DEV Community

Mahfuzur Rahman Saber
Mahfuzur Rahman Saber

Posted on

ChatGPT compared the memory usage with and without PHP generators for large datasets.

Visualizing memory usage with and without using generators can help understand the efficiency benefits. Below is a comparison of memory usage in two scenarios:

  1. Without using generators (loading all data into memory at once).
  2. Using generators (loading one item at a time).

Scenario 1: Without Generators

Let's say we have a simple function that returns an array of numbers from 0 to 999,999. This function loads all the data into memory at once.

<?php
function getNumbersArray() {
    $numbers = [];
    for ($i = 0; $i < 1000000; $i++) {
        $numbers[] = $i;
    }
    return $numbers;
}

$numbers = getNumbersArray();
foreach ($numbers as $number) {
    // Process each number
}
?>
Enter fullscreen mode Exit fullscreen mode

Memory Usage (Without Generators)

When the function getNumbersArray is called:

  • Initial memory: Low, only the function and environment setup are in memory.
  • During execution: As the loop runs, the memory usage increases linearly, holding all 1,000,000 numbers in an array.
  • Peak memory: Very high, holding all the data in memory simultaneously.
  • After execution: Memory remains high until the script ends or the array is explicitly unset.
|            Memory Usage Without Generators           |
|------------------------------------------------------|
| Start    | *                                         |
|          | **                                        |
|          | ***                                       |
|          | ****                                      |
|          | *****                                     |
|          | ******                                    |
|          | *******                                   |
| ...      | ******************************************|
| End      | ******************************************|
|------------------------------------------------------|
Enter fullscreen mode Exit fullscreen mode

Scenario 2: Using Generators

Now, we use a generator function to yield numbers one at a time.

<?php
function numberGenerator() {
    for ($i = 0; $i < 1000000; $i++) {
        yield $i;
    }
}

foreach (numberGenerator() as $number) {
    // Process each number
}
?>
Enter fullscreen mode Exit fullscreen mode

Memory Usage (Using Generators)

When the generator function numberGenerator is called:

  • Initial memory: Low, only the function and environment setup are in memory.
  • During execution: Memory usage remains low as only one number is held in memory at a time.
  • Peak memory: Low, only one item plus overhead for the generator.
  • After execution: Memory usage drops immediately after the iteration ends.
|            Memory Usage With Generators              |
|------------------------------------------------------|
| Start    | *                                         |
|          | *                                         |
|          | *                                         |
|          | *                                         |
|          | *                                         |
|          | *                                         |
|          | *                                         |
| ...      | *                                         |
| End      | *                                         |
|------------------------------------------------------|
Enter fullscreen mode Exit fullscreen mode

Summary

  • Without Generators: Memory usage increases with the size of the dataset and remains high throughout the script execution.
  • With Generators: Memory usage remains constant and low, regardless of the dataset size, because only one item is processed at a time.

Generators provide significant memory efficiency benefits, especially for large datasets, by yielding one item at a time and maintaining low memory usage throughout the script's execution.

Top comments (0)