Benchmark: Nested object proxy - getting

Script Preparation code:

var objs = [];
var proxProp = [];
var proxDyn = [];
var proxEval = [];
var proxFunc = [];
var count = 1000;
var a, b, c;

class ProxProp {
    constructor(data) { this.data = data }
    get a() { return this.data.a; }
    get b() { return this.data.b; }
    get c() { return this.data.c; }
    set a(v) { this.data.a = v; }
    set b(v) { this.data.b = v; }
    set c(v) { this.data.c = v; }
}

const ProxEval = eval(`class ProxEval {
    constructor(data) { this.data = data }
    get a() { return this.data.a; }
    get b() { return this.data.b; }
    get c() { return this.data.c; }
    set a(v) { this.data.a = v; }
    set b(v) { this.data.b = v; }
    set c(v) { this.data.c = v; }
};
ProxEval;
`);

const ProxFunc = new Function(`return class ProxFunc {
    constructor(data) { this.data = data }
    get a() { return this.data.a; }
    get b() { return this.data.b; }
    get c() { return this.data.c; }
    set a(v) { this.data.a = v; }
    set b(v) { this.data.b = v; }
    set c(v) { this.data.c = v; }
};
`);

class ProxDyn {
    constructor(data) { this.data = data }
}

["a", "b", "c"].forEach(n=> Object.defineProperty(ProxDyn.prototype, n, {
    get() { return this.data[n]; },
    set(v) { this.data[n] = v; },
}));

for (let i = 0; i < count; i++) {
    const obj = { a: Math.random(), b: new Date, c: Date.now().toString() };
    objs.push(obj);
    proxProp.push(new ProxProp(obj));
    proxDyn.push(new ProxDyn(obj));
    proxEval.push(new ProxEval(obj));
    proxFunc.push(new ProxFunc(obj));
}

Tests:

Get generated property with eval
for (let i = 0; i < proxEval.length; i++) { const obj = proxEval[i]; a = obj.a; b = obj.b; c = obj.c; }
Get property
for (let i = 0; i < proxProp.length; i++) { const obj = proxProp[i]; a = obj.a; b = obj.b; c = obj.c; }
Get dynamic property
for (let i = 0; i < proxDyn.length; i++) { const obj = proxDyn[i]; a = obj.a; b = obj.b; c = obj.c; }
Get field
for (let i = 0; i < objs.length; i++) { const obj = objs[i]; a = obj.a; b = obj.b; c = obj.c; }
Nothing
for (let i = 0; i < objs.length; i++) { const obj = objs[i]; }
Get generated property with Function ctor
for (let i = 0; i < proxFunc.length; i++) { const obj = proxFunc[i]; a = obj.a; b = obj.b; c = obj.c; }

Rendered benchmark preparation results:

Suite status: <idle, ready to run>

Previous results

Test case name	Result
Get generated property with eval
Get property
Get dynamic property
Get field
Nothing
Get generated property with Function ctor

Fastest: N/A

Slowest: N/A

Autogenerated LLM Summary (model llama3.2:3b, generated one month ago):

LLMs can make mistakes. Check important info.

Let's dive into the world of microbenchmarks and explore what's being tested in this specific benchmark.

**Benchmark Overview**

The provided JSON represents a JavaScript microbenchmarking test case on MeasureThat.net. The test measures the performance of different approaches to access properties of objects. There are six individual test cases, each with its own unique characteristics.

**Test Cases**

1. **"Nothing"**: This test case simply iterates over an empty array without accessing any properties.
2. **"Get field"**: This test case accesses a property directly on an object using dot notation (e.g., `obj.a`).
3. **"Get property"**: Similar to "Get field", but uses bracket notation (e.g., `obj['a']`) to access the property.
4. **"Get generated property with Function ctor"**: This test case creates a new function constructor and uses it to generate a new object with a property.
5. **"Get generated property with eval"**: Similar to the previous one, but uses the `eval` function to create a new object with a property.
6. **"Get dynamic property"**: This test case accesses a dynamically created property on an object using bracket notation.

**What's being tested?**

The primary focus of this benchmark is to measure the performance differences between various approaches to access properties on objects. The tests aim to evaluate:

* Direct property access (dot notation vs. bracket notation)
* Creation of new objects using function constructors and `eval`
* Accessing dynamically created properties

**Performance Insights**

From the provided benchmark results, we can observe the following:

* "Nothing" is the fastest test case, as expected, since it doesn't involve any property accesses.
* The order of performance for accessing properties directly on an object (dot notation vs. bracket notation) is:
	+ Dot notation: slower
	+ Bracket notation: faster
* Creating new objects using function constructors and `eval` results in slower performance compared to direct property access or bracket notation.
* Accessing dynamically created properties also shows a performance penalty, likely due to the overhead of string manipulation or parsing.

Overall, this benchmark provides valuable insights into the performance characteristics of different approaches to accessing properties on objects in JavaScript.

Latest run results:

Run details: (Test run date: one month ago)

User agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/18.0.1 Safari/605.1.15

Browser/OS: Safari 18 on Mac OS X 10.15.7

View result in a separate tab

Test name	Executions per second
Get generated property with eval	3741.8 Ops/sec
Get property	3745.4 Ops/sec
Get dynamic property	3721.2 Ops/sec
Get field	3787.8 Ops/sec
Nothing	12450.9 Ops/sec
Get generated property with Function ctor	3806.3 Ops/sec

Related benchmarks:

Nested object proxy - getting (version: 5)

Comparing performance of: Get generated property with eval vs Get property vs Get dynamic property vs Get field vs Nothing vs Get generated property with Function ctor

Created: 4 years ago by: Registered User

Jump to the latest result

Get generated property with eval

Get property

Get dynamic property

Get field

Nothing

Get generated property with Function ctor

Suite status: <idle, ready to run>

Experimental features:

Fastest: N/A

Slowest: N/A

Autogenerated LLM Summary (model llama3.2:3b, generated one month ago):