PC Bottleneck Calculator

This bottleneck calculator estimates whether your CPU or your GPU limits your frame rate, and by how much. Pick your processor, graphics card, RAM kit and resolution, and you get the bottleneck direction and percentage, the estimated FPS you will actually see, and the single cheapest upgrade that fixes it. It works as a bottleneck detector for any desktop build from about 2016 onwards, as a CPU bottleneck checker, a GPU bottleneck checker and a quick RAM sanity check, at 1080p, 1440p and 4K. Nothing to install and nothing to sign up for.

How to use this bottleneck calculator

1. Pick your processor and graphics card. The lists cover every mainstream desktop CPU and GPU from the GTX 10 series and Ryzen 1000 to the current generation. Laptop parts are not modelled yet.
2. Set your RAM and resolution. RAM capacity, single versus dual channel and speed all move the CPU ceiling. Resolution moves the GPU ceiling; the same pair can be CPU-bound at 1080p and GPU-bound at 4K.
3. Open "More options" if you play one game a lot. Competitive titles lean on the processor, big open-world games lean on the card, and a few games cap themselves. Your monitor's refresh rate adds a third ceiling.
4. Read the verdict. The animation shows frames flowing through your parts and stalling at the limiter. The cards under it explain the result, price the idle capacity, and name the fix.

Your result explained: bottleneck %, direction and FPS

The percentage is the gap between the two frame ceilings, as a share of the higher one. A 20% CPU bottleneck means the processor can only prepare 80% of the frames the graphics card could draw. Under 10% is balanced and not worth acting on. 10 to 25% is mild: you would notice it mostly in 1% lows and in CPU-heavy scenes. Above 25% at the resolution you actually play at, the slower part is holding real money hostage and is worth upgrading.

Direction matters more than the number. GPU-bound is the normal, healthy state for a gaming PC: it means every graphics setting you raise costs frames and every one you lower gives them back. CPU-bound tends to feel worse than it measures, because the stalls show up as stutter and uneven frame pacing rather than a lower average.

The FPS figure is the lowest of your ceilings for the chosen game and resolution, with high settings and no upscaling. Treat it as a well-informed estimate, not a benchmark: the methodology page explains where the numbers come from and how far they can drift.

What is a PC bottleneck?

A bottleneck is the part of your PC that finishes its share of each frame last. The processor decides what happens in the game world and prepares draw calls; the graphics card turns those into pixels; the monitor shows them at its refresh rate. The frame rate you see is set by whichever of those three finishes last, and the other two spend the difference waiting.

That is why a faster part does not always mean more frames. Put a flagship graphics card behind a six-year-old processor at 1080p and the card idles for a third of every frame. Put a flagship processor in front of a budget card at 4K and the processor sits idle instead. Neither combination is broken; both are wasting money that could have gone into the part that actually limits.

Motherboards, power supplies and SSDs almost never limit average frame rate. A slow drive lengthens loading and can cause streaming hitches in open-world games, a weak power supply causes crashes rather than low frame rates, and the motherboard matters only through the memory it supports. That is why the calculator asks for four things and not twelve: everything else is a comfort or a reliability decision, not a frame-rate one.

CPU bottleneck vs GPU bottleneck: the quick way to tell

Open a performance overlay (Steam, RivaTuner, or the tool built into your graphics driver) and watch two numbers while you play. If GPU usage sits at 97 to 100%, the graphics card is the limiter and the processor is keeping up. If GPU usage keeps dropping into the 60s and 70s while one or two CPU threads stay pinned, the processor is the limiter.

The same pair flips depending on the scene. Crowded cities, big battles and simulation-heavy games push the processor; wide vistas, ray tracing and high resolutions push the card. That is why the calculator lets you choose the game: a build that is CPU-bound in a competitive shooter at 1080p can be firmly GPU-bound in an open-world game at 1440p. And 100% GPU usage is not a warning sign, it is the goal: it means the most expensive part in the box is fully busy.

Does RAM bottleneck a PC?

Yes, in three ways, and all three are modelled here. Too little capacity for the game forces it to page textures and assets through your SSD, which is where hitching comes from; 16 GB is the floor for current titles and 32 GB is comfortable. A single stick halves memory bandwidth and costs a modern processor a noticeable share of its gaming performance. Slow kits, or kits running at default speed because XMP or EXPO was never enabled, trim a few percent more.

RAM shows up in the result as a lower CPU ceiling and a separate memory section that names the problem and the kit that fixes it. X3D processors are far less sensitive to memory speed than other chips, which the model accounts for. When your memory is not the problem, the result says so plainly rather than selling you more of it.

Bottleneck at 1080p vs 1440p vs 4K

The processor's ceiling barely changes with resolution; the graphics card's ceiling falls sharply as the pixel count rises. So the same pair moves from CPU-bound at 1080p, through balanced at 1440p, to GPU-bound at 4K. High-refresh 1080p gaming is a processor problem; 4K gaming is a graphics card problem; 1440p is where balanced pairs live. Every bottleneck check on this site shows all three resolutions side by side, so you can see exactly where a pairing stops being processor-limited.

Bottleneck calculator by CPU

Choosing the processor first tells you how much graphics card the rest of the build can feed. A processor's frame ceiling is set almost entirely by single-thread speed and cache, and it barely moves when you change resolution, so the number you see at 1080p is roughly the number you are stuck with at 4K as well. That is what makes an old CPU behind a new card so expensive: the card cannot draw frames the processor never prepared.

Pick your exact chip in the calculator above and the result names the ceiling it sets in your game, how much of your graphics card is idling behind it, and the cheapest processor that clears the gap. Socket and memory type come with it, because a CPU upgrade that needs a new motherboard and new RAM is a different decision from a drop-in swap.

Bottleneck calculator by GPU

Choosing the graphics card first tells you which resolution the build is really for. Unlike a processor, a card's ceiling collapses as the pixel count rises, so the same GPU can be starved by your CPU at 1080p and be the limiter itself at 4K. VRAM is the other half of the answer: once a game needs more than the card has, frame times get spiky in a way an average frame rate never shows.

Pick your card in the calculator above and the result shows the ceiling it sets at your resolution, whether the processor is keeping up with it, and the cheapest card that would actually add frames rather than land on the same number.

Bottleneck detector, checker or calculator: is there a difference?

No. A bottleneck detector, a bottleneck checker, a bottleneck test and a bottleneck calculator all describe the same thing: a tool that estimates which of your parts limits your frame rate from published performance data. None of them measures your PC. The only way to measure a bottleneck is to run a game with an overlay and watch usage and frame times yourself, which takes about five minutes. This calculator is the estimate; your overlay is the confirmation.

How accurate is this bottleneck calculator?

Every part in the database carries a performance index derived from published benchmark aggregates: a 1080p gaming index for processors and a per-resolution index for graphics cards. Frame-rate estimates for each game come from how CPU-heavy or GPU-heavy that game is, scaled by those indices. In practice the estimates land within roughly 10 to 15% of review averages for mainstream settings, and further out for unusual settings, heavy mods, or games patched after the dataset was built.

What the model cannot see: your driver version, background load, RAM timings, overclocks, thermal throttling and 1% lows. It is honest about that in two ways. Each result shows a confidence label based on how well-documented the two parts are, and every page shows the data version it was built from. The full model, its sources and its limitations are on the methodology page, including how to sanity-check any calculator against a real benchmark.