Graphics Processing Units (GPUs) are the accelerator of choice for performance-critical applications, yet optimizing for performance requires mastery of the complex interactions between its memory architecture and its execution model. Existing static analysis tools for GPU kernels either identify performance bugs without quantifying costs or cannot handle thread-divergent control flow, leading to significant over-approximations. We present the first static relational-cost analysis for GPU warp-level parallelism that can give exact bounds even in the presence of thread divergence. Our analysis is general and flexible, as it is parametric on the resource metric (uncoalesced accesses, bank conflicts) and on the cost relation (=, ≤, ≥). We establish a soundness theorem for our technique, provide mechanized proofs in Rocq and implement our theory in a tool called Pico . In a reproducibility experiment, Pico produced the tightest bounds in every input, outperforming the state-of-the-art tool RaCUDA in 10 kernels (1.7× better), while RaCUDA produced 4 incorrect bounds and crashed on 2 kernels. In an experiment to measure the accuracy of Pico , we studied the impact of threaddivergence in control-flow in a dataset of 226 kernels. We found that at least 75.3% of conditionals and 85.4% of loops can be captured exactly, without introducing approximation.

Abstract GPUs offer parallelism as a commodity, but they are difficult to program correctly. Static analyzers that guarantee data-race freedom (DRF) are essential to help programmers establish the correctness of their programs (kernels). However, existing approaches produce too many false alarms and struggle to handle larger programs. To address these limitations we formalize a novel compositional analysis for DRF, based on access memory protocols. These protocols are behavioral types that codify the way threads interact over shared memory. Our work includes fully mechanized proofs of our theoretical results, the first mechanized proofs in the field of DRF analysis for GPU kernels. Our theory is implemented in , a tool that outperforms the state-of-the-art. Notably, it can correctly verify at least $$1.42{\times }$$ 1.42 × more real-world kernels, and it exhibits a linear growth in 4 out of 5 experiments, while others grow exponentially in all 5 experiments.