The Dataflow Abstract Machine Simulator Framework

The growing interest in novel dataflow architectures and streaming execution paradigms has created the need for a simulator optimized for modeling dataflow systems.

To fill this need, we present three new techniques that make it feasible to simulate complex systems consisting of thousands of components. First, we introduce an interface based on Communicating Sequential Processes which allows users to simultaneously describe functional and timing characteristics. Second, we introduce a scalable point-to-point synchronization scheme that avoids global synchronization. Finally, we demonstrate a technique to exploit slack in the simulated system, such as FIFOs, to increase simulation parallelism.

We implement these techniques in the Dataflow Abstract Machine (DAM), a parallel simulator framework for dataflow systems. We demonstrate the benefits of using DAM by highlighting three case studies using the framework. First, we use DAM directly as an exploration tool for streaming algorithms on dataflow hardware. We simulate two different implementations of the attention algorithm used in large language models, and use DAM to show that the second implementation only requires a constant amount of local memory. Second, we re-implement a simulator for a sparse tensor algebra accelerator, resulting in 57% less code and a simulation speedup of up to four orders of magnitude. Finally, we demonstrate a general technique for timemultiplexing real hardware to simulate multiple virtual copies of the hardware using DAM.

Modern applications such as large language models (LLMs), data analytics, and sparse machine learning have ignited a flurry of research in both dataflow architectures, such as Reconfigurable Dataflow Accelerators (RDAs) and Coarse-Grained Reconfigurable Arrays (CGRAs) [9], [16], [26], [42], [43], [45], [47], and streaming abstractions such as the Sparse Abstract Machine [29]. To explore the functional behavior and performance characteristics of their proposed dataflow systems, many researchers develop bespoke simulators.

To simulate dataflow systems, a framework must (1) support communication among thousands of coupled units, (2) model fine-grained channel behaviors, and (3) simulate heterogenous user-defined models. Furthermore, the scale of these systems demands efficient parallelization. Providing these properties in sequential simulation is straightforward; however, no efficient method exists for parallel simulation.