An OpenGL backend for Halide
Name
868679502-MIT.pdf
Description
Full printable version
Size
1.57 MB
Format
Adobe PDF
Checksum (MD5)
247907ef0f1fa5ec4e711b6fc8ef7d4d
Author(s)
Chornay, Nicholas J
Advisor(s)
Frédo Durand.
Date Issued
2013
Publisher
Massachusetts Institute of Technology
Abstract
High performance image processing requires not only an efficient underlying algorithm but also an implementation tailored to maximally exploit the available hardware resources. In practice, this requires low-level optimization, platform-specific instructions, and, when available, the use of special purpose hardware such as GPU. Halide is a domain-specific programming language targeted at image processing applications. Its programming model decouples an algorithm from the details of its execution, vastly simplifying development and optimization. We present an OpenGL backend for the Halide compiler, which enables Halide programs to run GPU computation on devices that support the OpenGL API. In particular, this paves the way for GPU computation on mobile devices using OpenGL ES. In doing so, we demonstrate how a general image processing framework can be built upon functionality designed for 3D graphics applications.
Description
Thesis (M. Eng.)--Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2013.
Cataloged from PDF version of thesis.
Includes bibliographical references (pages 37).
Subjects
Electrical Engineering and Computer Science.
MIT Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Terms of Use
M.I.T. theses are protected by
copyright. They may be viewed from this source for any purpose, but
reproduction or distribution in any format is prohibited without written
permission. See provided URL for inquiries about permission.
copyright. They may be viewed from this source for any purpose, but
reproduction or distribution in any format is prohibited without written
permission. See provided URL for inquiries about permission.
Persistent DSpace Link
