Proceedings of the 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems 2022
DOI: 10.1145/3503222.3507746
|View full text |Cite
|
Sign up to set email alerts
|

BiSon-e: a lightweight and high-performance accelerator for narrow integer linear algebra computing on the edge

Abstract: Linear algebra computational kernels based on byte and sub-byte integer data formats are at the base of many classes of applications, ranging from Deep Learning to Pattern Matching. Porting the computation of these applications from cloud to edge and mobile devices would enable significant improvements in terms of security, safety, and energy efficiency. However, despite their low memory and energy demands, their intrinsically high computational intensity makes the execution of these workloads challenging on h… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1

Citation Types

0
3
0

Year Published

2023
2023
2024
2024

Publication Types

Select...
5
1

Relationship

1
5

Authors

Journals

citations
Cited by 6 publications
(3 citation statements)
references
References 33 publications
0
3
0
Order By: Relevance
“…In Bison-e [58], the authors propose an area-efficient hardware microarchitecture that enables support for binary segmentation on RISC-V based edge processors. Although both Mix-GEMM and Bison-e exploit binary segmentation to reduce the arithmetic complexity of narrow-integer operations, we identify four critical weaknesses of Bison-e when compared to Mix-GEMM.…”
Section: Physical Design and Energy Efficiencymentioning
confidence: 99%
“…In Bison-e [58], the authors propose an area-efficient hardware microarchitecture that enables support for binary segmentation on RISC-V based edge processors. Although both Mix-GEMM and Bison-e exploit binary segmentation to reduce the arithmetic complexity of narrow-integer operations, we identify four critical weaknesses of Bison-e when compared to Mix-GEMM.…”
Section: Physical Design and Energy Efficiencymentioning
confidence: 99%
“…The content of this thesis and the main contributions have been published in the following conference papers: [128] Enrico Reggiani, Cristóbal Ramírez Lazo, Roger Figueras Bagué, Adrián Cristal, Mauro Olivieri, and Osman Sabri Unsal. BiSon-e: a lightweight and highperformance accelerator for narrow integer linear algebra computing on the edge.…”
Section: Publicationsmentioning
confidence: 99%
“…BiSon-e: a lightweight and highperformance accelerator for narrow integer linear algebra computing on the edge. In particular, the contributions made in Bison-e [128] and Mix-GEMM [129] aim to improve the challenges highlighted in Figure 1.2, considering quantized DNNs with a focus on edge and mobile scenarios. These contributions exploit mathematical transformations (i.e., quantization and binary segmentation), algorithmic improvements of SoA linear algebra frameworks, ISA extensions, and custom hardware accelerators to improve the DNNs computation on edge CPUs considering performance, as well as memory and energy consumption.…”
Section: Publicationsmentioning
confidence: 99%