TechPrufer
Get started
Shopify
Shopify logoShopify · Feature trail

Ruby Method Cache Tuning

1 post · Jan 2015

What this trail covers

Shopify has evolved its approach to optimizing Ruby's performance by tuning the global method cache size. This involved identifying performance bottlenecks using system-level profiling tools like `perf` and `ftrace`, and then dynamically configuring the cache size via the `RUBY_GLOBAL_METHOD_CACHE_SIZE` environment variable. This led to significant improvements in cache hit rates and a reduction in CPU cycles spent on method lookups, demonstrating the value of system-level performance tuning for large Ruby applications.

All Shopify trails
  1. Tuning Ruby's Global Method Cache - ShopifyThis post details the process of identifying and addressing a performance bottleneck in Ruby's global method cache. It describes the use of `perf` and `ftrace` (specifically uprobes) to analyze method cache hit rates in production. The author then introduced a runtime configuration option (`RUBY_GLOBAL_METHOD_CACHE_SIZE`) and demonstrated how increasing the cache size from the default 2048 entries to 64K or higher significantly improved hit rates and reduced CPU usage. This work was contributed upstream to Ruby 2.2.