Understanding FPS and Frame Time: Complete Guide to Game Performance

FPS is one of the most commonly used measurements when discussing game performance. Players often use FPS to describe how smoothly a game is running, while developers use it as one of several measurements when profiling and optimizing a game.

However, FPS alone does not tell the entire story. Frame time provides another way of looking at the same rendering process by measuring how many milliseconds are required to produce each frame.

Understanding the relationship between FPS and frame time makes it easier to reason about performance budgets, identify bottlenecks, and determine how much time a game has available to perform its CPU and GPU work.

FPS stands for frames per second. It represents the number of frames a game renders or displays during one second of gameplay.

For example, a game running at 60 FPS produces approximately 60 frames every second. A game running at 120 FPS produces approximately 120 frames during the same amount of time.

FPS is commonly used as a high-level measurement of game performance because it provides an easy way to compare different performance targets.

Frame time is the amount of time required to produce a single frame. It is normally measured in milliseconds, abbreviated as ms.

A game running at 60 FPS has approximately 16.67 milliseconds available to produce each frame. At 120 FPS, that budget falls to approximately 8.33 milliseconds.

This makes frame time particularly useful for game developers. Instead of only asking how many frames are being produced each second, developers can look at how much time is available to perform the work required for each frame.

Frame time is also useful when investigating stuttering. A game may have a high average FPS while still experiencing occasional frame-time spikes that are noticeable during gameplay.

FPS and frame time are two different ways of describing the same rendering process. FPS tells you how many frames are produced each second, while frame time tells you how long it takes to produce one frame.

The two values are inversely related. When FPS increases, the amount of time available for each frame decreases. When FPS decreases, the frame-time budget becomes larger.

Because one second contains 1000 milliseconds, FPS and frame time can be converted using a simple mathematical relationship.

For example, if a game is running at 60 FPS:

This means a game targeting 60 FPS has approximately 16.67 milliseconds to complete the work required for each frame.

The reverse conversion can be used when you know the frame time and want to determine the corresponding FPS.

For example, if a frame takes 8.33 milliseconds to produce:

This relationship is particularly useful during game performance analysis because it lets developers translate between the FPS target players see and the frame-time budget developers work with while profiling.

Different games target different frame rates depending on their genre, platform, hardware, and desired player experience. Each FPS target has a corresponding frame-time budget.

As the target FPS increases, the available time for each frame decreases. Maintaining a higher frame rate therefore requires the game to complete its work within a smaller time window.

A frame budget is the amount of time available to produce one frame before the next frame needs to be ready.

The higher your target FPS, the smaller your frame budget becomes. At 60 FPS, a game has approximately 16.67 milliseconds to produce each frame. At 120 FPS, that budget drops to approximately 8.33 milliseconds.

This is why increasing the target frame rate can become increasingly demanding. Moving from 30 FPS to 60 FPS requires reducing frame time from approximately 33.33 ms to 16.67 ms. Moving from 120 FPS to 240 FPS requires reducing it again from approximately 8.33 ms to just 4.17 ms.

Thinking in terms of frame budgets gives developers a more practical way to evaluate performance than looking only at the FPS number.

Frame time provides a practical way to understand how much work a game can perform during each frame. Rendering, gameplay simulation, physics, animation, AI, input processing, and other systems all contribute to the work that needs to be completed.

If the work required to produce a frame exceeds the available frame budget, the game cannot consistently maintain its target FPS.

For example, suppose a game targets 60 FPS. Its frame budget is approximately 16.67 milliseconds. If the work required to produce a frame takes 20 milliseconds, the game cannot consistently produce 60 frames every second.

Looking at frame time therefore gives developers a direct way to reason about whether a performance target is achievable.

When profiling a game, developers commonly investigate whether the CPU or GPU is responsible for the frame-time bottleneck.

A CPU-bound game is limited primarily by the amount of work being performed on the CPU. This can include gameplay systems, physics, artificial intelligence, animation, scripting, and other simulation tasks.

A GPU-bound game is limited primarily by rendering work. Geometry, materials, lighting, shadows, post-processing, visual effects, and rendering resolution can all contribute to GPU frame time.

Average FPS is useful for getting a general idea of game performance, but it does not tell the entire story. Two games can have the same average FPS while providing very different experiences.

A game can report a high average FPS while occasionally taking significantly longer to produce individual frames. These sudden increases in frame time can appear as visible stutters or hitches during gameplay.

For example, imagine a game targeting 60 FPS. Ideally, each frame should take approximately 16.67 milliseconds. If most frames take around that amount of time but one frame suddenly takes 50 milliseconds, the player may notice a brief hitch.

This is why developers often examine frame-time graphs when investigating stuttering rather than relying only on the average FPS displayed by an FPS counter.

FPS and frame time are useful measurements when profiling games made with both Unity and Unreal Engine. While the specific profiling workflows differ between the engines, the underlying relationship between FPS and frame time remains the same.

The same frame-budget principle applies regardless of the engine being used. A 60 FPS target provides approximately 16.67 milliseconds per frame, while a 120 FPS target provides approximately 8.33 milliseconds.

When optimizing a Unity or Unreal Engine project, avoid making changes based only on assumptions. Profile the game, identify the largest contributors to frame time, make a targeted optimization, and then profile again to measure the result.

FPS is easy to understand at a basic level, but there are several common mistakes that can lead to incorrect conclusions when analyzing game performance.

FPS is a useful performance metric, but it should not be treated as the only measurement that matters.

For example, two games could both report an average of 60 FPS while one maintains a relatively stable frame time and the other experiences frequent frame-time spikes. The second game may feel noticeably less smooth despite having the same average FPS.

Frame time provides additional information by showing how consistently the game is producing frames and how much of the available performance budget is being used.

FPS and frame time are two different ways of describing the same rendering process. FPS tells you how many frames are produced per second, while frame time tells you how many milliseconds are available to produce each frame.

Understanding both measurements makes it easier to reason about performance targets, frame budgets, and the amount of work a game can perform before the next frame needs to be produced.

For developers, frame time is particularly useful during profiling because CPU and GPU workloads can be measured directly in milliseconds and compared against the available frame budget.

Remember that a high average FPS does not automatically guarantee smooth gameplay. Consistent frame times and the absence of large frame-time spikes are also important for a stable experience.