Benchmark: math.sin vs fast sin vs cached math.sin

Script Preparation code:

const PI = Math.PI;
const TAU = PI * 2;
const B = 4 / PI;
const C = -4 / TAU;
var cachedSin = Math.sin;
var sin = x =>  {
    return (B * x + C * x * ((x < 0) ? -x : x));
};

Tests:

Math.sin
let k = []; for (let i = 0; i < 100; ++i) k.push(Math.sin(i));
fast sin
let k = []; for (let i = 0; i < 100; ++i) k.push(sin(i));
cachedSin
let k = []; for (let i = 0; i < 100; ++i) k.push(cachedSin(i));

Rendered benchmark preparation results:

Suite status: <idle, ready to run>

Previous results

Test case name	Result
Math.sin
fast sin
cachedSin

Fastest: N/A

Slowest: N/A

Autogenerated LLM Summary (model llama3.2:3b, generated 27 days ago):

LLMs can make mistakes. Check important info.

The provided benchmark compares the performance of three different approaches to calculate the sine function:

1. **Native `Math.sin()`**: This is the built-in JavaScript function for calculating the sine of an angle, implemented in C++ by Mozilla.
2. **Custom "fast sin" function**: This is a custom implementation of the sine function, which uses a clever mathematical trick to reduce the number of multiplications required. The formula used is `B * x + C * x * ((x < 0) ? -x : x)`, where `B` and `C` are constants defined in the script preparation code.
3. **Cached `Math.sin()`**: This approach caches the result of a previous call to `Math.sin()`. In this benchmark, the variable `cachedSin` is assigned the value of `Math.sin()`, but with the actual calculation done only once during script preparation.

**Options Compared:**

* Native `Math.sin()` vs Custom "fast sin" function
* Native `Math.sin()` vs Cached `Math.sin()`

**Pros and Cons:**

1. **Native `Math.sin()`**:
	* Pros:
		+ Widely supported by most browsers.
		+ Optimized for performance by Mozilla.
	* Cons:
		+ May not be as efficient due to overhead of browser implementation.
2. **Custom "fast sin" function**:
	* Pros:
		+ Potentially faster than native `Math.sin()` due to reduced number of multiplications.
	* Cons:
		+ Requires custom implementation, which may add complexity.
		+ May not be supported by all browsers or JavaScript engines.
3. **Cached `Math.sin()`**:
	* Pros:
		+ Can potentially be faster than native `Math.sin()` due to caching effect.
	* Cons:
		+ Requires an extra variable to store the cached result, which can lead to memory usage.

**Other Considerations:**

* The custom "fast sin" function uses a mathematical trick that is specific to floating-point arithmetic. This trick may not work correctly for integer arithmetic or other data types.
* The benchmark does not consider other factors like input range, precision, or rounding mode when comparing the performance of these approaches.

**Libraries and Special JS Features:**

* The custom "fast sin" function uses a library-like approach to implement its formula. However, it is not a traditional library, as it is embedded directly in the script.
* There are no special JavaScript features used beyond the `let`, `const`, and arrow functions syntax.

**Alternatives:**

* Other optimization techniques for calculating sine, such as using polynomial approximations or specialized libraries like FFTW.
* Using SIMD instructions to perform calculations on multiple values simultaneously, if supported by the browser or JavaScript engine.
* Compiling custom JavaScript code to machine code using tools like WebAssembly, which can potentially lead to significant performance improvements.

Latest run results:

Run details: (Test run date: 6 months ago)

User agent: Mozilla/5.0 (X11; Linux x86_64; rv:126.0) Gecko/20100101 Firefox/126.0

Browser/OS: Firefox 126 on Linux

View result in a separate tab

Test name	Executions per second
Math.sin	184695.0 Ops/sec
fast sin	1239976.1 Ops/sec
cachedSin	234276.3 Ops/sec

Related benchmarks:

math.sin vs fast sin vs cached math.sin (version: 0)

Comparing performance of: Math.sin vs fast sin vs cachedSin

Created: 2 years ago by: Guest

Jump to the latest result

Math.sin

fast sin

cachedSin

Suite status: <idle, ready to run>

Experimental features:

Fastest: N/A

Slowest: N/A

Autogenerated LLM Summary (model llama3.2:3b, generated 27 days ago):