Flexible High-Level Synthesis Library for Linear Transformations

被引：0

作者：

Zhao, Wuqiong ^{[1
,2
]}

Li, Changhan ^{[1
,2
]}

Ji, Zhenhao ^{[1
,2
]}

Guo, Zhichen ^{[1
,2
]}

Chen, Xuanbo ^{[1
,2
]}

You, You ^{[1
,2
]}

Huang, Yongming ^{[1
,2
]}

You, Xiaohu ^{[1
,2
]}

Zhang, Chuan ^{[1
,2
]}

机构：

[1] Southeast Univ, LEADS, Natl Mobile Commun Res Lab, Nanjing 211100, Peoples R China

[2] Southeast Univ, Frontiers Sci Ctr Mobile Informat Commun & Secur, Nanjing 211100, Peoples R China

来源：

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS | 2024年 / 71卷 / 07期

关键词：

Libraries; Hardware; Fires; Random access memory; Signal processing algorithms; C plus plus languages; Tensors; High-level synthesis (HLS); linear transformations; compressed sensing; hardware implementation; field programmable gate array (FPGA); MATRICES;

D O I：

10.1109/TCSII.2024.3366282

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

Despite decades pursuing efficient hardware design for signal processing based on linear algebra, traditional hardware description languages (HDLs)-based design workflows remain challenging and time-consuming. High-level synthesis (HLS) provides an easier approach but still requires thorough designs of basic modules concerning linear transformations to achieve acceptable hardware efficiency. To simplify the HLS workflow, we propose the FLAMES library, which provides efficient ready-to-use linear transformation modules. Users can implement algorithms with significantly higher code-writing efficiency via the FLAMES library. We demonstrate its effectiveness by implementing the orthogonal matching pursuit list (OMPL) algorithm for compressed sensing in FPGA, achieving 1.56 x and 1.12 x throughput/slice compared with traditional HLS for the sequential and parallel architecture, respectively.

引用

页码：3348 / 3352

页数：5

共 50 条

[1] RESCHEDULING TRANSFORMATIONS FOR HIGH-LEVEL SYNTHESIS
PAPACHRISTOU, CA
1989 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOLS 1-3, 1989, : 766 - 769
[2] BEHAVIOR-PRESERVING TRANSFORMATIONS FOR HIGH-LEVEL SYNTHESIS
CAMPOSANO, R
LECTURE NOTES IN COMPUTER SCIENCE, 1990, 408 : 106 - 128
[3] HIGH-LEVEL ALGORITHM AND ARCHITECTURE TRANSFORMATIONS FOR DSP SYNTHESIS
PARHI, KK
JOURNAL OF VLSI SIGNAL PROCESSING, 1995, 9 (1-2): : 121 - 143
[4] POLSCA: Polyhedral High-Level Synthesis with Compiler Transformations
Zhao, Ruizhe
Cheng, Jianyi
Luk, Wayne
Constantinides, George A.
2022 32ND INTERNATIONAL CONFERENCE ON FIELD-PROGRAMMABLE LOGIC AND APPLICATIONS, FPL, 2022, : 235 - 242
[5] A Dynamic Memory Allocation Library for High-Level Synthesis
Giamblanco, Nicholas V.
Anderson, Jason H.
2019 29TH INTERNATIONAL CONFERENCE ON FIELD-PROGRAMMABLE LOGIC AND APPLICATIONS (FPL), 2019, : 314 - 320
[6] Library binding for high-level synthesis of analog systems
Ganesan, S
Vemuri, R
VLSI DESIGN 2001: FOURTEENTH INTERNATIONAL CONFERENCE ON VLSI DESIGN, 2001, : 261 - 268
[7] SNAKES: A Flexible High-Level Petri Nets Library (Tool Paper)
Pommereau, Franck
APPLICATION AND THEORY OF PETRI NETS AND CONCURRENCY, 2015, 9115 : 254 - 265
[8] Transformations of High-Level Synthesis Codes for High-Performance Computing
de Fine Licht, Johannes
Besta, Maciej
Meierhans, Simon
Hoefler, Torsten
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS, 2021, 32 (05) : 1014 - 1029
[9] Coordinated transformations for high-level synthesis of high performance microprocessor blocks
Gupta, S
Kam, T
Kishinevsky, M
Rotem, S
Savoiu, N
Dutt, N
Gupta, R
Nicolau, A
39TH DESIGN AUTOMATION CONFERENCE, PROCEEDINGS 2002, 2002, : 898 - 903
[10] Formal Verification of Optimizing Transformations during High-level Synthesis
Chouksey, Ramanuj
Karfa, Chandan
Bhaduri, Purandar
PROCEEDINGS OF THE 12TH INNOVATIONS ON SOFTWARE ENGINEERING CONFERENCE (ISEC), 2019,

← 1 2 3 4 5 →