Benchmark: slice vs subarray vs set

Script Preparation code:

var data = [1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9];
var f32 = new Float32Array(data);

Tests:

subarray direct
var copy = f32.subarray(10, 20);
slice direct
var copy = f32.slice(10, 20);
subarray copy
var copy = new Float32Array(f32.subarray(10, 20));
slice copy
var copy = new Float32Array(f32.slice(10, 20));
subarray set
var copy = new Float32Array(10).set(f32.subarray(10, 20));
slice set
var copy = new Float32Array(10).set(f32.slice(10, 20));

Rendered benchmark preparation results:

Suite status: <idle, ready to run>

Previous results

Test case name	Result
subarray direct
slice direct
subarray copy
slice copy
subarray set
slice set

Fastest: N/A

Slowest: N/A

Latest run results:

Run details: (Test run date: 11 months ago)

User agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/137.0.0.0 Safari/537.36

Browser/OS: Chrome 137 on Linux

View result in a separate tab

Test name	Executions per second
subarray direct	19992284.0 Ops/sec
slice direct	15777429.0 Ops/sec
subarray copy	10085197.0 Ops/sec
slice copy	9348755.0 Ops/sec
subarray set	11258312.0 Ops/sec
slice set	9724500.0 Ops/sec

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

LLMs can make mistakes. Check important info.

**Benchmark Overview**

The provided benchmark compares the performance of three approaches to create a subset of an array in JavaScript: `subarray()`, `slice()`, and creating a new array using `set()`.

**Options Compared**

1. **`subarray()`**: This method creates a new array containing the elements from the original array, starting at the specified index (10) and ending at the specified end index (20).
2. **`slice()`**: Similar to `subarray()`, but returns a shallow copy of the specified range of elements in the original array.
3. **`subarray()` + `new Float32Array()`**: Creates a new `Float32Array` instance from the subset of the original array using `subarray()`.
4. **`slice()` + `new Float32Array()`**: Similar to the previous option, but uses `slice()` instead of `subarray()`.

**Pros and Cons**

* **`subarray()`**: Creates a new array, which can be inefficient if the resulting array is smaller than the original.
* **`slice()`**: Returns a shallow copy of the specified range of elements, which can lead to performance issues if the copied array needs to be modified later.
* **`subarray()` + `new Float32Array()`**: Creates a new `Float32Array` instance from the subset, which is more memory-efficient than creating a full array using `slice()`.
* **`slice()` + `new Float32Array()`**: Similar to the previous option, but uses `slice()` instead of `subarray()`.

**Library and Special Features**

None mentioned in the provided benchmark.

**Other Considerations**

* The benchmark only tests the performance of these four approaches, so other factors like memory allocation or garbage collection may not be considered.
* The test data is a large array with duplicate values, which can impact performance due to caching effects.

**Alternatives**

If you need to create an array subset in JavaScript, consider using `Array.prototype.slice()` or `Array.prototype.subarray()`, as they are more efficient than creating a new array using `new Float32Array()`.

For more complex scenarios, you may want to explore other libraries like Lodash's `_.slice()` or _.tail(), which provide optimized implementations for common array operations.

Related benchmarks:

slice vs subarray vs set (version: 0)

Comparing performance of: subarray direct vs slice direct vs subarray copy vs slice copy vs subarray set vs slice set

Created: 8 years ago by: Guest

Jump to the latest result

subarray direct

slice direct

subarray copy

slice copy

subarray set

slice set

Suite status: <idle, ready to run>

Fastest: N/A

Slowest: N/A

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