wrk

Utilizes multiple threads and scalable event systems like epoll/kqueue to generate massive load efficiently, as shown in benchmarks with high requests per second (e.g., 748k RPS in the example).

Integrates LuaJIT for customizable HTTP request generation, response processing, and reporting, enabling complex test scenarios via scripts in the 'scripts/' directory.

Provides comprehensive statistics including latency distribution, requests per second, and transfer rates, aiding precise bottleneck identification as seen in the output with avg latency and stdev.

Designed for minimal overhead on a single multi-core CPU, maximizing load generation capabilities without heavy resource consumption.