Modeling and Library Support for Early-stage Exploration of Sparse Tensor Accelerator Designs

Techniques, like pruning and dimension reduction, and characteristics of data for applications, like natural language processing and object detection, introduce sparsity in deep learning models inherently. Sparse tensor accelerators leverage sparsity (0's) in data in order to remove ineffectual...

Full description

Saved in:
Bibliographic Details
Published in:IEEE access 2023-01, Vol.11, p.1-1
Main Authors: Ha, Whoi Ree, Kim, Hyunjun, Paek, Yunheung
Format: Article
Language:English
Subjects:
Accelerators
AI accelerators
Computational modeling
Costs
Deep learning
Design
design space exploration
Encoding
Estimation
Libraries
Mathematical analysis
Modelling
Natural language processing
Object recognition
Space missions
sparse tensor accelerators
Sparsity
sparsity-aware accelerators
Tensors
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Techniques, like pruning and dimension reduction, and characteristics of data for applications, like natural language processing and object detection, introduce sparsity in deep learning models inherently. Sparse tensor accelerators leverage sparsity (0's) in data in order to remove ineffectual computations to speed-up overall run-time. Numerous approaches have been suggested by other researchers such as encoding, decoding, non-zero extraction, load balancing and etc. However, since each implementation requires specialized hardware to accommodate the unique features, the design space becomes much larger compared to regular tensor accelerators when designing a new sparse accelerator. Also these features are hard to compare since the efficiency varies according to application and data sparsity. In this paper, we classify and support the modeling of popularly used features for sparse tensor accelerators. These features enable us to model much larger design space and to estimate their cost more accurately. Library support for these features is also included to make early-stage exploration more realistic. Overall, our experiments show that we can analytically estimate the previously un-modeled components with 93% accuracy on average and provide 19 features as a library support.
ISSN:2169-3536
2169-3536
DOI:10.1109/ACCESS.2023.3278274