Extending halide to improve software development for imaging DSPs

S. Vocke; H. Corporaal; R. Jordans; R. Corvino; Rick J.M. Nas

doi:10.1145/3106343

Extending halide to improve software development for imaging DSPs

S. Vocke, H. Corporaal, R. Jordans, R. Corvino, Rick J.M. Nas

Electronic Systems

Onderzoeksoutput: Bijdrage aan tijdschrift › Tijdschriftartikel › Academic › peer review

13 Citaten (Scopus)

5 Downloads (Pure)

Samenvatting

Specialized Digital Signal Processors (DSPs) play an important role in power-efficient, high-performance image processing. However, developing applications for DSPs is more time-consuming and error-prone than for general-purpose processors. Halide is a domain-specific language (DSL) which enables low-effort development of portable, high-performance imaging pipelines. We propose a set of extensions to Halide to support DSPs in combination with arbitrary C compilers, including a template solution to support scratchpad memories. Using a commercial DSP, we demonstrate that this solution can achieve performance within 20% of highly tuned C code, while leading to a reduction in development time and code complexity.

Originele taal-2	Engels
Artikelnummer	21
Aantal pagina's	25
Tijdschrift	ACM Transactions on Architecture and Code Optimization
Volume	14
Nummer van het tijdschrift	3
DOI's	https://doi.org/10.1145/3106343
Status	Gepubliceerd - sep. 2017

Toegang tot document

10.1145/3106343

Andere bestanden en links

Link to publication in Scopus

Citeer dit

@article{3d29378834944408a879badf002075e1,

title = "Extending halide to improve software development for imaging DSPs",

abstract = "Specialized Digital Signal Processors (DSPs) play an important role in power-efficient, high-performance image processing. However, developing applications for DSPs is more time-consuming and error-prone than for general-purpose processors. Halide is a domain-specific language (DSL) which enables low-effort development of portable, high-performance imaging pipelines. We propose a set of extensions to Halide to support DSPs in combination with arbitrary C compilers, including a template solution to support scratchpad memories. Using a commercial DSP, we demonstrate that this solution can achieve performance within 20% of highly tuned C code, while leading to a reduction in development time and code complexity.",

keywords = "Accelerator, Asip, DSP, Embedded processor, Halide, IPU, Image processing unit, Optimization, Parallel computing, Performance optimization, SIMD, Scratchpad memory, VLIW",

author = "S. Vocke and H. Corporaal and R. Jordans and R. Corvino and Nas, {Rick J.M.}",

year = "2017",

month = sep,

doi = "10.1145/3106343",

language = "English",

volume = "14",

journal = "ACM Transactions on Architecture and Code Optimization",

issn = "1544-3566",

publisher = "Association for Computing Machinery, Inc",

number = "3",

}

TY - JOUR

T1 - Extending halide to improve software development for imaging DSPs

AU - Vocke, S.

AU - Corporaal, H.

AU - Jordans, R.

AU - Corvino, R.

AU - Nas, Rick J.M.

PY - 2017/9

Y1 - 2017/9

N2 - Specialized Digital Signal Processors (DSPs) play an important role in power-efficient, high-performance image processing. However, developing applications for DSPs is more time-consuming and error-prone than for general-purpose processors. Halide is a domain-specific language (DSL) which enables low-effort development of portable, high-performance imaging pipelines. We propose a set of extensions to Halide to support DSPs in combination with arbitrary C compilers, including a template solution to support scratchpad memories. Using a commercial DSP, we demonstrate that this solution can achieve performance within 20% of highly tuned C code, while leading to a reduction in development time and code complexity.

AB - Specialized Digital Signal Processors (DSPs) play an important role in power-efficient, high-performance image processing. However, developing applications for DSPs is more time-consuming and error-prone than for general-purpose processors. Halide is a domain-specific language (DSL) which enables low-effort development of portable, high-performance imaging pipelines. We propose a set of extensions to Halide to support DSPs in combination with arbitrary C compilers, including a template solution to support scratchpad memories. Using a commercial DSP, we demonstrate that this solution can achieve performance within 20% of highly tuned C code, while leading to a reduction in development time and code complexity.

KW - Accelerator

KW - Asip

KW - DSP

KW - Embedded processor

KW - Halide

KW - IPU

KW - Image processing unit

KW - Optimization

KW - Parallel computing

KW - Performance optimization

KW - SIMD

KW - Scratchpad memory

KW - VLIW

UR - http://www.scopus.com/inward/record.url?scp=85029543730&partnerID=8YFLogxK

U2 - 10.1145/3106343

DO - 10.1145/3106343

M3 - Article

SN - 1544-3566

VL - 14

JO - ACM Transactions on Architecture and Code Optimization

JF - ACM Transactions on Architecture and Code Optimization

IS - 3

M1 - 21

ER -

Extending halide to improve software development for imaging DSPs

Samenvatting

Toegang tot document

Andere bestanden en links

Vingerafdruk

Citeer dit