About smplmark
A home for benchmarks, built on a simple trust model.
smplmark is a general-purpose benchmark host. A publisher uploads benchmark data on a recurring basis; a visitor reviews benchmarks from the publishers they choose to trust. That is the whole model.
smplmark does not validate the truth of the data — only its shape. What the publisher publishes is what you get. Every benchmark declares the metrics it reports, and every number is fetchable as JSON or downloadable as CSV from the same public endpoint that powers the charts. If you trust the publisher, you can trust the numbers; if you don't, the raw data is right there to check.
Each benchmark is organized as targets (the things being measured), runs (a series of measurements from a target), and samples (the individual data points). Some values are stored exactly as the publisher sent them; others are derived — computed from the raw sample when it is read, so the stored data never commits to an interpretation and definitions can evolve without rewriting history.
Every published benchmark has an Overview, its Data, a Methodology writeup, and a Publisher — so you always know what you're looking at and who stands behind it. Browse what's published on the benchmarks page.