Towards a Linear-Algebraic Hypervisor

TL;DR

This paper introduces a parallel virtual machine optimized for array programs, achieving significant speedups on GPUs for concurrent program evaluation.

Contribution

It presents a novel linear-algebraic hypervisor enabling efficient parallel execution of array programs on GPUs, outperforming serial evaluation.

Findings

Achieved up to 147x speedup on GPU for concurrent array programs.

Demonstrated the hypervisor's effectiveness in program synthesis and superoptimization tasks.

Evaluated performance with millions of concurrent programs.

Abstract

Many techniques in program synthesis, superoptimization, and array programming require parallel rollouts of general-purpose programs. GPUs, while capable targets for domain-specific parallelism, are traditionally underutilized by such workloads. Motivated by this opportunity, we introduce a pleasingly parallel virtual machine and benchmark its performance by evaluating millions of concurrent array programs, observing speedups up to $147\times$ relative to serial evaluation.