Skip to main content

Tungsten Project for Apache Spark

What is Tungsten?‚Äč

The Tungsten Project is a major initiative to enhance Apache Spark's execution engine. Its goal is to significantly boost the efficiency of memory and CPU usage for Spark applications, bringing performance closer to the limits of modern hardware. The project includes several key initiatives, such as memory management and binary processing, cache-aware computation, code generation, and more. By leveraging these techniques, Tungsten reduces overhead and improves performance by reducing CPU calls and optimizing data storage and retrieval.

With a focus on CPU efficiency, the Tungsten Project addresses the growing challenge of bottlenecked CPU and memory use in big data workloads. Recent research has shown that these factors are increasingly limiting performance, making Tungsten's improvements more important than ever.

ON THIS PAGE