Snafu: An Ultra-Low-Power, Energy-Minimal CGRA-Generation Framework and Architecture
Graham Gobieski, Ahmet Oguz Atli, Kenneth Mai, Brandon Lucia, Nathan Beckmann
Abstract
Ultra-low-power (ULP) devices are becoming pervasive, enabling many emerging sensing applications. Energy-efficiency is paramount in these applications, as efficiency determines device lifetime in battery-powered deployments and performance in energy-harvesting deployments. Unfortunately, existing designs fall short because ASICs’ upfront costs are too high and prior ULP architectures are too inefficient or inflexible.We present Snafu, the first framework to flexibly generate ULP coarse-grain reconfigurable arrays (CGRAs). Snafu provides a standard interface for processing elements (PE), making it easy to integrate new types of PEs for new applications. Unlike prior high-performance, high-power CGRAs, Snafu is designed from the ground up to minimize energy consumption while maximizing flexibility. Snafu saves energy by configuring PEs and routers for a single operation to minimize switching activity; by minimizing buffering within the fabric; by implementing a statically routed, bufferless, multi-hop network; and by executing operations in-order to avoid expensive tag-token matching.We further present Snafu-Arch, a complete ULP system that integrates an instantiation of the Snafu fabric alongside a scalar RISC-V core and memory. We implement Snafu in RTL and evaluate it on an industrial sub-28 nm FinFET process across a suite of common sensing benchmarks. Snafu-Arch operates at <1 mW, orders-of-magnitude less power than most prior CGRAs. Snafu-Arch uses 41% less energy and runs 4.4× faster than the prior state-of-the-art general-purpose ULP architecture. Moreover, we conduct three comprehensive case-studies to quantify the cost of programmability in Snafu. We find that Snafu-Arch is close to ASIC designs built in the same technology, using just 2.6× more energy on average.
Ask about this paper
Ask your agent about it.
Lune has read the top-tier papers around this one, so every answer names the papers it rests on.
Your agent calls
Lunesearch_papers
Free to start. No credit card required.
Terminal
Install the CLIlune papers get 231fd7c8-8f7e-4925-858a-bc6bcca323c3Cited by top-tier papers18
- REVAMP: a systematic framework for heterogeneous CGRA realizationThilini Kaushalya Bandara, Dhananjaya Wijerathne, Tulika Mitra, Li-Shiuan PehASPLOS 2022 · 64 citations
- SCALO: An Accelerator-Rich Distributed System for Scalable Brain-Computer InterfacingKarthik Sriram, Raghavendra Pradyumna Pothukuchi, Michal Gerasimiuk, Muhammed Ugur et al.ISCA 2023 · 18 citations
- PICACHU: Plug-In CGRA Handling Upcoming Nonlinear Operations in LLMsJiajun Qin, Tianhua Xia, Cheng Tan, Jeff Zhang et al.ASPLOS 2025 · 17 citations
- Pipestitch: An energy-minimal dataflow architecture with lightweight threadsNathan Serafin, Souradip Ghosh, Harsh Desai, Nathan Beckmann et al.MICRO 2023 · 16 citations
- Towards Efficient Control Flow Handling in Spatial Architecture via Architecting the Control Flow PlaneJinyi Deng, Xinru Tang, Jiahao Zhang, Yuxuan Li et al.MICRO 2023 · 15 citations
Related papers
- A programmable, energy-minimal dataflow compiler and architectureGraham Gobieski, Souradip Ghosh, Marijn Heule, Todd C. Mowry et al.MICRO 2022 · 66 citations
- Ultra-Elastic CGRAs for Irregular Loop SpecializationChristopher Torng, Peitian Pan, Yanghui Ou, Cheng Tan et al.HPCA 2021 · 68 citations
- APEX: A Framework for Automated Processing Element Design Space Exploration using Frequent Subgraph AnalysisJackson Melchert, Kathleen Feng, Caleb Donovick, Ross Daly et al.ASPLOS 2023 · 12 citations
- ML-CGRA: An Integrated Compilation Framework to Enable Efficient Machine Learning Acceleration on CGRAsYixuan Luo, Cheng Tan, Nicolas Bohm Agostini, Ang Li et al.DAC 2023 · 39 citations
- PT-Map: Efficient Program Transformation Optimization for CGRA MappingBizhao Shi, Tuo Dai, Jiaxi Zhang, Xuechao Wei et al.DAC 2024 · 1 citation
