Towards a Linear-Algebraic Hypervisor
summary
Describes a linear-algebraic hypervisor that compiles a tiny array language into a branch-free multilinear transition operator and executes thousands to millions of tiny virtual machines on GPU threads. The pipeline samples syntactic programs, compiles to fixed-word polynomials, dispatches VMs in chunks, and measures wall-clock throughput, reporting speedups up to 147× relative to serial evaluation across sampled programs and typical halting within <100 steps for the majority.
AI-written summary · about