#Benchmarking
6 notas · última:
Optimizing Javascript for fun and for profit
The page collects practical JavaScript optimization habits: avoid unnecessary work, string comparisons, shape changes, indirection, cache misses, large objects, and weak data structures, while insisting that benchmarking and profiling come first.
Price per 1M tokens is meaningless
The post argues that per-token pricing is a poor way to compare AI models because tokenizers differ and token efficiency varies. It favors cost per completed task as the more useful metric.
Planetary Scale Rust and Golang by embedding WebAssembly in Elixir
The post describes Assemblex, a toy Elixir platform for deploying and running WebAssembly binaries across nodes, with hot reloads, supervision, and benchmarked execution overhead.
Everything You Know About Latency Is Wrong
The post argues that latency must be understood as a full distribution, not as averages or a few percentiles, and that coordinated omission makes many benchmarks and monitors lie about response time.
Micro Benchmarking in Elixir using Benchfella
The post compares file-hashing approaches in Elixir with Benchfella, showing that larger chunk sizes perform better, while reading a whole file at once can be faster than a single oversized chunk.
Elixir vs Ruby Showdown - Phoenix vs Rails
The post compares Phoenix and Rails on the same routed, rendered request and reports much higher throughput for Phoenix, with lower CPU load and steadier latency in localhost and Heroku benchmarks.