← back to paper
arxiv: 2605.03566 · 2 revisions
Lifting to tensors when compiling scientific computing workloads for AI Engines