Skip to main content
CorgibanAlpha

Benchmark Suite

Build a repeatable solver test set, review saved results, and check browser save health without burying the main benchmarking tools.

Suite Builder

Choose the algorithms, levels, warm-up runs, and run limits before you start the benchmark.

Algorithms
iBreadth-first search checks the simplest push plans first. It is steady and reliable, but it can get slow on larger puzzles.
iA-star search uses a score to guess which pushes are most promising. It usually reaches good answers faster than breadth-first search.
iIDA-star search repeats a depth-first search with tighter score limits. It saves memory, but it may repeat work to find a solution.
iGreedy search always follows the move that looks best right now. It can be fast, but it is less careful and may miss better paths.
iTunnel Macro search groups forced tunnel moves together. That cuts down busywork in narrow spaces and can speed up puzzle solving.
iPI-Corral search looks for boxed-in areas that strongly limit what matters next. It can prune hard puzzles well, but the extra checks cost work.
Levels
These are the runs that get saved in your benchmark history for each level and algorithm.These are the runs that get saved in your benchmark history for each level and algorithm.
These practice runs happen before the saved runs. They help settle performance and are not saved.These practice runs happen before the saved runs. They help settle performance and are not saved.
This is the time limit for each run. If a solve takes longer, it times out.This is the time limit for each run. If a solve takes longer, it times out.
This is the search limit for each run. Use it when you want a hard cap on how much work the solver can do.This is the search limit for each run. Use it when you want a hard cap on how much work the solver can do.

Results

Saved benchmark runs (0). Sort the table, then reopen a puzzle in Play or Lab when a result needs a closer look.

No saved benchmark runs yet.

Analytics & Comparison

Compare saved suite runs when they were measured under the same settings and browser conditions.

Run at least one suite to see summary stats and comparisons.

Diagnostics

Track run progress and check whether this browser can save your benchmark history reliably.

Execution status
Idle
Execution progress
0/0
Latest result ID
None
Browser storage permissionThis shows whether the browser agreed to keep saved benchmark data around instead of treating it as easy to remove later.
pending
Save reliabilityThis shows how safe your saved benchmark history is right now. Memory fallback means it only lasts until this page reloads.
pending

Import / Export

History files contain past benchmark results. Level packs contain puzzle selections you can reopen in Play, Lab, and Bench.