angel-dart-archive/soniq_benchmark

Benchmarks of Angel compared to other frameworks.

benchmark

Auto-generated benchmarks, using soniq.

Please consider forking/cloning and running the benchmarks yourself. No, really.

To run:

dart bin/benchmark.dart

It will generate benchmark summary files.

Give it several minutes to run. Each benchmark runs three trials per framework, per test case, each lasting 1 minute.

The simulated load numbers are determined based on the number of cores your processor has.

In a rush? Run tests for less time by specifying a number of milliseconds:

# Run for 5 seconds only
dart bin/benchmark.dart 5000

Interpretation

As you can imagine, the numbers are dependent on a number of factors, which greatly vary across systems. Thus, latency numbers alone are only comparable to numbers from benchmarks run on the same machine.

However, the latency numbers should be proportional, so in general, the top framework will perform accordingly on any machine.

Test Cases

hello

Basic test, just tests a framework's ability to JSON encode data. Expected response: "Hello, world!"".

route_param

Tests a framework's ability to parse route parameters. The route should be declared with a pattern like /route_param/:name, and when called, return the following dynamic response:

{"name": "<the name parameter>"}