PyTorch prototypes AOTI registration for DSL kernels via TVM-FFI
A new torch.library.register_aot_impl path aims to compile custom DSL kernels into PyTorch’s ahead-of-time Inductor stack.
By tensorA new torch.library.register_aot_impl path aims to compile custom DSL kernels into PyTorch’s ahead-of-time Inductor stack.
By tensorLorenzo Stoakes targets single-threaded bottlenecks in kallsyms, modpost, nm, and make dependency checks.
By kexec