Radar ·
Formalizing PTX in Lean to prove GPU kernel correctness
A Koska team piece on formalizing NVIDIA's PTX instruction set in the Lean prover to formally verify GPU kernels, including code written by agents.
Published on November 3, 2026 in Radar, this item is a Koska team piece on formalizing NVIDIA's PTX instruction set in the Lean prover. The stated goal is to formally verify low-level GPU kernels with machine-checked correctness. According to the author, the aim is to have confidence in kernels written by agents without having to read them line by line.
The text presents this approach as useful for verifying generated code, but the available summary gives no further details on method, results or tools. To consult the original, search for the exact title and the Koska team in public publication channels, check the date and read the full version before drawing technical conclusions.
Original source: Formalizando PTX em Lean para provar a correção de kernels de GPU