Matt Hess
714721874e
- Full RandomX implementation (light mode + full mode) - WASM-accelerated Argon2d cache init (37x faster than pure JS) - WASM-accelerated SuperscalarHash with SIMD support - Parallel dataset generation using worker threads (8 workers) - Light mode: ~4s init, Full mode: ~8min init - Mining utilities (difficulty calculation, block construction) - Progress callbacks for long-running operations Performance targets 2023+ platforms (WASM SIMD: Chrome 91+, Firefox 89+, Safari 16.4+, Node 16.4+)
87 lines
3.2 KiB
JavaScript
87 lines
3.2 KiB
JavaScript
#!/usr/bin/env node
|
|
/**
|
|
* Test: Parallel Dataset Generation
|
|
*
|
|
* Tests the worker-based parallel dataset generation.
|
|
*
|
|
* Usage:
|
|
* source ~/.bash_profile && bun test/test-parallel-dataset.js
|
|
*/
|
|
|
|
import { cpus } from 'os';
|
|
import { generateDatasetParallel } from '../src/randomx/parallel-dataset.js';
|
|
import { RANDOMX_DATASET_ITEM_COUNT } from '../src/randomx/config.js';
|
|
|
|
console.log('Parallel Dataset Generation Test');
|
|
console.log('=================================\n');
|
|
|
|
const numCpus = cpus().length;
|
|
console.log(`Available CPUs: ${numCpus}`);
|
|
console.log(`Dataset items: ${RANDOMX_DATASET_ITEM_COUNT.toLocaleString()}`);
|
|
console.log(`Dataset size: ${((RANDOMX_DATASET_ITEM_COUNT * 64) / 1024 / 1024 / 1024).toFixed(2)} GB\n`);
|
|
|
|
// For testing, we'll use a smaller dataset
|
|
// Uncomment the full run for production testing
|
|
const testItems = 100000; // 100K items for quick test
|
|
|
|
console.log(`Test mode: ${testItems.toLocaleString()} items`);
|
|
console.log(`Using ${numCpus} workers\n`);
|
|
|
|
const testKey = new TextEncoder().encode('test key');
|
|
|
|
try {
|
|
// Override item count for testing
|
|
const originalCount = RANDOMX_DATASET_ITEM_COUNT;
|
|
|
|
// Monkey-patch the config for testing
|
|
// In production, remove this and use full dataset
|
|
const config = await import('../src/randomx/config.js');
|
|
|
|
console.log('Starting parallel generation...\n');
|
|
|
|
const startTime = Date.now();
|
|
|
|
// Note: For full production use, remove the test limit
|
|
// For now, we're testing with a subset
|
|
const dataset = await generateDatasetParallel(testKey, {
|
|
workers: numCpus,
|
|
itemCount: testItems, // Use test subset instead of full 34M items
|
|
onProgress: (stage, percent, details) => {
|
|
if (typeof process !== 'undefined' && process.stdout) {
|
|
const bar = '█'.repeat(Math.floor(percent / 5)) + '░'.repeat(20 - Math.floor(percent / 5));
|
|
|
|
if (stage === 'cache') {
|
|
const info = details.pass !== undefined
|
|
? `pass ${details.pass + 1}/3, slice ${details.slice + 1}/4`
|
|
: details.message || '';
|
|
process.stdout.write(`\rCache: [${bar}] ${percent}% ${info}`.padEnd(80));
|
|
} else if (stage === 'programs') {
|
|
process.stdout.write(`\rPrograms: [${bar}] ${percent}% ${details.message || ''}`.padEnd(80));
|
|
} else if (stage === 'dataset') {
|
|
const info = details.eta !== undefined
|
|
? `${details.itemsPerSec?.toLocaleString()} items/s, ETA: ${Math.floor(details.eta / 60)}m ${details.eta % 60}s`
|
|
: details.message || '';
|
|
process.stdout.write(`\rDataset: [${bar}] ${percent}% ${info}`.padEnd(80));
|
|
} else if (stage === 'complete') {
|
|
process.stdout.write('\r' + ' '.repeat(80) + '\r');
|
|
console.log(`\nDataset initialized in ${details.totalTime?.toFixed(1)}s`);
|
|
}
|
|
}
|
|
}
|
|
});
|
|
|
|
const totalTime = (Date.now() - startTime) / 1000;
|
|
|
|
console.log(`\nTotal time: ${totalTime.toFixed(1)}s`);
|
|
console.log(`Dataset size: ${(dataset.length / 1024 / 1024).toFixed(2)} MB`);
|
|
|
|
// Verify first few items
|
|
console.log('\nFirst item (hex):');
|
|
console.log(Buffer.from(dataset.slice(0, 64)).toString('hex'));
|
|
|
|
} catch (error) {
|
|
console.error('Error:', error.message);
|
|
console.error(error.stack);
|
|
process.exit(1);
|
|
}
|