V-Sync Lag vs. FPS Capping in Hardware Benchmarks

This article compares the input lag penalty of V-Sync against framerate capping within the context of computer hardware benchmarking. It explains the mechanics behind how V-Sync introduces significant latency via frame buffering, whereas framerate capping restricts frame output at the CPU level to maintain responsiveness. Furthermore, it details how both settings impact benchmark data and why hardware testers vastly prefer framerate capping over V-Sync when conducting controlled latency, thermal, and frame-pacing evaluations.

The Mechanics of V-Sync Input Lag

Vertical Synchronization (V-Sync) is designed to eliminate screen tearing by forcing the graphics card (GPU) to wait until the monitor finishes its current refresh cycle before displaying a new frame.

When the GPU renders frames faster than the monitor's refresh rate, V-Sync holds those extra frames in a buffer (the render queue). Because the frames wait in a queue before appearing on screen, the visual result of your mouse movement or keystroke is delayed. This backpressure can add anywhere from 15 to over 50 milliseconds of input lag, depending on the refresh rate.

In hardware benchmarking, enabling V-Sync invalidates both raw throughput tests and system latency measurements. It artificially limits the hardware to the monitor's refresh rate and inflates latency metrics with buffer delays that do not reflect the actual speed of the CPU or GPU.

How Framerate Capping Controls Latency

Framerate capping limits the maximum frames per second (FPS) a system can produce, either through an in-game setting or an external tool like RivaTuner Statistics Server (RTSS) or GPU driver controls.

Unlike V-Sync, a framerate cap does not rely on a display buffer to slow down the output. Instead, it tells the CPU to wait before gathering input and preparing the next frame. Because the frame is generated closer to the exact moment it is displayed, the render queue remains empty.

In-game framerate limiters typically offer the lowest input lag because they control the simulation loop directly at the engine level. External driver or software limiters may add up to one frame of latency compared to in-game caps, as they intercept the frame later in the pipeline, but they still induce drastically less input lag than V-Sync.

Impact on Hardware Benchmarking

When benchmarking computer hardware, the goal is to isolate and measure the exact performance of the components. V-Sync and framerate capping affect this data in vastly different ways.

1. Raw Throughput Testing

For standard FPS benchmarks, both V-Sync and framerate caps must be disabled. Benchmarking requires uncapped framerates to push the CPU and GPU to their absolute limits to discover the maximum average FPS and 1% low metrics.

2. System Latency Benchmarking

Modern hardware reviews often measure "click-to-photon" latency using specialized tools.

  • With V-Sync: Latency spikes dramatically because the measurement includes the time frames spend waiting in the V-Sync buffer. This hides the true responsiveness of the hardware.
  • With Framerate Capping: If aGPU hits 99% to 100% utilization during an uncapped benchmark, the render queue backs up, which naturally increases input lag. By applying a framerate cap just below the GPU's maximum output, benchmarkers can keep GPU utilization under 95%. This prevents a bottleneck, resulting in lower input lag than running the game completely uncapped.

3. Power, Thermal, and Frame-Pacing Benchmarks

Benchmarking isn't just about maximum speed; it is also about efficiency and consistency. When testing how much power a GPU draws or how hot a CPU gets under a standardized load (for example, locking a game to 60 or 120 FPS to compare two graphics cards equally), a framerate cap is the required method.

Using a framerate cap ensures the hardware is only doing the exact work required to hit that target, while maintaining flat, consistent frame times and minimal input lag. Using V-Sync for this purpose would introduce micro-stutters if the framerate ever dropped below the refresh rate, while simultaneously corrupting any latency data collected during the thermal test.

Summary of Differences

  • V-Sync: Synchronizes to the display, eliminates screen tearing, fills the render queue, and causes a massive input lag penalty. It is unusable for accurate hardware benchmarking.
  • Framerate Capping: Restricts the CPU/GPU workload directly, leaves screen tearing unmanaged, keeps the render queue empty, and reduces input lag when preventing a 100% GPU bottleneck. It is a vital tool for standardized efficiency and latency benchmarking.