Ahead of its January 19th beta launch, the new ETL engine PardoX v0.1 is moving from Windows to Linux. The developer found Microsoft's OS created a 'sticky floor' that throttled high-performance Rust code. By switching to Ubuntu 24.04 on dedicated AMD hardware, the engineer eliminated kernel-level latency, allowing the engine to finally access the raw power of the CPU and GPU.

In a direct speed test against the popular Polars library, PardoX processed 640 million rows in 182 seconds. That result beat Polars by nearly 22 seconds while consuming only 1.13 GB of RAM. The engine also introduces a native PostgreSQL connector that bypasses Python drivers, cutting memory use by 70% and eliminating the translation overhead that usually slows down data ingestion.

Future updates will bring native support for MySQL, SQL Server, and Arrow Flight SQL for cloud databases. The developer is also finalizing a custom .prdx persistence format that writes at 3.5 GB/s and a vectorized arithmetic engine for instant math operations. The goal is a universal tool that works across Python, PHP, and Go, releasing next week.