BEYOND STANDARDS:
A NEW GPU-AWARE IMAGE CODING SYSTEM
Pablo Enfedaque
BEYOND STANDARDS: A NEW GPU-AWARE IMAGE CODING SYSTEM Pablo - - PowerPoint PPT Presentation
BEYOND STANDARDS: A NEW GPU-AWARE IMAGE CODING SYSTEM Pablo Enfedaque GPU Technology Conference April 4 - 7, 2016 Ph.D. Advisors Francesc Aul-Llins Juan C. Moure Universitat Autnoma de Barcelona Table of contents 4. RESULTS PROBLEM
A NEW GPU-AWARE IMAGE CODING SYSTEM
Pablo Enfedaque
April 4 - 7, 2016
Universitat Autònoma de Barcelona
Ph.D. Advisors
Francesc Aulí-Llinàs Juan C. Moure
4.
CONCLUSIONS
RESULTS
SOLUTION
PROBLEM
+ Enhanced compression + Error resilience + Region-of-interest coding + Interactive transmission + Progressive lossy-to-lossless Computational complexity
Image Coding Evolution
Now 1990 Image data JPEG2000
80% Computational time
JPEG2000 Coding Pipeline
Data Transformation
Compressed file
BPC Arithmetic Coder
Bitstream Reorganization
Data Transformation
Bitplane Coder
Bitplane Coder
JPEG2000
(BPC with Parallel Coefficient processing)
Arithmetic coder
N Arithmetic coders
BPC-PaCo
N Stripes
The Proposal
generation
New mechanisms
generation
New mechanisms BPC-PaCo
CUDA implementation
Coding performance
GTX TITAN X 8x Intel Xeon E5-4620 (64 HW threads)
GPU
Decoder
CPU
Encoder
BPC-PaCo GPU JPEG2K Kakadu
Computational performance
5 satellite images 10240 x 10240 8 bit grey scale
BPC-PaCo GPU JPEG2K Kakadu
Power Efficiency
GTX TITAN X Tegra X1
CPU
JPEG2000 complexity BPC-PaCo parallelism 2% coding performance
x25 SpeedUp
Contact
Pablo Enfedaque pablo.enfedaque@gmail.com LinkedIn Francesc Aulí-Llinàs fauli@deic.uab.es Juan C. Moure juancarlos.moure@uab.cat
Publications
Strategies of SIMD computing for image coding in GPU IEEE International Conference on High Performance Computing (HiPC) 2016 Bitplane Image Coding With Parallel Coefficient Processing IEEE Transactions on Image Processing 2016 Strategy of microscopic parallelism for bitplane image coding IEEE Data Compression Conference (DCC) 2015 Implementation of the DWT in a GPU through a Register-based Strategy IEEE Transactions on Parallel & Distributed Systems 2014
About us
http://gici.uab.es/GiciWebPage/ http://grupsderecerca.uab.cat/hpca4se/en/content/gpu