Interpreting Latency Metrics: From Mouse Click to Final Photon

Latency is not a single number. It is a chain of delays, each caused by a different piece of hardware or software. For competitive gaming, understanding every link in this chain is the difference between reacting at 80ms total latency and reacting at 22ms. This guide breaks down every stage from the moment your finger presses a button to the moment a photon leaves your monitor.

1. Stage One: Input Device Polling (0–4ms)

The journey begins in your mouse or keyboard's micro-controller. The polling rate determines how frequently the device reports its state to the host OS. A 125Hz mouse reports every 8ms. A 1000Hz mouse reports every 1ms. An 8000Hz mouse (such as Razer's Viper V3 HyperSpeed) reports every 0.125ms.

However, polling rate alone does not tell the full story. Most mice apply an internal debounce algorithm to prevent false triggers. This debounce — typically 1–4ms — adds a fixed floor to your input latency regardless of polling rate. High-end mice like the Pulsar X2V2 have reduced this to as low as 0.2ms.

2. Stage Two: OS Interrupt and HID Stack (0.1–2ms)

When the mouse reports a button state, the OS processes a hardware interrupt. The USB HID stack parses the report descriptor, populates an event in the input queue, and notifies applications. On Windows 11, this path has been optimized significantly — with a high-priority real-time thread handling gaming peripherals — but background processes and driver quality still affect consistency.

Linux (with usbhid) and macOS (CoreHID) handle this differently. Linux with a PREEMPT_RT kernel patch achieves <1ms interrupt latency consistently, making it preferred for pro esports setups.

3. Stage Three: Game Engine Processing (1–8ms)

Once the input reaches the game engine, the CPU must process the logic frame: read the input, calculate character movement, simulate physics, and prepare a draw call list for the GPU. This is the most variable stage: a single-threaded bottleneck at this stage adds 5–15ms regardless of how fast your GPU or display is.

This is why raw GPU performance does not always improve your perceived responsiveness. A game that runs at 240fps but has a 12ms CPU game thread is much less responsive than a game running at 144fps with a 2ms game thread. More importantly, this stage is why you should always profile with GPU-Z or similar tools to identify whether your bottleneck is CPU or GPU.

4. Stage Four: GPU Render Pipeline (2–20ms)

The GPU receives the draw call list and begins rendering. This is where your TFLOPS score and VRAM bandwidth directly affect latency. Higher TFLOPS means the GPU can rasterize and shade more pixels per unit time. Higher VRAM bandwidth means textures transfer faster to the shader cores.

At 240fps, each frame has only 4.17ms to render. An RTX 4090 renders a typical esports title in ~1ms per frame. An aging GTX 1070 might take 8–12ms, forcing the game to drop frames or queue multiple frames in the driver — adding render latency even when the GPU "keeps up" on FPS.

5. Stage Five: Display Scan-Out and Panel Response (0.5–7ms)

Once the GPU finishes a frame and writes it to the framebuffer, the display's scan-out circuit begins reading and transmitting the pixel data line by line from top to bottom. At 240Hz, the entire screen refreshes in 4.17ms. At 144Hz, it's 6.94ms. At 60Hz, it's a painful 16.67ms — a full frame of delay before any pixel can change.

On top of scan-out, each pixel must physically change state. Grey-to-Grey (GtG) response time measures this. A TN panel achives ~1ms. A modern fast IPS panel achieves ~1ms with overdrive. An OLED panel achieves ~0.03ms — which is why OLED has become the professional standard for low-latency display in 2025–2026.

6. Measuring Your Own Total System Latency

The only accurate way to measure total system latency (from input to photon) is with a high-speed camera filming your input device and your monitor simultaneously. Tools like NVIDIA's LDAT (Latency Display Analysis Tool) automate this. For GPU render latency specifically, GearVerify's frame-time analysis provides the render portion of this chain directly from your browser.