Pushing Java to the Limits: Processing a Billion Rows in Under 2 Seconds

Dive into an in-depth conference talk exploring how to optimize Java performance for processing massive datasets. Learn about the journey from a 4-minute reference implementation to parsing one billion rows of weather data in under two seconds. Discover advanced techniques including parallelism, memory-mapped files, SWAR (SIMD Within A Register), bit twiddling, branchless code, and mechanical sympathy. Explore the benefits of Graal native compilation and the controversial use of sun.misc.Unsafe. Gain insights into pushing Java to its limits and unlocking unprecedented processing speeds for big data applications.

Pushing Java to the Limits: Processing a Billion Rows in under 2 Seconds by Thomas Wuerthinger