Popular CPU + GPU Bottleneck Pairings
Honest balance verdicts at 1440p for common combinations — or test your exact pair in the estimator.
Quick Summary · TL;DR
A bottleneck is not a fixed percentage — it depends on resolution and game type. These verdicts are for 1440p; use the full estimator for 1080p, 4K, upscaling and game-type tuning.
How to read these verdicts
Each pairing below carries a band rather than a percentage: balanced, GPU-limited, or one of two degrees of CPU limit. The band comes from comparing two tier indices at 1440p, and every pair page shows both indices so you can see the gap the verdict is based on.
A GPU-limited result is the normal and desirable outcome for gaming. It means the graphics card is the part working hardest, which is what you paid it to do. The verdicts worth acting on are the CPU-limited ones, and even those are conditional: a processor that limits a fast card in a simulation game at 1080p may be entirely adequate in the same build at 1440p.
What these verdicts cannot tell you
They cannot predict a frame rate, and they cannot tell you how a specific game behaves. Tier indices summarise broad gaming performance; individual titles disagree with them regularly, particularly large-cache processors in simulation-heavy games and older cards in titles that exceed their VRAM.
Getting a real answer instead
If you already own the parts, an overlay showing GPU utilisation and per-core CPU load will answer the question properly in about five minutes. The procedure is written out step by step in why a bottleneck percentage is not a real number, along with the full set of weights the estimator applies for resolution, game type and target frame rate.
Frequently asked questions
How do I know if my CPU and GPU are balanced?
Compare their relative tiers at your resolution. At 1440p a mismatch shows as one part far ahead of the other; pick a pairing rated balanced, or test your exact combo in the estimator.
Is a small bottleneck a problem?
No — a mild imbalance is normal and barely noticeable. Only a severe mismatch, where one part is far weaker, meaningfully costs you frames.
Test any exact combination in the Bottleneck Estimator.