How Accurate is Python\'s `time.sleep()` Function, and How Can I Improve Its Precision?

Measuring the Precision of Python's time.sleep() Function

Python's time.sleep() function is a useful tool for pausing the execution of a program. However, there are questions regarding its accuracy, especially when using floating-point numbers.

Precision of time.sleep()

The accuracy of time.sleep() depends on the underlying operating system (OS). Non-real-time OSs, such as Windows, have a minimum sleep interval of around 10-13 milliseconds.

What does this mean? If you attempt to sleep for 50 milliseconds (0.05 seconds), time.sleep() may only pause for about 50 milliseconds from the minimum interval (10-13 ms).

Looping for Accuracy

To improve accuracy, it's recommended to implement a loop around time.sleep(). If the function wakes up before the intended time, it will go back to sleep for the remaining duration.

Real-Time Kernels

If you need more precise sleep intervals, consider using a pseudo real-time kernel, such as Ubuntu's rt kernel (RT_PREEMPT patch set). This can reduce the minimum sleep interval to around 1 millisecond.

Further Considerations

Non-real-time Linux kernels have a minimum sleep interval closer to 1 millisecond, but it can vary in a non-deterministic manner.

Example with Looping for Accuracy

Here's an example with a loop to ensure accuracy:

import time

target_sleep_time = 0.05  # 50 milliseconds

start_time = time.time()
while True:
    elapsed_time = time.time() - start_time
    if elapsed_time >= target_sleep_time:
        break
    time.sleep(target_sleep_time - elapsed_time)

The above is the detailed content of How Accurate is Python\'s `time.sleep()` Function, and How Can I Improve Its Precision?. For more information, please follow other related articles on the PHP Chinese website!