Benchmark: Instantiation of ArrayBuffer vs Normal Part 3

Script Preparation code:

var last = 100000;

Tests:

Uint32
for (let i; i < last; i++) { const x = new ArrayBuffer(4); const y = new Uint32Array(x); y[0] = 255; }
Normal
for (let i; i < last; i++) { var y = 255; }
Uint32Array
const x = new ArrayBuffer(4 * last); const y = new Uint32Array(x); last = y.length; for (let i; i < last; i++) { y[0] = 255; }
Uint8Array
const x = new ArrayBuffer(last); const y = new Uint8Array(x); for (let i; i < last; i++) { y[0] = 255; }
Uint8Array (maybe faster due to garbage collector)
for (let i; i < last; i++) { const x = new ArrayBuffer(1); const y = new Uint8Array(x); y[0] = 255; }
Number
for (let i; i < last; i++) { var y = Number(255); }
Instantiation Number
for (let i; i < last; i++) { var y = new Number(255); }

Rendered benchmark preparation results:

Suite status: <idle, ready to run>

Previous results

Test case name	Result
Uint32
Normal
Uint32Array
Uint8Array
Uint8Array (maybe faster due to garbage collector)
Number
Instantiation Number

Fastest: N/A

Slowest: N/A

Autogenerated LLM Summary (model llama3.2:3b, generated 25 days ago):

LLMs can make mistakes. Check important info.

Let's dive into the benchmark and explain what's being tested.

**Benchmark Description**

The benchmark is testing different ways to create and use arrays for storing numbers in JavaScript. Specifically, it compares:

1. Instantiation of `ArrayBuffer` (with `Uint32Array`, `Uint8Array`, or `Uint8Array` with a small size)
2. Normal creation of variables (`var y = 255;`)
3. Instantiation of the built-in `Number` type
4. Instantiation of wrapper objects around `Number` (not used in this benchmark)

**Options Compared**

The options being compared are:

* `Uint8Array`: an array buffer with a single byte, potentially benefiting from garbage collection.
* `Uint32Array`, `Uint8Array` with larger size: an array buffer for storing 32-bit or 8-bit numbers directly.
* Normal variables (`var y = 255;`): creating a variable without initializing it.

**Pros and Cons**

Here's a brief overview of the pros and cons of each option:

* **Uint8Array**: Pros:
	+ May be faster due to garbage collection.
	+ More memory-efficient compared to `Uint32Array`.
	+ Can store smaller values (1 byte).
*   Cons: Limited precision, no built-in support for larger numbers.
*   **Uint32Array**, `Uint8Array` with larger size: Pros:
	+ Offers higher precision and support for larger numbers.
	+ Built-in array buffer type for direct access to memory.
*   Cons: May not be as fast due to overhead of initializing an array buffer.

**Test Case Analysis**

The benchmark tests the execution speed of each option. The test cases are designed to create arrays or variables with a large number of elements (100,000), and then iterate over them to set their values.

For `Uint8Array` (with small size), Chrome 104 is slightly faster, likely due to garbage collection benefits.
For `Number`, Chrome 104 is slower than the other options, likely because it involves more overhead when creating a new variable.

**Library and Special JS Features**

* No external libraries are used in this benchmark.
* There are no special JavaScript features or syntax that require explanation, as they don't impact the performance of these specific test cases.

**Alternatives**

If you want to explore alternatives, here are some options:

1. **Native arrays**: Instead of using `Uint8Array`, `Uint32Array`, or `Number`, consider native array types like `Float64Array` or `BigInt64Array`. These may offer better performance and precision.
2. **Typed arrays with typed arrays API**: The Typed Array API provides a more modern way to work with arrays for specific data types (e.g., `Float64Array`, `Int8Array`). This might be an alternative, but it's not directly comparable to the options in this benchmark.
3. **Custom solutions**: Depending on your use case, you might need to create custom solutions that utilize specialized data structures or algorithms.

Keep in mind that these alternatives might require additional setup and understanding of their behavior.

Latest run results:

Run details: (Test run date: 2 years ago)

User agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/104.0.0.0 Safari/537.36

Browser/OS: Chrome 104 on Windows

View result in a separate tab

Test name	Executions per second
Uint32	7042838.5 Ops/sec
Normal	7542070.0 Ops/sec
Uint32Array	30673.1 Ops/sec
Uint8Array	59587.3 Ops/sec
Uint8Array (maybe faster due to garbage collector)	7994510.5 Ops/sec
Number	7608067.0 Ops/sec
Instantiation Number	7385961.5 Ops/sec

Related benchmarks:

Instantiation of ArrayBuffer vs Normal Part 3 (version: 0)

normal var is faster but using a Uint__Array is more memory efficient

Comparing performance of: Uint32 vs Normal vs Uint32Array vs Uint8Array vs Uint8Array (maybe faster due to garbage collector) vs Number vs Instantiation Number

Created: 2 years ago by: Guest

Jump to the latest result

Uint32

Normal

Uint32Array

Uint8Array

Uint8Array (maybe faster due to garbage collector)

Number

Instantiation Number

Suite status: <idle, ready to run>

Experimental features:

Fastest: N/A

Slowest: N/A

Autogenerated LLM Summary (model llama3.2:3b, generated 25 days ago):