infrastructure
The Kernel Test Ended in a Proof
RESOLVE perturbed execution, reduced concurrency and used formal analysis to validate AI-written GPU kernels.
Summary
RESOLVE perturbed execution, reduced concurrency and used formal analysis to validate AI-written GPU kernels.
Random tests can miss races, while numerical tolerances can conceal wrong GPU kernels. RESOLVE layers three checks: timing perturbations to expose nondeterminism, agent-written reduced-concurrency versions that remain bitwise identical in tests, and an F*/Pulse proof that the reduced candidate and reference compute the same real-number function. The researchers applied the pipeline across KernelBench and fused GEMMs written in CUTLASS, Triton and Gluon. They report four previously unreported issues, including two clear bugs, and show agent repairs with little performance loss. The important shift is operational: aggressive optimization becomes safer when the validation path is stronger than another pile of examples.
Why it matters
RESOLVE perturbed execution, reduced concurrency and used formal analysis to validate AI-written GPU kernels.
Limits and context
- RESOLVE layers three checks: timing perturbations to expose nondeterminism, agent-written reduced-concurrency versions that remain bitwise identical in tests, and an F*/Pulse proof that the reduced candidate and reference compute the same real-number function.
Key claims
RESOLVE perturbed execution, reduced concurrency and used formal analysis to validate AI-written GPU kernels.
Qualification: RESOLVE layers three checks: timing perturbations to expose nondeterminism, agent-written reduced-concurrency versions that remain bitwise identical in tests, and an F*/Pulse proof that the reduced candidate and reference compute the same real-number function.
Evidence: source-2026-10-06-007
Sources
- arXiv preprint 2610.05683arXiv · primary research
Corrections
No corrections have been recorded for this story.