Algebraic Transformation of Descriptive Vector Byte-code Sequences

Mads Ohm Larsen

doi:10.1145/3009925.3009926

Algebraic Transformation of Descriptive Vector Byte-code Sequences

Mads Ohm Larsen

eScience

1 Citation (Scopus)

50 Downloads (Pure)

Abstract

Both high-productivity and high-performance are two often sought after aspects of scientific programming. Python gives the programmer high-productivity, but even with NumPy it is often not high-performant because of the GIL1, which makes it inherently single threaded. Bohrium intercepts NumPy calls and generates an intermediate language, Bohrium byte-code, before being compiled to OpenCL kernels. It thus grants Python/NumPy the ability to be easily run on multicore systems or GPUs, without changing the source code. The Bohrium byte-code can be optimized, by transforming byte-code sequences into more performant ones. This way, the scientific programmer will not need to change her code to utilize special performant constructs.

Original language	English
Title of host publication	Middleware Doctoral Symposium'16 Proceedings of the Doctoral Symposium of the 17th International Middleware Conference, Trento, Italy — December 12 - 16, 2016
Place of Publication	New York
Publication date	12 Dec 2016
Edition	ACM
Article number	No.1
ISBN (Print)	978-1-4503-4665-8
DOIs	https://doi.org/10.1145/3009925.3009926
Publication status	Published - 12 Dec 2016

Access to Document

10.1145/3009925.3009926

a1-larsenFinal published version, 225 KB

Cite this

Algebraic Transformation of Descriptive Vector Byte-code Sequences. / Larsen, Mads Ohm.

Middleware Doctoral Symposium'16 Proceedings of the Doctoral Symposium of the 17th International Middleware Conference, Trento, Italy — December 12 - 16, 2016 . ACM. ed. New York, 2016. No.1.

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

@inproceedings{cc5a5abe2d504132a6a827248d267c87,

title = "Algebraic Transformation of Descriptive Vector Byte-code Sequences",

abstract = "Both high-productivity and high-performance are two often sought after aspects of scientific programming. Python gives the programmer high-productivity, but even with NumPy it is often not high-performant because of the GIL1, which makes it inherently single threaded. Bohrium intercepts NumPy calls and generates an intermediate language, Bohrium byte-code, before being compiled to OpenCL kernels. It thus grants Python/NumPy the ability to be easily run on multicore systems or GPUs, without changing the source code. The Bohrium byte-code can be optimized, by transforming byte-code sequences into more performant ones. This way, the scientific programmer will not need to change her code to utilize special performant constructs.",

author = "Larsen, {Mads Ohm}",

year = "2016",

month = dec,

day = "12",

doi = "10.1145/3009925.3009926",

language = "English",

isbn = "978-1-4503-4665-8",

booktitle = "Middleware Doctoral Symposium'16 Proceedings of the Doctoral Symposium of the 17th International Middleware Conference, Trento, Italy — December 12 - 16, 2016",

edition = "ACM",

}

TY - GEN

T1 - Algebraic Transformation of Descriptive Vector Byte-code Sequences

AU - Larsen, Mads Ohm

PY - 2016/12/12

Y1 - 2016/12/12

N2 - Both high-productivity and high-performance are two often sought after aspects of scientific programming. Python gives the programmer high-productivity, but even with NumPy it is often not high-performant because of the GIL1, which makes it inherently single threaded. Bohrium intercepts NumPy calls and generates an intermediate language, Bohrium byte-code, before being compiled to OpenCL kernels. It thus grants Python/NumPy the ability to be easily run on multicore systems or GPUs, without changing the source code. The Bohrium byte-code can be optimized, by transforming byte-code sequences into more performant ones. This way, the scientific programmer will not need to change her code to utilize special performant constructs.

AB - Both high-productivity and high-performance are two often sought after aspects of scientific programming. Python gives the programmer high-productivity, but even with NumPy it is often not high-performant because of the GIL1, which makes it inherently single threaded. Bohrium intercepts NumPy calls and generates an intermediate language, Bohrium byte-code, before being compiled to OpenCL kernels. It thus grants Python/NumPy the ability to be easily run on multicore systems or GPUs, without changing the source code. The Bohrium byte-code can be optimized, by transforming byte-code sequences into more performant ones. This way, the scientific programmer will not need to change her code to utilize special performant constructs.

U2 - 10.1145/3009925.3009926

DO - 10.1145/3009925.3009926

M3 - Article in proceedings

SN - 978-1-4503-4665-8

BT - Middleware Doctoral Symposium'16 Proceedings of the Doctoral Symposium of the 17th International Middleware Conference, Trento, Italy — December 12 - 16, 2016

CY - New York

ER -

Algebraic Transformation of Descriptive Vector Byte-code Sequences

Abstract

Access to Document

Fingerprint

Cite this