1.5k citations
- Rice UniversityUS7 papers
- Xilinx (Ireland)IE5 papers
- Cornell UniversityUS3 papers
- Norwegian University of Science and TechnologyNO3 papers
- The University of SydneyAU2 papers
- Amazon (Germany)DE1 paper
- Atlanta Technical CollegeUS1 paper
- Chinese Academy of SciencesCN1 paper
- Delft University of TechnologyNL1 paper
- École Polytechnique Fédérale de LausanneCH1 paper
- Imperial College LondonGB1 paper
- KU LeuvenBE1 paper
24 papers
Bring Your Own Codegen to Deep Learning Compiler
Zhi Chen, Cody Hao Yu, Trevor Morris +6
Deep neural networks (DNNs) have been ubiquitously applied in many applications, and accelerators are emerged as an enabler to support the fast and efficient inference tasks of the…
Comparison of Different Methods for Time Sequence Prediction in Autonomous Vehicles
Teng Liu, Bin Tian, Yunfeng Ai +3
As a combination of various kinds of technologies, autonomous vehicles could complete a series of driving tasks by itself, such as perception, decision-making, planning, and contro…
Compiler Directed Speculative Intermittent Computation
Jongouk Choi, Qingrui Liu, Changhee Jung
This paper presents CoSpec, a new architecture/compiler co-design scheme that works for commodity in-order processors used in energy-harvesting systems. To achieve crash consistenc…
Vyasa: A High-Performance Vectorizing Compiler for Tensor Convolutions on the Xilinx AI Engine
Prasanth Chatarasi, Stephen Neuendorffer, Samuel Bayliss +2
Xilinx's AI Engine is a recent industry example of energy-efficient vector processing that includes novel support for 2D SIMD datapaths and shuffle interconnection network. The cur…
Evolutionary Bin Packing for Memory-Efficient Dataflow Inference Acceleration on FPGA
Mairin Kroes, Lucian Petrica, Sorin Cotofana +1
Convolutional neural network (CNN) dataflow inference accelerators implemented in Field Programmable Gate Arrays (FPGAs) have demonstrated increased energy efficiency and lower lat…
Learning Accurate Integer Transformer Machine-Translation Models
Ephrem Wu
We describe a method for training accurate Transformer machine-translation models to run inference using 8-bit integer (INT8) hardware matrix multipliers, as opposed to the more co…