Lisp-like DSL for Benchmarking
summary
MZBench provides a constrained Lisp-like benchmarking DSL embedded in Erlang that runs on localhost or a cloud cluster and ships on GitHub. It uses Erlang parse terms and a simple graph-based static analysis to detect loops and synchronization bugs, limits the language to about ten core operations to prevent runaway 100-node experiments, and adds a Python CLI and alternate front end so protocol workers can be written in Python. A prototype takes about a month and the system runs in production for two years.
AI-written summary · about
chapters (5)
related talks
46mMZBench - Distributed Load Testing the Erlang Way
31mMaking it Fast: The Power of Benchmarking
46mStop Guessing and Start Measuring - Benchmarking in Practice
02h 44mMeasure, don’t guess: benchmarking stories from the trenches
25mPrescriptively Benchmarking Kubernetes System and Application Using K-Bench
21mBenchmarking Like a Boss
06mIGNITE: How to benchmark … poorly
11mLightning Talk – Testing Bazel’s Remote Execution API