Removing Host Interventions from GPU Accelerated Neural Network

被引：0

作者：

Yoo, Jeongjoon ^{[1
]}

Oh, Kyongjoo ^{[1
]}

Jun, Jaehee ^{[1
]}

Cho, Hoonhee ^{[1
]}

Kim, Kyeongmin ^{[1
]}

机构：

[1] Samsung Elect, Suwon, South Korea

来源：

2023 IEEE INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS, ICCE | 2023年

关键词：

OpenCL; GPU; Neural Network; Performance; Delegator; Host intervention;

D O I：

10.1109/ICCE56470.2023.10043523

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

In this paper, we present a novel scheme removing host interventions from OpenCL execution in GPUaccelerated neural network. Our scheme must satisfy two folds; i) remove host interventions between CPU and GPU interactions, ii) use previous OpenCL kernels without modification. To do so, we propose a Delegator-based OpenCL execution method providing a good solution in such situation. Experimental result shows that our scheme reduces execution latency into 0.43.

引用

页数：2

共 50 条

[21] Neural network accelerated process design of polycrystalline microstructures
Lin, Junrong
Hasan, Mahmudul
Acar, Pinar
Blanchet, Jose
Tarokh, Vahid
MATERIALS TODAY COMMUNICATIONS, 2023, 36
[22] Bayesian Optimization for Auto-tuning Convolution Neural Network on GPU
Zhu, Huming
Liu, Chendi
Zhang, Lingyun
Dong, Ximiao
ALGORITHMS AND ARCHITECTURES FOR PARALLEL PROCESSING, ICA3PP 2023, PT VI, 2024, 14492 : 478 - 489
[23] Parallelizing Neural Network Models Effectively on GPU by Implementing Reductions Atomically
Zhao, Jie
Bastoul, Cedric
Yi, Yanzhi
Hu, Jiahui
Nie, Wang
Zhang, Renwei
Geng, Zhen
Li, Chong
Tachon, Thibaut
Gan, Zhiliang
PROCEEDINGS OF THE 2022 31ST INTERNATIONAL CONFERENCE ON PARALLEL ARCHITECTURES AND COMPILATION TECHNIQUES, PACT 2022, 2022, : 451 - 466
[24] A Power Efficient Neural Network Implementation on Heterogeneous FPGA and GPU Devices
Tu, Yuexuan
Sadiq, Saad
Tao, Yudong
Shyu, Mei-Ling
Chen, Shu-Ching
2019 IEEE 20TH INTERNATIONAL CONFERENCE ON INFORMATION REUSE AND INTEGRATION FOR DATA SCIENCE (IRI 2019), 2019, : 193 - 199
[25] A GPU-Based Convolutional Neural Network Approach for Image Classification
Cengil, Emine
Cinar, Ahmet
Guler, Zafer
2017 INTERNATIONAL ARTIFICIAL INTELLIGENCE AND DATA PROCESSING SYMPOSIUM (IDAP), 2017,
[26] Neural Network Reliability Enhancement Approach Using Dropout Underutilization in GPU
Lee, Dongsu
Lim, Hyunyul
Kim, Tae Hyun
Kang, Sungho
PROCEEDINGS OF TENCON 2018 - 2018 IEEE REGION 10 CONFERENCE, 2018, : 2281 - 2286
[27] A Method for Reducing Noise Radiated from Structures with Vibration Absorbers by Using an Accelerated Neural Network
李连进
葛为民
TransactionsofTianjinUniversity, 2004, (01) : 9 - 15
[28] Accelerated gradient learning algorithm for neural network weights update
Željko Hocenski
Mladen Antunoviæ
Damir Filko
Neural Computing and Applications, 2010, 19 : 219 - 225
[29] Accelerated gradient learning algorithm for neural network weights update
Hocenski, Zeljko
Antunovic, Mladen
Filko, Damir
KNOWLEDGE - BASED INTELLIGENT INFORMATION AND ENGINEERING SYSTEMS, PT 1, PROCEEDINGS, 2008, 5177 : 49 - 56
[30] Novel accelerated methods for convolution neural network with matrix core
Guo, Yijie
Lu, Lu
Zhu, Songxiang
JOURNAL OF SUPERCOMPUTING, 2023, 79 (17) : 19547 - 19573

← 1 2 3 4 5 →