Search papers, labs, and topics across Lattice.
The paper introduces GauS, a differentiable framework for operator scheduling that models schedules as stochastic relaxations using Gaussian distributions to better capture the ordinal nature of time and leverage GPU parallelism. This approach addresses the limitations of existing differentiable scheduling methods that rely on categorical distributions and suffer from scalability issues. Experiments on various benchmarks demonstrate that GauS achieves Pareto-optimal results, highlighting its effectiveness in complex pipelined scheduling problems.
Differentiable operator scheduling gets a Gaussian makeover, shrinking the optimization space and unlocking Pareto-optimal results on complex pipelined problems.
Efficient operator scheduling is a fundamental challenge in software compilation and hardware synthesis. While recent differentiable approaches have sought to replace traditional ones like exact solvers or heuristics with gradient-based search, they typically rely on categorical distributions that fail to capture the ordinal nature of time and suffer from a parameter space that scales poorly. In this paper, we propose a novel differentiable framework, GauS, that models operator scheduling as a stochastic relaxation using Gaussian distributions, which fully utilize modern parallel computing devices like GPUs. By representing schedules as continuous Gaussian variables, we successfully capture the ordinal nature of time and reduce the optimization space by orders of magnitude. Our method is highly flexible to represent various objectives and constraints, which provides the first differentiable formulation for the complex pipelined scheduling problem. We evaluate our method on a range of benchmarks, demonstrating that Gaus achieves Pareto-optimal results.
