Low-Power Parallel Computing on GPUs 2

Informations projet

LPGPU2

N° de convention de subvention: 688759

DOI

10.3030/688759

Projet clôturé

Date de signature de la CE 10 Novembre 2015

Date de début 1 Janvier 2016

Date de fin 30 Septembre 2018

Financé au titre de

• INDUSTRIAL LEADERSHIP - Leadership in enabling and industrial technologies - Information and Communication Technologies (ICT)

Coût total

€ 3 954 846,25

Contribution de l’UE

€ 2 975 786,88

Coordonné par

TECHNISCHE UNIVERSITAT BERLIN
Germany

CORDIS fournit des liens vers les livrables publics et les publications des projets HORIZON.

Les liens vers les livrables et les publications des projets du 7e PC, ainsi que les liens vers certains types de résultats spécifiques tels que les jeux de données et les logiciels, sont récupérés dynamiquement sur OpenAIRE .

Livrables

Documents, reports (7)

Report on power model for mobile SoCs based on hardware performance counters

This report will contain a description of the power model developed in T4.2, the microbenchmark suite and the power measurement testbed from T4.1. It will also provide insight into the capabilities of the various performance counters for power estimation. The report will also provide an evaluation of the power estimation tool. The power consumption of applications will be measured using the testbed and compared to the estimates provided by the tool

Public report on LPGPU2 applications

Public report on LPGPU2 applications. This report will describe the public part of the applications optimizations in T3.1, T3.4, T3.5, and T3.6.

Profiling-driven DVFS report

This report will summarize the profile driven DVFS algorithm

Final Report on Tool Validation, Application Optimizations, and GPU Customization

This deliverable will report the outcome of the tool validation phase. All the tasks that are active in the last semester of the project (T7.3, T7.4, and T7.5) will contribute to this deliverable

Delivering the LPGPU2 tool to the Open Community

This deliverable will be prepared as part of Task 7.7. It will contain instructions to download (and install) the LPGPU2 tool from a public repository. In addition, it will include a part in the form of application note about best-practices of the tool’s usage.

White paper on Tool Validation, Application Optimizations, and GPU Customization

White paper on Tool Validation, Application Optimizations, and GPU Customization. This deliverable will prepare the material from D7.2 as a white paper, to ensure wider dissemination.

Final Periodic Report

Final Report of the LPGPU2 project

Websites, patent fillings, videos etc. (4)

Final press release
First press release

Initial press release

Marketing materials (poster, flyers and document templates, Project web-site, creation of social media accounts, Project logo)
Public Summary

A 4-page public summary of the project targeted at the general public

Demonstrators, pilots, prototypes (1)

Industry standard performance monitoring API

This deliverable will contain a proposed definition of performance monitoring APIs as part of Khronos standards, including OpenGL ES and OpenCL

Publications

Conference proceedings (6)

Optimal DC/AC data bus inversion coding

Auteurs: Jan Lucas, Sohan Lal, Ben Juurlink
Publié dans: 2018 Design, Automation & Test in Europe Conference & Exhibition (DATE), Numéro 2018, 2018, Page(s) 1063-1068, ISBN 978-3-9819263-0-9
Éditeur: IEEE
DOI: 10.23919/DATE.2018.8342169

Enabling GPU software developers to optimize their applications — The LPGPU 2 approach

Auteurs: Ben Juurlink, Jan Lucas, Nadjib Mammeri, Georgios Keramidas, Katerina Pontzolkova, Ignacio Aransay, Chrysa Kokkala, Martyn Bliss, Andrew Richards
Publié dans: 2017 Conference on Design and Architectures for Signal and Image Processing (DASIP), Numéro 2017, 2017, Page(s) 1-6, ISBN 978-1-5386-3534-6
Éditeur: IEEE
DOI: 10.1109/DASIP.2017.8122116

The LPGPU2 Project - Low-Power Parallel Computing on GPUs: Extended Abstract

Auteurs: Ben Juurlink, Jan Lucas, Nadjib Mammeri, Martyn Bliss, Georgios Keramidas, Chrysa Kokkala, Andrew Richards
Publié dans: Proceedings of the 20th International Workshop on Software and Compilers for Embedded Systems - SCOPES '17, Numéro 2017, 2017, Page(s) 76-80, ISBN 9781-450350396
Éditeur: ACM Press
DOI: 10.1145/3078659.3078672

E^2MC: Entropy Encoding Based Memory Compression for GPUs

Auteurs: Sohan Lal, Jan Lucas, Ben Juurlink
Publié dans: 2017 IEEE International Parallel and Distributed Processing Symposium (IPDPS), Numéro 2017, 2017, Page(s) 1119-1128, ISBN 978-1-5386-3914-6
Éditeur: IEEE
DOI: 10.1109/IPDPS.2017.101

VComputeBench: A Vulkan Benchmark Suite for GPGPU on Mobile and Embedded GPUs

Auteurs: Nadjib Mammeri, Ben Juurlink
Publié dans: Proceedings 2018 IEEE International Symposium on Workload Characterization (IISWC), Numéro 2018, 2018
Éditeur: IEEE
DOI: 10.14279/depositonce-7346

SLC: Memory Access Granularity Aware Selective Lossy Compression for GPUs

Auteurs: Sohan Lal, Jan Lucas, Ben Juurlink
Publié dans: 2019 Design, Automation & Test in Europe Conference & Exhibition (DATE), Numéro 2019, 2019
Éditeur: IEEE

Peer reviewed articles (1)

Highly parallel HEVC decoding for heterogeneous systems with CPU and GPU

Auteurs: Biao Wang, Diego Felix de Souza, Mauricio Alvarez-Mesa, Chi Ching Chi, Ben Juurlink, Aleksandar Ilić, Nuno Roma, Leonel Sousa
Publié dans: Signal Processing: Image Communication, Numéro 62, 2018, Page(s) 93-105, ISSN 0923-5965
Éditeur: Elsevier BV
DOI: 10.1016/j.image.2017.12.009

Other (1)

VENDOR-AGNOSTIC TOOLS FOR ASSESSING GPU PERFORMANCE/POWER

Auteurs: Georgios Keramidas, Graham Mudd, Andrew Richards, Ben Juurlink
Publié dans: HiPEAC Info, Numéro 49, 2017, Page(s) p27
Éditeur: HiPEAC