LangBench - JFF benchmark of some programming languages

Just For Fun :)

Background

Everyone was constantly trying to convince me of certain stereotypes about the performance and speed of programming languages. At some point, I decided everyone was wrong and decided to test how REALLY true this was, as was commonly believed.

How does it work?

The `competitors`` directory contains simple API implementations based on the requirements (the requirements are described below).

I tried (and I encourage anyone wanted to contribuet) to implement them optimally and at the same time in a way that a typical developer would implement them on a typical project (without ASM, following common frameworks, without php://input, etc.).

Each competitor should also provide valid Dockerfile in root directory.

API requirements

API should contain route POST /api/calc with following input

{
    "request_id": "some_string",
    "items": [
        {
            "id": "unique_id",
            "value": 3.14,
            "tags": ["foo", "bar"]
        },
        ...
    ]
}

Fields:

• request_id: random string, in test generates andomly to prevent caching
• items.id: random string. Must be unique withing items list
• items.value: any float value. Can overflow the limits. In this case, consider it's equal 0
• items.tags: list of random strings

{
    "request_id": "some_string",
    "checksum": "M2ZiNWI2ZDY2YjMzODY1YTliZDE4YTg0MTgxYjNjY2RlM2I1ZjA1MWQxZTk5ODc3ZTZkN2ZkNjYxOWIyYWVmNg==",
    "stats": {
        "foo": {
            "count": 1,
            "sum": 3.14
        },
        "bar": {
            "count": 1,
            "sum": 3.14
        },
        ...
    }
}

Fields:

• request_id: the same string, provided in input
• checksum: sha256 hash of json with only request_id and stats
• stats.*: each tag mentioned in items.tags
• stats.*.count: how many times that particular tag was mentioned in input
• stats.*.sum: summary of values in items containing this particular tag

Contribution

If you found some bottleneck, don't agree with testing approach, etc - feel free to contribute :)

TODOs

• [] Workflow for automatic comparison
• [] Add JS-based API
• [] who-knows