Serving Recurrent Neural Networks Efficiently with a Spatial Accelerator

September 26, 2019 Β· Declared Dead Β· πŸ› USENIX workshop on Tackling computer systems problems with machine learning techniques

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

"No code URL or promise found in abstract"

Evidence collected by the PWNC Scanner

Authors Tian Zhao, Yaqi Zhang, Kunle Olukotun arXiv ID 1909.13654 Category cs.DC: Distributed Computing Cross-listed cs.LG, cs.PF Citations 16 Venue USENIX workshop on Tackling computer systems problems with machine learning techniques Last Checked 4 months ago
Abstract
Recurrent Neural Network (RNN) applications form a major class of AI-powered, low-latency data center workloads. Most execution models for RNN acceleration break computation graphs into BLAS kernels, which lead to significant inter-kernel data movement and resource underutilization. We show that by supporting more general loop constructs that capture design parameters in accelerators, it is possible to improve resource utilization using cross-kernel optimization without sacrificing programmability. Such abstraction level enables a design space search that can lead to efficient usage of on-chip resources on a spatial architecture across a range of problem sizes. We evaluate our optimization strategy on such abstraction with DeepBench using a configurable spatial accelerator. We demonstrate that this implementation provides a geometric speedup of 30x in performance, 1.6x in area, and 2x in power efficiency compared to a Tesla V100 GPU, and a geometric speedup of 2x compared to Microsoft Brainwave implementation on a Stratix 10 FPGA.
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