CHIPSIM: A Co-Simulation Framework for Deep Learning on Chiplet-Based Systems
By: Lukas Pfromm , Alish Kanani , Harsh Sharma and more
Potential Business Impact:
Tests computer chips faster and more accurately.
Due to reduced manufacturing yields, traditional monolithic chips cannot keep up with the compute, memory, and communication demands of data-intensive applications, such as rapidly growing deep neural network (DNN) models. Chiplet-based architectures offer a cost-effective and scalable solution by integrating smaller chiplets via a network-on-interposer (NoI). Fast and accurate simulation approaches are critical to unlocking this potential, but existing methods lack the required accuracy, speed, and flexibility. To address this need, this work presents CHIPSIM, a comprehensive co-simulation framework designed for parallel DNN execution on chiplet-based systems. CHIPSIM concurrently models computation and communication, accurately capturing network contention and pipelining effects that conventional simulators overlook. Furthermore, it profiles the chiplet and NoI power consumptions at microsecond granularity for precise transient thermal analysis. Extensive evaluations with homogeneous/heterogeneous chiplets and different NoI architectures demonstrate the framework's versatility, up to 340% accuracy improvement, and power/thermal analysis capability.
Similar Papers
Simulation-Driven Evaluation of Chiplet-Based Architectures Using VisualSim
Hardware Architecture
Builds better computer chips from smaller pieces.
NeuroScalar: A Deep Learning Framework for Fast, Accurate, and In-the-Wild Cycle-Level Performance Prediction
Hardware Architecture
Tests new computer chips super fast.
Chiplet-Based RISC-V SoC with Modular AI Acceleration
Hardware Architecture
Makes smart gadgets faster and use less power.