Optimizing batched winograd convolution on GPUs
Authors:
Open Access
- 19 February 2020
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 32-44
Abstract
No abstract availableKeywords
Funding Information
- HK Research Grants Council (26213818)
This publication has 8 references indexed in Scilit:
- Anatomy of High-Performance Deep Learning Convolutions on SIMD Architectures2018
- Optimizing N-dimensional, winograd-based convolution for manycore CPUs2018
- Understanding the GPU Microarchitecture to Achieve Bare-Metal Performance Tuning2017
- Benchmarking State-of-the-Art Deep Learning Software Tools2016
- Fast Algorithms for Convolutional Neural Networks2016
- Benchmarking the Memory Hierarchy of Modern GPUs2014
- The cache performance and optimizations of blocked algorithms1991
- Arithmetic Complexity of Computations1980