By Walter Jobs, Technology writer and editor · Published 10 October 2026
Yes, but not in the way the question implies, and the exceptions are the interesting part. A higher frame rate shortens one link in a chain of about eight. Shorten that link far enough and the links either side of it start to dominate, which is why a machine running 400 fps can feel no more responsive than one running 200, and why a machine running 60 fps can feel worse than its frame time suggests.
This article does three things. It gives you the vocabulary the hardware vendors actually use, because “input lag” is one word for several different delays with different owners. It shows where frame rate enters that chain and where it does not. And it names three widely repeated claims about latency that the vendors’ own documentation contradicts.
The vocabulary, from the company that defined it
NVIDIA published a six-term breakdown specifically to retire the word everyone uses. Its own framing: “Let’s take a moment to dive one level deeper and define some terms that are more accurate than ‘Input lag’.” The definitions, verbatim:
| Term | NVIDIA’s definition | Who controls it |
|---|---|---|
| Peripheral latency | “The time it takes your input device to process your mechanical input and send those input events to the PC” | Your mouse or keyboard |
| Game latency | “The time it takes for the CPU to process input or changes to the world and submit a new frame to the GPU to be rendered” | The game engine and your processor |
| Render latency | “The time from when the frame gets in line to be rendered to when the GPU completely renders the frame” | Your graphics card, plus the queue |
| PC latency | “The time it takes a frame to travel through the PC. This includes both Game and Render Latency” | Everything inside the case |
| Display latency | “The time it takes for the display to present a new image after the GPU has finished rendering the frame” | Your monitor |
| System latency | “The time encompassing the whole end-to-end measurement – from the start of peripheral latency to the end of display latency” | The total you feel |
One small correction while the table is in front of you, because it appears in a great many articles on this subject. NVIDIA does not call the end-to-end figure “click-to-photon” on that page. It says click-to-display. The phrase motion-to-photon belongs to a different NVIDIA document, about its hardware measurement tool. If you see “NVIDIA’s click-to-photon latency” presented as vendor terminology, the writer did not read the vendor.
Where frame rate enters, and the arithmetic
Frame rate sets the duration of one stage. At N frames per second, each frame occupies 1000 divided by N milliseconds, so 60 fps is 16.7 ms per frame and 240 fps is 4.2 ms. No vendor publishes that formula, because it is division rather than a finding, but it is the whole reason frame rate matters to responsiveness at all.
What it does not do is scale the total. Two reasons. The first is that the other stages do not shrink when the frame does: your mouse still takes its own time, and NVIDIA’s own figures for peripheral hardware are wide, noting that mice and keyboards “can range anywhere from 1ms of latency to ~20ms of latency” and that a 125 Hz polling rate “adds up to 3ms of system latency on average compared to a 1000Hz polling rate”. A 20 ms mouse on a 400 fps machine is spending five times longer in the peripheral than in the frame.
The second reason is the queue, and it is the part that makes this subject counterintuitive.
The queue: why high frame rate and low latency can pull against each other
Between the driver and the GPU sits a buffer. NVIDIA describes its purpose without euphemism: the render queue “is designed to keep the GPU constantly fed by always having work buffered for the GPU to do. This helps to maximize FPS (throughput), but can introduce latency.”
Read that twice. The structure that keeps your frame rate high is the structure that delays your input. When the processor is producing frames faster than the graphics card can draw them, the queue fills, and every frame in it is a frame of your input sitting in a waiting room. NVIDIA’s description of the symptom: “The CPU is processing frames faster than the GPU can render them causing this backup, resulting in an increase of render latency.”
Microsoft explains the same mechanism from the operating system side, and because Microsoft is not selling graphics cards the phrasing is unusually clean: “An inherent side effect of buffering between CPU and GPU is that the user experiences increased latency. User input is picked up by the CPU during ‘frame N+1’ but is not rendered by the GPU until the following frame. There is a fundamental tension between latency reduction and submission/scheduling overhead.”
A fundamental tension, in Microsoft’s words. Not a bug anyone is going to patch out. Every latency feature shipped in the last six years is an attempt to manage it.
What the three vendors actually do about it
| Feature | What the vendor says it does | Requirements the vendor states |
|---|---|---|
| NVIDIA Reflex | “aligns game engine work to complete just-in-time for rendering, eliminating the GPU render queue and reducing CPU back pressure in GPU intensive scenes” | Needs the game to integrate it. NVIDIA: supports GPUs “all the way back to 2014’s GeForce GTX 900 Series products” |
| NVIDIA Low Latency Mode (driver) | “reduces latency by limiting the number of frames the CPU can prepare before the frames are processed by the GPU”; On “limiting queued frames to one”, Ultra “fully minimizing queued frames” | Driver-level, no game support needed. NVIDIA says Reflex takes priority if both are on |
| AMD Anti-Lag 2 | “reduces the system latency by applying frame alignment between the CPU and GPU jobs”, inserted “just before the user controls are sampled” | Needs game integration. AMD names “RDNA 1 architecture-based products, including the AMD Radeon RX 5000 Series and newer” |
| Intel XeLL | “render queue reduction, which is always available”; the added delay “is calculated so that latency is minimized, while the application’s performance is preserved” | Intel lists Arc A-Series or later, DirectX 12, and the DXGI flip model |
Intel’s one-line summary of the whole phenomenon is the best sentence any vendor has written on it: “Having a frame ready earlier adds to increased latency.” A frame finished before the system can use it is not a head start, it is a wait.
NVIDIA’s own characterisation of what Reflex achieves is worth keeping too, because it tells you where the floor is: “latency is only increased by the raw GPU render time.” Remove the queue and what remains is the actual work. That is the limit no setting gets under.
Three claims the documentation does not support
That you can add the stages together. You cannot, and NVIDIA says so explicitly: “Often game latency and render latency overlap, which means simply adding them together won’t produce a correct latency sum.” Frames are built in pieces, so stages run concurrently. Any stacked bar chart that sums latency stages into a total is overstating the total, and NVIDIA’s own definition of PC latency (“This includes both Game and Render Latency”) invites exactly that error.
That Hardware-Accelerated GPU Scheduling reduces input lag. Microsoft’s engineering blog, announcing the feature, says the opposite in effect: “The transition should be transparent, and users should not notice any significant changes. Although the new scheduler reduces the overhead of GPU scheduling, most applications have been designed to hide scheduling costs through buffering.” Microsoft claims lower scheduling overhead. It never claims lower input lag. Its only latency-adjacent commitment is negative: “We do not expect customers to experience performance regressions.”
That a higher mouse DPI lowers latency. NVIDIA addresses this directly: “Note: your mouse sensitivity will not significantly impact the latency of the mouse. Higher DPI does not mean lower latency.” Polling rate is the mouse setting that moves latency. Sensitivity is not.
The one stage nobody mentions: composition
After your graphics card finishes a frame, the operating system may still have work to do on it. NVIDIA: “Depending on your display mode (Fullscreen, Borderless, Windowed), the Desktop Windows Manager (DWM) in the OS has to submit some additional rendering work to composite the rest of the desktop for a particular frame. This can add latency. It’s recommended to always be in exclusive fullscreen mode to minimize compositing latency!”
Here the two companies disagree, and it is a live disagreement rather than one side being out of date. Microsoft tells developers that a modern windowed presentation can match or beat exclusive fullscreen: “Flip model presents go as far as making windowed mode effectively equivalent or better when compared to the classic ‘fullscreen exclusive’ mode.” Microsoft also states that with the right flag “you can get down to 1 frame of latency on recent versions of Windows”.
Our reading, stated as a reading rather than a fact: both are true of different games. A title using the modern flip presentation path loses little or nothing in borderless. A title that does not, loses something. You cannot tell from the outside which one you are running, which makes exclusive fullscreen the safer default and borderless a thing to test rather than assume. NVIDIA itself concedes movement here, noting that “in recent Windows updates, the latency of borderless windowed (windowed fullscreen) mode has slightly improved”.
Measuring it, rather than arguing about it
There are three tiers, and only one is free.
In-game metrics. Any game that integrates Reflex can expose game latency and render latency in its own display, and the NVIDIA app overlay can show PC and system latency figures. NVIDIA is honest about the limits of the overlay figure: render present latency “will be slightly smaller than render latency measured with the NVIDIA Reflex SDK”, and the in-game numbers are “not the full latency you feel, but can get you started on your path to latency optimization”. Free, approximate, useful for before-and-after comparisons on one machine.
Hardware in the monitor. The Reflex Latency Analyzer is built into specific 360 Hz G-SYNC displays. You plug your mouse into a dedicated passthrough port and it “works by detecting the clicks coming from your mouse and measuring the time it takes for a resulting display pixel change (i.e. gun fire) to happen on the screen”. It splits the result into mouse latency, PC plus display latency, and system latency. Accurate, and tied to a short list of monitors.
A sensor pointed at the screen. NVIDIA’s LDAT is “a discrete hardware analyzer that uses a luminance sensor to quickly and accurately measure the motion-to-photon (click-to-muzzle flash) latency”, and NVIDIA notes it “is cross platform and works with GPUs from all vendors”. This is how reviewers produce latency charts. It is not a consumer purchase, and NVIDIA’s own context for the cost of doing this before such tools existed was “over $7000 in specialized high-speed cameras and equipment”.
What to do with all this
In order of how much each is worth, for a person who wants a more responsive machine:
- Turn on Reflex, Anti-Lag 2 or XeLL if the game offers it. This attacks the queue, which is the largest controllable stage in a GPU-bound game. It is a checkbox and it is free.
- Check your mouse polling rate before buying anything. NVIDIA’s own figure puts up to 3 ms on the table between 125 Hz and 1000 Hz, which is more than the entire gap between 144 and 240 fps.
- Get the frame rate out of the basement. Going from 30 to 60 fps removes 16.7 ms of frame time. Nothing else on this list is that large. This is where our fps calculator and the frame rate drops article earn their place.
- Stop chasing frame rate once it is high. Past roughly 120 fps the per-frame saving is small and the other stages dominate. NVIDIA’s own extreme case makes the point: at over 800 fps in VALORANT on an RTX 5090 with its newest latency feature, it reports PC latency “under 3 ms”. The curve has flattened long before that.
- Decide the tearing question deliberately. Whether you cap your frame rate and how you configure VSync against a variable refresh display has a real latency cost either way. We cover the specifics, including a place where NVIDIA’s own documents disagree, in should you cap your fps and VSync, G-SYNC and FreeSync.
The honest summary: frame rate is one term in the equation, it matters most when it is low, and the single biggest win available to most people is a setting rather than a purchase.