Search papers, labs, and topics across Lattice.
This paper introduces MeshReduce-U, a compiler-guided framework designed to optimize communication reduction and routing for irregular neural workloads on mesh NoCs. By coalescing colocated sources and forming local aggregation islands, MeshReduce-U achieves significant reductions in communication latency and link usage metrics across a diverse set of neural network workloads. The framework demonstrates a mean latency reduction of 40.3% and total link usage reduction of 56.0%, outperforming traditional mapping techniques and highlighting the benefits of pre-routing communication optimization.
Achieving over 40% reductions in latency and link usage reveals that optimizing neural communication before routing can dramatically enhance performance in irregular workloads.
Many irregular neural workloads induce skewed many-to-one reductions with repeated neighborhoods and nonlocal communication. Conventional NoC mappers optimize placement and routes for a fixed communication graph, even though associative reductions expose legal opportunities to eliminate traffic before routing. We present MeshReduce-U, a compiler-guided communication-reduction and routing framework for mesh-NoC-based spatial accelerators. MeshReduce-U coalesces colocated sources, forms local aggregation islands, blocks channels with compatible fan-in structure, selects capacity-feasible sinks, and routes the remaining fixed-width carriers using fused usage-aware costs. A deterministic route-replay model reports schedule-derived communication latency, total link usage (TLU), and fused link usage (FusedTLU) separately. Across a 20-workload lowerable neural-network zoo, MeshReduce-U reduces mean latency, TLU, and FusedTLU by 40.3%, 56.0%, and 48.7%, respectively, relative to an ABC-style source-aggregation baseline, improving all three metrics on every workload. Across 40 synthetic irregular reductions, it reduces mean latency and TLU by 12.3% and 19.7%. A new 30-instance pass-by-pass study further shows that the structural rewrites reduce the global carrier count by 60.9% and replay latency by 63.0%. These results show that rewriting reducible neural communication before routing can be more effective than searching harder over an unreduced traffic graph.
