Guide - UUID Library v1.3.0

Understanding performance characteristics and optimization strategies for high-performance UUID generation.

Benchmark Comparisons

Performance Optimizations

1. Efficient Random Generation

// Modern Random\Randomizer (11% faster than random_bytes)
$randomizer = new Random\Randomizer();
$bytes = $randomizer->getBytes(16);
 
// Optimized bit manipulation
$bytes[6] = chr(ord($bytes[6]) & 0x0f | 0x40); // Version 4
$bytes[8] = chr(ord($bytes[8]) & 0x3f | 0x80); // Variant 10

2. Optimized Time Handling

// hrtime() for nanosecond precision (faster than microtime)
$timestamp = hrtime(true);
$uuid_time = ($timestamp / 100) + 0x01b21dd213814000;
 
// Direct hex conversion without intermediate string operations
$hex = str_pad(dechex($uuid_time), 15, '0', STR_PAD_LEFT);
return substr($hex, 7, 8) . '-' .
       substr($hex, 3, 4) . '-' .
       '1' . substr($hex, 0, 3) . '-' .
       substr($hex, 15, 4) . '-' .
       substr($hex, 20, 12);

3. Memory Efficiency

Best Practices for Performance

use Webpatser\Uuid\Uuid;
 
// ✅ DO: Batch generation for better performance
$uuids = [];
$start = microtime(true);
for ($i = 0; $i < 10000; $i++) {
    $uuids[] = Uuid::v4();
}
$elapsed = (microtime(true) - $start) * 1000;
echo "Generated 10k UUIDs in {$elapsed:.2f}ms";

Memory Usage

// Efficient batch generation
$uuids = [];
$start = memory_get_usage();
for ($i = 0; $i < 1000; $i++) {
    $uuids[] = Uuid::v7(); // Time-ordered, better for databases
}
$memory = memory_get_usage() - $start;
echo "Memory used for 1000 UUIDs: " . number_format($memory / 1024, 1) . " KB";
 
// Database insertion performance
$stmt = $pdo->prepare("INSERT INTO users (id, email, data) VALUES (?, ?, ?)");
for ($i = 0; $i < 10000; $i++) {
    $stmt->execute([$uuids[$i], "user{$i}@example.com", $data]);
}

Version-Specific Performance

Real-World Performance Tips

Database Optimization

• Use v7 for primary keys: Sequential ordering improves B-tree indexing
• Store as BINARY(16): 50% smaller than VARCHAR(36)
• Index strategically: v7 UUIDs cluster well in indexes

Application Optimization

• Cache namespace UUIDs: Don't regenerate the same namespace UUID
• Use v4 for session IDs: Maximum entropy with good performance
• Batch operations: Generate UUIDs in loops rather than one-by-one

Profiling Your Application

// Simple benchmarking function
function benchmarkUuid($version, $iterations = 10000) {
    $start = microtime(true);
    $memory_start = memory_get_usage();
 
    for ($i = 0; $i < $iterations; $i++) {
        switch ($version) {
            case 4: Uuid::v4(); break;
            case 7: Uuid::v7(); break;
            case 1: Uuid::generate(1); break;
        }
    }
 
    $time = (microtime(true) - $start) * 1000;
    $memory = memory_get_usage() - $memory_start;
 
    echo "v{$version}: {$time:.2f}ms, " .
         number_format($memory / 1024, 1) . "KB\n";
}
 
benchmarkUuid(4);
benchmarkUuid(7);
benchmarkUuid(1);