Sums of
three cubes.
Which integers k are a sum of three integer cubes? Numbers like 33 and 42 hid for decades behind enormous solutions. Grinders grow the search box one radius at a time, and the server confirms every triple with exact big-integer arithmetic before it counts.
Three-cubes workbench
CPU-only: each lease fixes one target k and one search radius. The worker sweeps x and y over the box, solves for the cube root of the remaining offset, and returns the first exact integer triple it finds.
Idle. No box is being swept.
A single cube-root rounding error would produce a false solution, so the server recomputes x³ + y³ + z³ − k in exact arithmetic and also rejects any target the residue class k ≡ ±4 (mod 9) proves impossible.
k ∈ [1, 99]
Residues first
Cubes are 0, 1 or 8 modulo 9, so no k ≡ ±4 (mod 9) can ever be written as a sum of three cubes. Those targets are excluded up front; every remaining k is searched inside an expanding symmetric box.
Exact, not floating-point
A candidate is only accepted when x³ + y³ + z³ equals k exactly in big-integer arithmetic. The coordinator recomputes the identity itself, so rounding artefacts on the worker's side can never be recorded as a discovery.