Mercurial

comparison benchmark/bun-http-framework-benchmark/README.md @ 186:8cf4ec5e2191 hg-web

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Fixed merge conflict.
author MrJuneJune <me@mrjunejune.com>
date Fri, 23 Jan 2026 22:38:59 -0800
parents a8976a008a9d
children
comparison
equal deleted inserted replaced
176:fed99fc04e12 186:8cf4ec5e2191
1 # Bun HTTP Framework Benchmark
2
3 Compare throughput benchmarks from various JavaScript HTTP framework
4
5 Test method: Average throughput
6
7 1. Ping
8 - Request to [GET] `/`
9 - Return `hi`
10 - Headers must contains text `Content-Type: text/plain`, additional context is acceptable eg. `Content-Type: text/plain; charset=utf-8`
11 2. Query
12 - Request to [GET] `/id/:id`
13 - Extract path parameter, query string and setting headers.
14 - For this benchmark, the request URL will be send as: `/id/1?name=bun`
15 - Headers must contains `x-powered-by` to `benchmark`
16 - Expected response: **"1 bun"** (`${id} ${query}`)
17 - You **MUST NOT use hardcode string or index** to extract querystring.
18 - In a real-world situation, there's no enforcement that the request will follow the specification, using hardcode index to extract `name=bun` querystring will be prone to error.
19 - To test if it pass the requirement, the implementation should be able to extract querystring **dynamically** (please treat the value of 'name=bun' can be any value beside 'bun', for example 'alice', 'hina'), which means that the same code should be able to extract querystring, for example:
20 - `/id/1?name=bun&id=1` -> should return `1 bun` not `1 bun&id=1`
21 - `/id/1?id=1` -> should return `1 `
22 - Query beside `name` maybe not need to be extracted and is optional
23 - Headers must contains text `Content-Type: text/plain`, additional context is acceptable eg. `Content-Type: text/plain; charset=utf-8`
24 3. Body
25 - [POST] `/json`
26 - Mirror body to response
27 - Server **MUST parse body to JSON and serialize back to string**
28 - For the benchmark, the request body will be sent as: `{ "hello": "world" }`
29 - Expected response: `{ "hello": "world" }`
30 - Headers must contains text `Content-Type: application/json`, additional context is acceptable eg. `Content-Type: application/json; charset=utf-8`.
31
32 ## requirement
33
34 - The framework must at-least has latest published in less than 9 month otherwise will be classified as unmaintained and removed unless is an industry standard (Express).
35
36 # Prerequistes
37
38 - [bombardier](https://github.com/codesenberg/bombardier)
39 - Nodejs
40 - Deno
41 - Bun
42
43 # Run Test
44
45 ```typescript
46 bun benchmark
47 ```
48
49 Dump result will be available at `results/[benchmark-name].txt`
50
51 ## Benchmark Condition
52
53 This benchmark is tested under the following condition:
54
55 - Intel I7-13700K, DDR5 32GB 5600MHz
56 - Arch Linux 6.10.9-arch1-2 (btw I use Arch)
57 - Bun 1.2.11
58 - Node 22.15.0
59 - Deno 2.2.0
60
61 Tested on 6 May 2025 20:55 (GMT+7)
62
63 ## Results
64
65 These results are measured in req/s:
66
67 | Framework | Runtime | Average | Ping | Query | Body |
68 | ---------------- | ------- | ------- | ---------- | ---------- | ---------- |
69 | uws | node | 526,295.397 | 535,079.45 | 509,759.27 | 534,047.47 |
70 | ultimate-express | node | 416,645.31 | 526,907.98 | 508,436.34 | 214,591.61 |
71 | elysia | bun | 397,259.163 | 533,560.05 | 300,716.41 | 357,501.03 |
72 | bun | bun | 366,716.193 | 386,961.93 | 317,958.96 | 395,227.69 |
73 | hyper-express | node | 315,844.147 | 370,809.65 | 299,522.59 | 277,200.2 |
74 | bun-web-standard | bun | 303,568.81 | 338,462.84 | 255,297.65 | 316,945.94 |
75 | nhttp | bun | 283,734.777 | 373,377.37 | 244,767.65 | 233,059.31 | | wobe | bun | 282,858.987 | 282,509.03 | 250,687.47 | 315,380.46 |
76 | hono | bun | 253,646.173 | 294,945.14 | 236,066.25 | 229,927.13 |
77 | deno | deno | 247,770.227 | 269,520.16 | 233,460.27 | 240,330.25 |
78 | nbit | bun | 229,226.153 | 252,902.81 | 196,843.1 | 237,932.55 |
79 | deno-web-standard | deno | 204,908.633 | 236,235.96 | 183,458.05 | 195,031.89 |
80 | hono | deno | 201,548.523 | 250,592.98 | 195,744.04 | 158,308.55 |
81 | fastify | node | 142,695.487 | 155,142.07 | 148,078.23 | 124,866.16 |
82 | hono | node | 129,234.593 | 144,842.92 | 134,412.44 | 108,448.42 |
83 | express | bun | 126,674.41 | 145,427.54 | 134,805.54 | 99,790.15 |
84 | oak | bun | 120,481.7 | 117,743.45 | 109,295.56 | 134,406.09 |
85 | h3 | bun | 109,739.16 | 132,720.75 | 104,854.13 | 91,642.6 |
86 | h3 | node | 101,237.803 | 136,023.12 | 111,361.83 | 56,328.46 |
87 | oak | deno | 100,205.983 | 105,824.67 | 93,335 | 101,458.28 | | koa | node | 96,456.887 | 103,217.62 | 97,582.12 | 88,570.92 |
88 | acorn | deno | 64,951.337 | 95,734.91 | 71,099.63 | 28,019.47 |
89 | express | node | 25,079.01 | 26,286.22 | 24,796.49 | 24,154.32 |
90
91 See more detail in [results](https://github.com/SaltyAom/bun-http-framework-benchmark/tree/main/results)
92
93 ## Notice
94
95 I highly recommended testing this benchmark on your machine yourself as performance in likely to vary between machine.
96
97 If you are unable to run Deno, please run each Deno app individually first until the Deno finish installing the package, then proceed to run benchmark using `bench.sh` or `npm run benchmark`