A combined ranking from Tom's Hardware, PCGH and ComputerBase. Each site tests different scenes, so their raw FPS numbers don't line up directly. The Performance Index puts every review on the same 0 to 100 scale so you can compare them. Tap any chip to compare against it.
Just want to know what to buy? → Complete builds at every budget
Same game, but re-measured over time and across sites, so a CPU's FPS drift between epochs and scenes — absolute numbers aren't comparable across sources. By source keeps you inside one comparable set. The Performance Index rescales every enabled dataset onto a shared scale using a two-way additive fit (per-dataset offset + per-CPU effect), so a CPU's score reflects its own speed, not which reviewer happened to test it. ·N next to a score is how many datasets cover that CPU. One sanity prior is then applied within each Intel microarchitecture: a higher-tier or higher-binned part (a 14900KS over a 14900K, the Core Ultra 285K over a 235) has strictly more cores, cache or clock, so it can't be slower in a CPU-bound game. Where a thinly-tested SKU's noisy score violates that, a coverage-weighted isotonic fit snaps the family back into order — well-tested parts barely move, thin ones fall into line. It's deliberately not applied across architectures or to AMD's ladder, where the spec sheet doesn't track gaming order (the 7800X3D beats the 7950X3D) — the lone exception being a few same-die, clock-only pairs (e.g. the 5800X3D over the 5700X3D) that are guaranteed by physics. Same-silicon siblings may legitimately tie.