Time in C++: std::chrono::high_resolution_clock: Myths and Realities -- Sandor Dargo

std::chrono::high_resolution_clock sounds like the obvious choice when you care about precision, but its name hides some important caveats. In this article, we’ll demystify what “high resolution” really means in <chrono>, why this clock is often just an alias, and when—if ever—it’s actually the right tool to use.

Time in C++: std::chrono::high_resolution_clock - Myths and Realities

by Sandor Dargo

From the article:

If there’s one clock in <chrono> that causes the most confusion, it’s std::chrono::high_resolution_clock. The name sounds too tempting — who wouldn’t want “the highest resolution”? But like many things in C++, the details matter.

In the earlier parts of this series, we looked at system_clock as the wall-clock time source, and at steady_clock as the reliable choice for measuring intervals. This time, we’ll tackle the so-called “high-resolution” clock, separate fact from myth, and see why it’s not always the right choice — even when you think you need precision.

What “high resolution” actually means

A clock’s resolution (or precision) is the granularity of its tick period — i.e., the smallest representable step in time for that clock’s time_point. In <chrono> it’s exposed via Clock::period, a std::ratio.

Notice an important difference. I didn’t mention accuracy, only precision. A clock might represent nanoseconds, but still be inaccurate due to hardware or OS scheduling. A higher resolution doesn’t necessarily mean better measurement. For timing, stability and monotonicity matter much more than how fine-grained the tick is.

View the full article