Opencl kernel vectorization on the cpu, GPU, and FPGA: A case study with frequent pattern compression

Zheming Jin; Hal Finkel

doi:10.1109/FCCM.2019.00071

Opencl kernel vectorization on the cpu, GPU, and FPGA: A case study with frequent pattern compression

Zheming Jin
, Hal Finkel

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Scopus citations

Abstract

OpenCL promotes code portability, and natively supports vectorized data types, which allows developers to potentially take advantage of the single-instruction-multiple-data instructions on CPUs, GPUs, and FPGAs. FPGAs are becoming a promising heterogeneous computing component. In our study, we choose a kernel used in frequent pattern compression as a case study of OpenCL kernel vectorizations on the three computing platforms. We describe different pattern matching approaches for the kernel, and manually vectorize the OpenCL kernel by a factor ranging from 2 to 16. We evaluate the kernel on an Intel Xeon 16-core CPU, an NVIDIA P100 GPU, and a Nallatech 385A FPGA card featuring an Intel Arria 10 GX1150 FPGA. Compared to the optimized kernel that is not vectorized, our vectorization can improve the kernel performance by a factor of 16 on the FPGA. The performance improvement ranges from 1 to 11.4 on the CPU, and from 1.02 to 9.3 on the GPU. The effectiveness of kernel vectorization depends on the work-group size.

Original language	English
Title of host publication	Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	330
Number of pages	1
ISBN (Electronic)	9781728111315
DOIs	https://doi.org/10.1109/FCCM.2019.00071
State	Published - Apr 2019
Externally published	Yes
Event	27th Annual IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019 - San Diego, United States Duration: Apr 28 2019 → May 1 2019

Publication series

Name	Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019

Conference

Conference	27th Annual IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019
Country/Territory	United States
City	San Diego
Period	04/28/19 → 05/1/19

Funding

The research was supported by the U.S. Department of Energy, Office of Science, under contract DE-AC02-06CH11357 and made use of the Argonne Leadership Computing Facility, a DOE Office of Science User Facility XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

Keywords

FPGA
OpenCL
Vectorization

Access to Document

10.1109/FCCM.2019.00071

Cite this

Jin, Z., & Finkel, H. (2019). Opencl kernel vectorization on the cpu, GPU, and FPGA: A case study with frequent pattern compression. In Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019 (pp. 330). Article 8735499 (Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/FCCM.2019.00071

Jin, Zheming ; Finkel, Hal. / Opencl kernel vectorization on the cpu, GPU, and FPGA : A case study with frequent pattern compression. Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 330 (Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019).

@inproceedings{d102243b06cd4cb6b59028d9de8bd5c9,

title = "Opencl kernel vectorization on the cpu, GPU, and FPGA: A case study with frequent pattern compression",

abstract = "OpenCL promotes code portability, and natively supports vectorized data types, which allows developers to potentially take advantage of the single-instruction-multiple-data instructions on CPUs, GPUs, and FPGAs. FPGAs are becoming a promising heterogeneous computing component. In our study, we choose a kernel used in frequent pattern compression as a case study of OpenCL kernel vectorizations on the three computing platforms. We describe different pattern matching approaches for the kernel, and manually vectorize the OpenCL kernel by a factor ranging from 2 to 16. We evaluate the kernel on an Intel Xeon 16-core CPU, an NVIDIA P100 GPU, and a Nallatech 385A FPGA card featuring an Intel Arria 10 GX1150 FPGA. Compared to the optimized kernel that is not vectorized, our vectorization can improve the kernel performance by a factor of 16 on the FPGA. The performance improvement ranges from 1 to 11.4 on the CPU, and from 1.02 to 9.3 on the GPU. The effectiveness of kernel vectorization depends on the work-group size.",

keywords = "FPGA, OpenCL, Vectorization",

author = "Zheming Jin and Hal Finkel",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 27th Annual IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019 ; Conference date: 28-04-2019 Through 01-05-2019",

year = "2019",

month = apr,

doi = "10.1109/FCCM.2019.00071",

language = "English",

series = "Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "330",

booktitle = "Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019",

}

Jin, Z & Finkel, H 2019, Opencl kernel vectorization on the cpu, GPU, and FPGA: A case study with frequent pattern compression. in Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019., 8735499, Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019, Institute of Electrical and Electronics Engineers Inc., pp. 330, 27th Annual IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019, San Diego, United States, 04/28/19. https://doi.org/10.1109/FCCM.2019.00071

Opencl kernel vectorization on the cpu, GPU, and FPGA: A case study with frequent pattern compression. / Jin, Zheming; Finkel, Hal.
Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019. Institute of Electrical and Electronics Engineers Inc., 2019. p. 330 8735499 (Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Opencl kernel vectorization on the cpu, GPU, and FPGA

T2 - 27th Annual IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019

AU - Jin, Zheming

AU - Finkel, Hal

PY - 2019/4

Y1 - 2019/4

N2 - OpenCL promotes code portability, and natively supports vectorized data types, which allows developers to potentially take advantage of the single-instruction-multiple-data instructions on CPUs, GPUs, and FPGAs. FPGAs are becoming a promising heterogeneous computing component. In our study, we choose a kernel used in frequent pattern compression as a case study of OpenCL kernel vectorizations on the three computing platforms. We describe different pattern matching approaches for the kernel, and manually vectorize the OpenCL kernel by a factor ranging from 2 to 16. We evaluate the kernel on an Intel Xeon 16-core CPU, an NVIDIA P100 GPU, and a Nallatech 385A FPGA card featuring an Intel Arria 10 GX1150 FPGA. Compared to the optimized kernel that is not vectorized, our vectorization can improve the kernel performance by a factor of 16 on the FPGA. The performance improvement ranges from 1 to 11.4 on the CPU, and from 1.02 to 9.3 on the GPU. The effectiveness of kernel vectorization depends on the work-group size.

AB - OpenCL promotes code portability, and natively supports vectorized data types, which allows developers to potentially take advantage of the single-instruction-multiple-data instructions on CPUs, GPUs, and FPGAs. FPGAs are becoming a promising heterogeneous computing component. In our study, we choose a kernel used in frequent pattern compression as a case study of OpenCL kernel vectorizations on the three computing platforms. We describe different pattern matching approaches for the kernel, and manually vectorize the OpenCL kernel by a factor ranging from 2 to 16. We evaluate the kernel on an Intel Xeon 16-core CPU, an NVIDIA P100 GPU, and a Nallatech 385A FPGA card featuring an Intel Arria 10 GX1150 FPGA. Compared to the optimized kernel that is not vectorized, our vectorization can improve the kernel performance by a factor of 16 on the FPGA. The performance improvement ranges from 1 to 11.4 on the CPU, and from 1.02 to 9.3 on the GPU. The effectiveness of kernel vectorization depends on the work-group size.

KW - FPGA

KW - OpenCL

KW - Vectorization

UR - https://www.scopus.com/pages/publications/85068330001

U2 - 10.1109/FCCM.2019.00071

DO - 10.1109/FCCM.2019.00071

M3 - Conference contribution

AN - SCOPUS:85068330001

T3 - Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019

SP - 330

BT - Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 28 April 2019 through 1 May 2019

ER -

Jin Z, Finkel H. Opencl kernel vectorization on the cpu, GPU, and FPGA: A case study with frequent pattern compression. In Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019. Institute of Electrical and Electronics Engineers Inc. 2019. p. 330. 8735499. (Proceedings - 27th IEEE International Symposium on Field-Programmable Custom Computing Machines, FCCM 2019). doi: 10.1109/FCCM.2019.00071

Opencl kernel vectorization on the cpu, GPU, and FPGA: A case study with frequent pattern compression

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this