Image of Charles Nutter

Charles Nutter

Red hat Inc.

Over the past 6 years, my focus on the JRuby project has usually been performance. I've worked on JRuby's bytecode compiler, optimizations to the runtime and core classes, improvements to memory use and allocation rates, and the removal of concurrency bottlenecks. Over that same time we've managed to increase JRuby's performance by orders of magnitude, in some cases approaching the speed of compiled Java.

This talk will go over the major performance leaps we've had during that time, starting with the disabling of ObjectSpace and proceeding through recent enhancements like instance-variable packing and backref framing. You will gain a better understanding of what makes Ruby fast or slow, how you can write more efficient Ruby code, and what implementations like JRuby are doing to make Ruby perform even better in the future.

“Optimizing Ruby”

Over the past 6 years, my focus on the JRuby project has usually been performance. I've worked on JRuby's bytecode compiler, optimizations to the runtime and core classes, improvements to memory use and allocation rates, and the removal of concurrency bottlenecks. Over that same time we've managed to increase JRuby's performance by orders of magnitude, in some cases approaching the speed of compiled Java.

This talk will go over the major performance leaps we've had during that time, starting with the disabling of ObjectSpace and proceeding through recent enhancements like instance-variable packing and backref framing. You will gain a better understanding of what makes Ruby fast or slow, how you can write more efficient Ruby code, and what implementations like JRuby are doing to make Ruby perform even better in the future.