Systolic Tensor Array: An Efficient Structured-Sparse GEMM Accelerator for Mobile CNN Inference

May 16, 2020 Β· Declared Dead Β· πŸ› IEEE computer architecture letters

πŸ‘» CAUSE OF DEATH: Ghosted
No code link whatsoever

"No code URL or promise found in abstract"

Evidence collected by the PWNC Scanner

Authors Zhi-Gang Liu, Paul N. Whatmough, Matthew Mattina arXiv ID 2005.08098 Category cs.DC: Distributed Computing Cross-listed cs.AR, cs.LG, eess.SP Citations 89 Venue IEEE computer architecture letters Last Checked 3 months ago
Abstract
Convolutional neural network (CNN) inference on mobile devices demands efficient hardware acceleration of low-precision (INT8) general matrix multiplication (GEMM). The systolic array (SA) is a pipelined 2D array of processing elements (PEs), with very efficient local data movement, well suited to accelerating GEMM, and widely deployed in industry. In this work, we describe two significant improvements to the traditional SA architecture, to specifically optimize for CNN inference. Firstly, we generalize the traditional scalar PE, into a Tensor-PE, which gives rise to a family of new Systolic Tensor Array (STA) microarchitectures. The STA family increases intra-PE operand reuse and datapath efficiency, resulting in circuit area and power dissipation reduction of as much as 2.08x and 1.36x respectively, compared to the conventional SA at iso-throughput with INT8 operands. Secondly, we extend this design to support a novel block-sparse data format called density-bound block (DBB). This variant (STA-DBB) achieves a 3.14x and 1.97x improvement over the SA baseline at iso-throughput in area and power respectively, when processing specially-trained DBB-sparse models, while remaining fully backwards compatible with dense models.
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

πŸ“œ Similar Papers

In the same crypt β€” Distributed Computing

Died the same way β€” πŸ‘» Ghosted