Add parallel_for to ThreadPool and ParallelFor convenience function

[Copilot]

Adds a parallel_for interface to the custom async task system, enabling index-based parallel iteration distributed across the thread pool.

Changes

• ThreadPool::parallel_for(begin, end, func) — splits [begin, end) into min(count, numWorkers) balanced chunks, submits each via async(), blocks until complete. Falls back to serial for 0 workers or single element.
• ParallelFor(begin, end, func) — free function forwarding to the global TaskExecutor singleton's pool.
• ThreadPool::NumWorkers() — exposes worker thread count.
• Tests — basic, empty/reversed ranges, single element, large range (10k), convenience function.

Usage

// Direct pool usage
sky::ThreadPool pool(4);
pool.parallel_for(0, N, [&data](int i) {
    data[i] = compute(i);
});

// Via global executor
sky::ParallelFor(0, N, [&data](int i) {
    data[i] = compute(i);
});

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.