VISUAL QUALITY ASSESSMENT CHALLENGES FOR ARCHITECTURE DESIGN EXPLORATIONS. Wen-Fu Kao and Durgaprasad Bilagi. Intel Corporation Folsom, CA 95630

Size: px
Start display at page:

Download "VISUAL QUALITY ASSESSMENT CHALLENGES FOR ARCHITECTURE DESIGN EXPLORATIONS. Wen-Fu Kao and Durgaprasad Bilagi. Intel Corporation Folsom, CA 95630"

Transcription

1 Proceedings of Seventh International Workshop on Video Processing and Quality Metrics for Consumer Electronics January 30-February 1, 2013, Scottsdale, Arizona VISUAL QUALITY ASSESSMENT CHALLENGES FOR ARCHITECTURE DESIGN EXPLORATIONS Wen-Fu Kao and Durgaprasad Bilagi Intel Corporation Folsom, CA ABSTRACT Typical computer architecture design primarily addresses performance and power tradeoff. In modern GPU design, more and more video and imaging features are integrated as part of the GPU. Media architecture design requires image/video quality as the third domain and the tradeoff among them. Assessment of video quality plays a crucial role in design space exploration. Although many popular evaluation methods and various quality metrics have been proposed, conducting cross-platform comparison as well as architecture optimization and tuning continues to pose great challenge during product design. We present realworld case studies in designing Intel processor graphics architecture on how we apply known quality metrics in the experiments and studies to form appropriate interpretation and guide us on the optimization on the actual silicon. 1. INTRODUCTION Video quality assessment is fundamental to evaluate any codec or imaging algorithm design. Many research works investigated and proposed subjective and objective metrics and methodologies in order to provide better understanding about the visual quality. In typical computer architecture design, substantial amount of effort spent on the tradeoff of performance and power. When more and more video and imaging features are integrated as part of the graphics processor, it is essential to consider performance, power, and visual quality during the design exploration to achieve the optimal decision under the design constraints. Typically, the visual quality evaluation is completed in early algorithmic exploration, and the performance/power tradeoff is evaluated during the design implementation. However, the need to bring all three elements together to optimize the design is crucial to deliver the right product. There are many challenges to overcome for a better and fair assessment in order to achieve this goal. In this paper, we visit these challenges through real case studies we conducted in the lab to understand the implications of these challenges when designing processor graphics. 2. RELATED WORKS Visual quality evaluation is a highly complex problem due to two reasons as mentioned in [3]. One is the video system that includes capture & display hardware, converters, networks, codecs all affect the quality. The second one is solutions that can match how human visual system perceives video & its quality. Many papers describe metrics to assess video quality. [1][2] reviews many metrics and classify them as data metrics, picture metrics, packet metrics, and hybrid metrics. They can also be classified as No-Reference (NR), Reduced-Reference (RR), and Full-Reference (FR). The FR metric evaluates the content respect to the original reference. The NR metrics only analyze the test content without any reference available. The RR uses an alternative channel between senders and receivers to transmit the parameters of the delivered content in a reduced way or use watermarking schemes. In [4], it is discussed that traditional data metrics such as PSNR, SSIM, MSE, MOS etc are still widely used, but significant improvements in prediction performance and versatility can only be achieved by perceptual quality metrics. In our studies, these difficulties directly challenge us on how to better combine various metrics and approaches in order to integrate perceptual quality experience and quantifiable target. 3. OVERVIEW OF INTEL PGFX First, Figure 1 shows Intel s latest processor graphics (pgfx) design, code name Ivy Bridge. It depicts a highlevel function diagram. As the graphics is part of the CPU/GPU integration on the same die, there is an interface to communicate with shared last-level cache (LLC) and memory. Media functions include a highperformance multi-format codec engine, and a media sampler that handles heavy computation tasks and high quality imaging enhancement. The array of execution 27 VPQM2013

2 units (EU) is the primary interface to control and work with media sampler. All media tasks can benefit from local L3 cache and the shared LLC. Another key component in design exploration is to understand other competitive products on the market and their design choice. The information greatly contributes to shape the design strategy. Although there are many software solutions that provide excellent quality, it might not be feasible for realworld application if the computation cost is too high. For example, HD video calling application on mobile devices would benefit from using power efficient hardware solution. However, simply implementing the solution directly in the hardware is often not enough. It requires careful design to balance between performance and power. Having the right quality assessment is the key to understand the necessary quality level under the design constraints. Enabling these new applications also guide us on deciding between fixed function unit design and generic shader unit computation. provide improved video and imaging experience on the PC and mobile devices. More and more users directly watch videos and process images on these devices. Enabling these features come with a cost. Many of these features are compute-intensive operations. For high performance PC with power source, turning on these features is not a problem. When PC is operated on battery or users are using personal mobile devices that have very limited thermal budget, it is not trivial to decide when to turn these features on. We explore each feature s performance and power efficiency under various configurations from high performance product to ultra low power product in order to define the best possible combination to enable or disable features. The screenshots in Figure 2 are both Intel pgfx and competition product operate with AC power source. The second screenshots in Figure 3 are both machines operate on battery. Apparently, the competition product on the right hand side will need to choose to disable important image enhancement features such as noise reduction in exchange for smooth video playback under the thermal and power budget. Consequently, the image quality will need to be compromised. Figure 2. Screenshots on AC power (Right: Competition, Left:Intel) Figure 1. High-level function diagram of Intel pgfx 4. CASE STUDIES We present several case studies to help understand how we explore the design space and optimize HW/SW solution for multiple market segments. Objective metrics such as SSIM and PSNR are still our primary metrics whenever we can measure and subjective quality assessment is still required in many of the studies when measuring objective metric is not feasible. Figure 3. Screenshots on battery (Right: Competition, Left: Intel) 4.1. Image enhancement quality on battery There are many image enhancement features included in Intel s pgfx such as noise reduction, color enhancement, skin tone correction, etc. These features are targeted to 4.2. H.264 encode tuning 28

3 This is a major task that is fundamentally challenging in delivering a good hardware encode solution. In early phase of the codec design, we greatly rely on C-based model to implement the algorithm and evaluate the quality. The computation cost in the C model is generally based on analytical model. The actual performance and power optimization will need to be conducted after the silicon comes back from the fab. There are many different aspects and goals in the optimization and tuning effort. We are choosing one of the scenarios as an example in this paper. The popular x264 is currently one of the high quality H.264 encoder available today. Generally, software encoder has the greatest flexibility to run complicated algorithm because the CPU is the most general-purpose processor. HW encoder has its design limitation in order to be small, fast, and power efficient. Studying the quality difference between x264 and Intel hardware encoder can still guide us on the areas where we should improve such as different rate control policy, different motion estimation behavior and mode decisions based on the contents, etc. In Figure 4, the chart from MSU s 2012 video codec comparison report shows that using a laptop PC with weaker CPU and basic integrated processor graphics, Intel s hardware encoder is better in terms of speed and quality tradeoff than best pure software encoder x264 at very high-speed encoding [5]. The green dot at the lower right corner represents baseline 1.0 from x264. The red dot at the upper left corner indicates that Intel hardware encoder takes about 5% less bits to achieve the same quality at nearly 14% faster speed. In order to conduct such quality study, immediately, we face the miscorrelation between highlevel C model and actual silicon driver implementation. Therefore, the quality obtained from the actual system does not always represent the real quality from our encoder design. To determine where to spend the tuning effort, we have to investigate what causes miscorrelation. If the quality gap truly comes from the algorithm design, it is reasonable to optimize the algorithm. However, if the goal of the tuning is to optimize performance, power, and quality, we can only rely on the real system for true performance and power measurement.. The tuning is time consuming since separate effort to validate and correlate between C model and the real driver needs to be made in parallel. The next challenge is what configurations of the encoder to use. Ideally, comparing the quality difference under total controlled environment and exact matching configuration gives us the best understanding on how the encode behavior is. However, even both x264 and Intel hardware encoder allow a wide range of parameters to reconfigure, it is not possible to have a perfect match naturally. We generally need to settle with best possible match. In addition, we have the need to compare the two encoders based on their default and stock settings in order to understand general users experience. Another challenge is to bring user experience to the tuning effort. PSNR continues to be used heavily during evaluation. The norm is also shifting to rely more on SSIM or similar metric to better reflect the visual quality. To evaluate a tuning option, we cannot just target to obtain higher PSNR or SSIM, We need to be able to realize the tuning through subjective user experience in addition to quantify with the objective metric. Courtesy: MSU Graphics & Media Lab video Group Figure 4. Intel HW encoder vs. x264 performance/quality tradeoff 4.3. Video calling In this case study, we face a different challenge. We are evaluating better hardware encode solution directly when customers have taken current hardware encoder to provide their own solution. Microsoft s Skype and Apple s FaceTime are two popular video calling applications. In the recent update, Apple has enabled hardware acceleration in their FaceTime HD on the Macs running on the latest Intel s pgfx. Comparing the two totally different applications with different approaches is difficult but it is essential to understand how other people utilize our hardware. We have to rely on subjective quality assessment with several indirect metrics such as network bandwidth, CPU and GPU utilization, and package and platform power consumption in order to form an appropriate study. There are additional considerations when we evaluate potential optimizations after we see the result from third-party applications. A work-in-progress option is compared with a software encoder. When we look at the screenshots in Figure 5 and Figure 6, right hand side is Intel s hardware encoder from reference C model and left hand side is a software encoder of interest. The reference C model tells us that our hardware encoder did a fairly good job in preserving some of the detail when we inspected these frames of interest. However, we cannot draw conclusion too early when we do not know what the quality would be on the real silicon. Another question is 29

4 the difference may be spotted when performing frame-byframe inspection, but how often users would spot this and be bothered during real-time video calling. Video calling is unique because it is not just about tuning the encode quality. It is a platform solution that requires us to carefully balance the performance, power, and quality. While market trend pushes further to enable high quality HD 1080p video calling, the market also pushes longer battery life at the same time. To find the balance of reasonable bit rate, better bit-rate control approach, and low power encode solution is the key to deliver a practical HD video calling platform solution. difference. In both Figure 7 and Figure 8, the right-most screenshot is Intel pgfx, and the other two are other hardware encoders. From the zoom-in screenshots, users can easily observe a better visual quality from Intel s hardware encoder. The chart in Figure 9 also shows superior performance advantage over other hardware solutions that Intel s hardware encoder delivers significantly faster than real-time transcode speed. Figure 6. Screenshots of HW1 (left), HW2 (middle), and Intel pgfx (right) Figure 7. Zoome-in screenshots of HW1 (left), HW2 (middle), and Intel pgfx (right) Figure 5. Exmaple screenshots (Left: SW, Right: Intel) Figure 6. Exmaple screenshots (Left: SW, Right: Intel) 4.4. Cross-platform quality comparison Figure 9. Transcode speed comparison A popular type of study that is often seen in technical review is product review. Most companies also conduct their own competitive studies in order to better understand and shape strategy. In one of our studies, we target other hardware encoder solution to compare with Intel pgfx. The primary challenge is each hardware encoder solution is tailor made and not all of them have the same level of freedom to reconfigure. The best-case scenario is finding a cross-platform application that supports many different hardware venders so we can minimize the difference. Nevertheless, each hardware encoder is based on its unique design and philosophy on balancing between fixed function unit and generic compute shader, we are still comparing total different configurations even using the same application. We again rely more on subjective quality assessment and inspect the video frame to understand the 5. FUTURE WORKS We presented quality assessment challenges during the architecture exploration. The gap between C model reference and real silicon driver itself will be a huge topic in the validation area. However, integrating C model directly into cycle-accurate architecture simulator with real-world workload will greatly help early architecture exploration. It can provide a better early indicator the tradeoff among performance, power, and quality. Next challenge is image enhancement feature evaluation such as noise reduction, image stabilization, frame rate conversion, skin tone correction, etc. These features aim to improve the visual quality of the original image. Existing metrics are designed to evaluate the match of the original. New approach and metric are needed to assess 30

5 the quality. Finally, addressing the real user visual experience is equally important to the objective metrics. Integrating subjective assessment into objective metrics to form a realistic benchmark to represent user experience will ultimately improve and guide media architecture design. 6. REFERENCES [1] R.Serral-Graci`a et al, An Overview of Quality of Experience Measurement - Challenges for Video Applications in IP Networks [2] Stefan Winkler, Video Quality Measurement Standards Current Status and Trends, ICICS 2009 [3] S. Winkler: Digital Video Quality Vision Models and Metrics. John Wiley & Sons, [4] Stefan Winkler, Video quality and beyond. Proc. European Signal Processing Conference, Poznan, Poland, Sept. 3-7, 2007 [5] Dmitriy Vatolin et al, MPEG-4 Video Codecs Comparison 31

NOT FOR DISTRIBUTION OR REPRODUCTION

NOT FOR DISTRIBUTION OR REPRODUCTION www.pipelinepub.com Volume 10, Issue 11 Next-Generation Video Transcoding By Alexandru Voica The Emergence of H.265 (HEVC) and 10- Bit Color Formats Today s increasingly demanding applications, such as

More information

Intel Quick Sync Video Tec h nology on Intel Iris Graphics and Intel HD Gra p hics family Flexible Transcode Performance and Quality

Intel Quick Sync Video Tec h nology on Intel Iris Graphics and Intel HD Gra p hics family Flexible Transcode Performance and Quality WHITE PAPER Intel Quick Sync Video Intel Quick Sync Video Tec h nology on Intel Iris Graphics and Intel HD Gra p hics family Flexible Transcode Performance and Quality The 4th generation Intel Core Processor

More information

Intel Iris Graphics Quick Sync Video Innovation Behind Quality and Performance Leadership

Intel Iris Graphics Quick Sync Video Innovation Behind Quality and Performance Leadership Intel Iris Graphics Quick Sync Video Innovation Behind Quality and Performance Leadership Dr. Wen-Fu Kao and Dr. Ryan Lei Intel Visual & Parallel Group Media Architecture Legal INFORMATION IN THIS DOCUMENT

More information

Blind Prediction of Natural Video Quality and H.264 Applications

Blind Prediction of Natural Video Quality and H.264 Applications Proceedings of Seventh International Workshop on Video Processing and Quality Metrics for Consumer Electronics January 30-February 1, 2013, Scottsdale, Arizona 1 Blind Prediction of Natural Video Quality

More information

The Case for Content-Adaptive Optimization

The Case for Content-Adaptive Optimization The Case for Content-Adaptive Optimization Whitepaper In today's digital world, video consumers are more demanding than ever before. Congested networks and technical challenges that content owners face

More information

MiraVision Picture Quality Enhancement Technology for Displays MediaTek White Paper January 2015

MiraVision Picture Quality Enhancement Technology for Displays MediaTek White Paper January 2015 Picture Quality Enhancement Technology for Displays MediaTek White Paper January 2015 2015 MediaTek Inc. 1 The Total Solution to Picture Quality Enhancement In multi media technology the display interface

More information

Building an Area-optimized Multi-format Video Encoder IP. Tomi Jalonen VP Sales

Building an Area-optimized Multi-format Video Encoder IP. Tomi Jalonen VP Sales Building an Area-optimized Multi-format Video Encoder IP Tomi Jalonen VP Sales www.allegrodvt.com Allegro DVT Founded in 2003 Privately owned, based in Grenoble (France) Two product lines: 1) Industry

More information

ORBX 2 Technical Introduction. October 2013

ORBX 2 Technical Introduction. October 2013 ORBX 2 Technical Introduction October 2013 Summary The ORBX 2 video codec is a next generation video codec designed specifically to fulfill the requirements for low latency real time video streaming. It

More information

Voice Analysis for Mobile Networks

Voice Analysis for Mobile Networks White Paper VIAVI Solutions Voice Analysis for Mobile Networks Audio Quality Scoring Principals for Voice Quality of experience analysis for voice... 3 Correlating MOS ratings to network quality of service...

More information

Image Quality Assessment Techniques: An Overview

Image Quality Assessment Techniques: An Overview Image Quality Assessment Techniques: An Overview Shruti Sonawane A. M. Deshpande Department of E&TC Department of E&TC TSSM s BSCOER, Pune, TSSM s BSCOER, Pune, Pune University, Maharashtra, India Pune

More information

Enabling a Richer Multimedia Experience with GPU Compute. Roberto Mijat Visual Computing Marketing Manager

Enabling a Richer Multimedia Experience with GPU Compute. Roberto Mijat Visual Computing Marketing Manager Enabling a Richer Multimedia Experience with GPU Compute Roberto Mijat Visual Computing Marketing Manager 1 What is GPU Compute Operating System and most application processing continue to reside on the

More information

How Many Humans Does it Take to Judge Video Quality?

How Many Humans Does it Take to Judge Video Quality? How Many Humans Does it Take to Judge Video Quality? Bill Reckwerdt, CTO Video Clarity, Inc. Version 1.0 A Video Clarity Case Study page 1 of 5 Abstract for Subjective Video Quality Assessment In order

More information

There are two lights in the scene: one infinite (directional) light, and one spotlight casting from the lighthouse.

There are two lights in the scene: one infinite (directional) light, and one spotlight casting from the lighthouse. Sample Tweaker Ocean Fog Overview This paper will discuss how we successfully optimized an existing graphics demo, named Ocean Fog, for our latest processors with Intel Integrated Graphics. We achieved

More information

The Benefits of GPU Compute on ARM Mali GPUs

The Benefits of GPU Compute on ARM Mali GPUs The Benefits of GPU Compute on ARM Mali GPUs Tim Hartley 1 SEMICON Europa 2014 ARM Introduction World leading semiconductor IP Founded in 1990 1060 processor licenses sold to more than 350 companies >

More information

Crusoe Processor Benchmark Report

Crusoe Processor Benchmark Report Crusoe Processor Benchmark Report Platform Benchmark Results Transmeta Corporation February 3, 2000 Property of: Transmeta Corporation 3940 Freedom Circle Santa Clara, CA 95054 USA (408) 919-3000 http://www.transmeta.com

More information

Advanced Encoding Features of the Sencore TXS Transcoder

Advanced Encoding Features of the Sencore TXS Transcoder Advanced Encoding Features of the Sencore TXS Transcoder White Paper November 2011 Page 1 (11) www.sencore.com 1.605.978.4600 Revision 1.0 Document Revision History Date Version Description Author 11/7/2011

More information

ECE 172 Digital Systems. Chapter 15 Turbo Boost Technology. Herbert G. Mayer, PSU Status 8/13/2018

ECE 172 Digital Systems. Chapter 15 Turbo Boost Technology. Herbert G. Mayer, PSU Status 8/13/2018 ECE 172 Digital Systems Chapter 15 Turbo Boost Technology Herbert G. Mayer, PSU Status 8/13/2018 1 Syllabus l Introduction l Speedup Parameters l Definitions l Turbo Boost l Turbo Boost, Actual Performance

More information

HEVC/H.265 Video Codecs Comparison

HEVC/H.265 Video Codecs Comparison HEVC/H.265 Video Codecs Comparison Video group head: Project head: Measurements & analysis: Dr. Dmitriy Vatolin Dr. Dmitriy Kulikov Mikhail Erofeev, Stanislav Dolganov, Sergey Zvezdakov Codecs: H.265 Intel

More information

Performance Analysis and Culling Algorithms

Performance Analysis and Culling Algorithms Performance Analysis and Culling Algorithms Michael Doggett Department of Computer Science Lund University 2009 Tomas Akenine-Möller and Michael Doggett 1 Assignment 2 Sign up for Pluto labs on the web

More information

Quantifying Trends in Server Power Usage

Quantifying Trends in Server Power Usage Quantifying Trends in Server Power Usage Richard Gimarc CA Technologies Richard.Gimarc@ca.com October 13, 215 215 CA Technologies. All rights reserved. What are we going to talk about? Are today s servers

More information

The Edge: Delivering the Quality of Experience of Digital Content

The Edge: Delivering the Quality of Experience of Digital Content The Edge: Delivering the Quality of Experience of Digital Content 2016 EDITION By Conviva for EdgeConneX As video consumption evolves from single screen to multi-screen, the burden on the Internet and

More information

SDACCEL DEVELOPMENT ENVIRONMENT. The Xilinx SDAccel Development Environment. Bringing The Best Performance/Watt to the Data Center

SDACCEL DEVELOPMENT ENVIRONMENT. The Xilinx SDAccel Development Environment. Bringing The Best Performance/Watt to the Data Center SDAccel Environment The Xilinx SDAccel Development Environment Bringing The Best Performance/Watt to the Data Center Introduction Data center operators constantly seek more server performance. Currently

More information

Evaluation Report: Improving SQL Server Database Performance with Dot Hill AssuredSAN 4824 Flash Upgrades

Evaluation Report: Improving SQL Server Database Performance with Dot Hill AssuredSAN 4824 Flash Upgrades Evaluation Report: Improving SQL Server Database Performance with Dot Hill AssuredSAN 4824 Flash Upgrades Evaluation report prepared under contract with Dot Hill August 2015 Executive Summary Solid state

More information

THE STATE OF IT TRANSFORMATION FOR RETAIL

THE STATE OF IT TRANSFORMATION FOR RETAIL THE STATE OF IT TRANSFORMATION FOR RETAIL An Analysis by Dell EMC and VMware Dell EMC and VMware are helping IT groups at retail organizations transform to business-focused service providers. The State

More information

HEVC/H.265 Video Codecs Comparison

HEVC/H.265 Video Codecs Comparison HEVC/H.265 Video Codecs Comparison 4K Appendix Video group head: Project head: Measurements & analysis: Dr. Dmitriy Vatolin Dr. Dmitriy Kulikov Mikhail Erofeev, Stanislav Dolganov Codecs: H.265 Intel MSS

More information

Building a Brighter Future. Your Guide to EISA-Compliant Ballast and Lamp Solutions

Building a Brighter Future. Your Guide to EISA-Compliant Ballast and Lamp Solutions Building a Brighter Future Your Guide to EISA-Compliant Ballast and Lamp Solutions EISA Compliance Made Easy The Energy Independence and Security Act (EISA) goes into effect on January 1, 2009 and promises

More information

3D Graphics in Future Mobile Devices. Steve Steele, ARM

3D Graphics in Future Mobile Devices. Steve Steele, ARM 3D Graphics in Future Mobile Devices Steve Steele, ARM Market Trends Mobile Computing Market Growth Volume in millions Mobile Computing Market Trends 1600 Smart Mobile Device Shipments (Smartphones and

More information

White Paper. Visual Processing in Rugged Environments. By Joe Lin, General Manager, Industrial Computing Solutions.

White Paper. Visual Processing in Rugged Environments. By Joe Lin, General Manager, Industrial Computing Solutions. White Paper Visual Processing in Rugged Environments By Joe Lin, General Manager, Industrial Computing Solutions www.nexcom.com Complex visual processing is becoming an important part of many rugged industrial

More information

Measuring VDI Fitness and User Experience Technical White Paper

Measuring VDI Fitness and User Experience Technical White Paper Measuring VDI Fitness and User Experience Technical White Paper 3600 Mansell Road Suite 200 Alpharetta, GA 30022 866.914.9665 main 678.397.0339 fax info@liquidwarelabs.com www.liquidwarelabs.com Table

More information

The Past, Present and Future of USB

The Past, Present and Future of USB The Past, Present and Future of USB Introduction When the humble universal serial bus (USB) was first introduced, it took the electronics industry a few years to get on board, but now there are few connected

More information

NVMe SSDs Becoming Norm for All Flash Storage

NVMe SSDs Becoming Norm for All Flash Storage SSDs Becoming Norm for All Flash Storage Storage media has improved by leaps and bounds over the last several years. Capacity and performance are both improving at rather rapid rates as popular vendors

More information

Multi-threading technology and the challenges of meeting performance and power consumption demands for mobile applications

Multi-threading technology and the challenges of meeting performance and power consumption demands for mobile applications Multi-threading technology and the challenges of meeting performance and power consumption demands for mobile applications September 2013 Navigating between ever-higher performance targets and strict limits

More information

Unleashing the benefits of GPU Computing with ARM Mali TM Practical applications and use-cases. Steve Steele, ARM

Unleashing the benefits of GPU Computing with ARM Mali TM Practical applications and use-cases. Steve Steele, ARM Unleashing the benefits of GPU Computing with ARM Mali TM Practical applications and use-cases Steve Steele, ARM 1 Today s Computational Challenges Trends Growing display sizes and resolutions, richer

More information

Choosing a Processor: Benchmarks and Beyond (S043)

Choosing a Processor: Benchmarks and Beyond (S043) Insight, Analysis, and Advice on Signal Processing Technology Choosing a Processor: Benchmarks and Beyond (S043) Jeff Bier Berkeley Design Technology, Inc. Berkeley, California USA +1 (510) 665-1600 info@bdti.com

More information

Intro to: Ultra-low power, ultra-high bandwidth density SiP interconnects

Intro to: Ultra-low power, ultra-high bandwidth density SiP interconnects This work was supported in part by DARPA under contract HR0011-08-9-0001. The views, opinions, and/or findings contained in this article/presentation are those of the author/presenter

More information

Using Alluxio to Improve the Performance and Consistency of HDFS Clusters

Using Alluxio to Improve the Performance and Consistency of HDFS Clusters ARTICLE Using Alluxio to Improve the Performance and Consistency of HDFS Clusters Calvin Jia Software Engineer at Alluxio Learn how Alluxio is used in clusters with co-located compute and storage to improve

More information

High Performance Video Artifact Detection Enhanced with CUDA. Atul Ravindran Digimetrics

High Performance Video Artifact Detection Enhanced with CUDA. Atul Ravindran Digimetrics High Performance Video Artifact Detection Enhanced with CUDA Atul Ravindran Digimetrics Goals & Challenges Provide automated QC for digital video files with accuracy and minimum false positives Provide

More information

BANDWIDTH REDUCTION SCHEMES FOR MPEG-2 TO H.264 TRANSCODER DESIGN

BANDWIDTH REDUCTION SCHEMES FOR MPEG-2 TO H.264 TRANSCODER DESIGN BANDWIDTH REDUCTION SCHEMES FOR MPEG- TO H. TRANSCODER DESIGN Xianghui Wei, Wenqi You, Guifen Tian, Yan Zhuang, Takeshi Ikenaga, Satoshi Goto Graduate School of Information, Production and Systems, Waseda

More information

LINEAR VIDEO DELIVERY FROM THE CLOUD. A New Paradigm for 24/7 Broadcasting WHITE PAPER

LINEAR VIDEO DELIVERY FROM THE CLOUD. A New Paradigm for 24/7 Broadcasting WHITE PAPER WHITE PAPER LINEAR VIDEO DELIVERY FROM THE CLOUD A New Paradigm for 24/7 Broadcasting Copyright 2016 Elemental Technologies. Linear Video Delivery from the Cloud 1 CONTENTS Introduction... 3 A New Way

More information

ENSC 427 COMMUNICATION NETWORKS

ENSC 427 COMMUNICATION NETWORKS ENSC 427 COMMUNICATION NETWORKS Implementation of VoIP over a Campus Wide Network Spring 2010 FINAL PROJECT Xuan Lu Kevan Thompson Zhiyu Zhou http://www.sfu.ca/~zzhou/project.html xla9@sfu.ca kjthomps@sfu.ca

More information

A Capacity Planning Methodology for Distributed E-Commerce Applications

A Capacity Planning Methodology for Distributed E-Commerce Applications A Capacity Planning Methodology for Distributed E-Commerce Applications I. Introduction Most of today s e-commerce environments are based on distributed, multi-tiered, component-based architectures. The

More information

Intra-Mode Indexed Nonuniform Quantization Parameter Matrices in AVC/H.264

Intra-Mode Indexed Nonuniform Quantization Parameter Matrices in AVC/H.264 Intra-Mode Indexed Nonuniform Quantization Parameter Matrices in AVC/H.264 Jing Hu and Jerry D. Gibson Department of Electrical and Computer Engineering University of California, Santa Barbara, California

More information

Maximize automotive simulation productivity with ANSYS HPC and NVIDIA GPUs

Maximize automotive simulation productivity with ANSYS HPC and NVIDIA GPUs Presented at the 2014 ANSYS Regional Conference- Detroit, June 5, 2014 Maximize automotive simulation productivity with ANSYS HPC and NVIDIA GPUs Bhushan Desam, Ph.D. NVIDIA Corporation 1 NVIDIA Enterprise

More information

The Future of 3D Graphics NVIDIA

The Future of 3D Graphics NVIDIA The Future of 3D Graphics GPUs are Great at Graphics Crysis 2006 Crytek / Electronic Arts Hellgate: London 2005-2006 Flagship Studios, Inc. Licensed by NAMCO BANDAI Games America, Inc. Full Spectrum Warrior:

More information

The Different Types of UPS Systems

The Different Types of UPS Systems The Different Types of UPS Systems White Paper # 1 Revision 4 Executive Summary There is much confusion in the marketplace about the different types of UPS systems and their characteristics. Each of these

More information

2010 Intel Core processor family (Intel Core i3/i5/i7)

2010 Intel Core processor family (Intel Core i3/i5/i7) CLIENT GRAPHICS 2 nd Generation Intel Core now with BuiltIn Visuals, Available on Select Models of the 2 nd Generation Intel Core Family Built-In Visuals Built for Mainstream Desktop and Mobile PC Users

More information

An Implementation of Multiple Region-Of-Interest Models in H.264/AVC

An Implementation of Multiple Region-Of-Interest Models in H.264/AVC An Implementation of Multiple Region-Of-Interest Models in H.264/AVC Sebastiaan Van Leuven 1, Kris Van Schevensteen 1, Tim Dams 1, and Peter Schelkens 2 1 University College of Antwerp Paardenmarkt 92,

More information

Silicon Motion s Graphics Display SoCs

Silicon Motion s Graphics Display SoCs WHITE PAPER Silicon Motion s Graphics Display SoCs Enable 4K High Definition and Low Power Power and bandwidth: the twin challenges of implementing a solution for bridging any computer to any high-definition

More information

White paper ETERNUS Extreme Cache Performance and Use

White paper ETERNUS Extreme Cache Performance and Use White paper ETERNUS Extreme Cache Performance and Use The Extreme Cache feature provides the ETERNUS DX500 S3 and DX600 S3 Storage Arrays with an effective flash based performance accelerator for regions

More information

Encoding At Netflix Media Meetup, May 2014 Netflix Digital Supply Chain

Encoding At Netflix Media Meetup, May 2014 Netflix Digital Supply Chain Encoding At Netflix Media Meetup, May 2014 Netflix Digital Supply Chain 1 Introduction Netflix started streaming in 2007 with IE-plugin hosting Windows Media Player and ~2000 titles Today Award-winning

More information

FPGA Co-Processing Architectures for Video Compression

FPGA Co-Processing Architectures for Video Compression Co-Processing Architectures for Compression Overview Alex Soohoo Altera Corporation 101 Innovation Drive San Jose, CA 95054, USA (408) 544-8063 asoohoo@altera.com The push to roll out high definition video

More information

Apple ProRes RAW. White Paper April 2018

Apple ProRes RAW. White Paper April 2018 Apple ProRes RAW White Paper April 2018 Contents Introduction 3 About Raw Video 4 Data Rate 6 Performance 8 Using ProRes RAW in Final Cut Pro 12 Using Log Conversion with Built-in Camera LUTs 13 Using

More information

Collecting OpenCL*-related Metrics with Intel Graphics Performance Analyzers

Collecting OpenCL*-related Metrics with Intel Graphics Performance Analyzers Collecting OpenCL*-related Metrics with Intel Graphics Performance Analyzers Collecting Important OpenCL*-related Metrics with Intel GPA System Analyzer Introduction Intel SDK for OpenCL* Applications

More information

QOE ISSUES RELEVANT TO VIDEO STREAMING IN CABLE NETWORKS Jeremy Bennington Praveen Mohandas. Cheetah Technologies, Sunnyvale, CA

QOE ISSUES RELEVANT TO VIDEO STREAMING IN CABLE NETWORKS Jeremy Bennington Praveen Mohandas. Cheetah Technologies, Sunnyvale, CA QOE ISSUES RELEVANT TO VIDEO STREAMING IN CABLE NETWORKS Jeremy Bennington Praveen Mohandas Cheetah Technologies, Sunnyvale, CA Abstract This paper explores how monitoring video quality in a streaming

More information

STUDY AND IMPLEMENTATION OF VIDEO COMPRESSION STANDARDS (H.264/AVC, DIRAC)

STUDY AND IMPLEMENTATION OF VIDEO COMPRESSION STANDARDS (H.264/AVC, DIRAC) STUDY AND IMPLEMENTATION OF VIDEO COMPRESSION STANDARDS (H.264/AVC, DIRAC) EE 5359-Multimedia Processing Spring 2012 Dr. K.R Rao By: Sumedha Phatak(1000731131) OBJECTIVE A study, implementation and comparison

More information

IBM POWER SYSTEMS: YOUR UNFAIR ADVANTAGE

IBM POWER SYSTEMS: YOUR UNFAIR ADVANTAGE IBM POWER SYSTEMS: YOUR UNFAIR ADVANTAGE Choosing IT infrastructure is a crucial decision, and the right choice will position your organization for success. IBM Power Systems provides an innovative platform

More information

International Journal of Emerging Technology and Advanced Engineering Website: (ISSN , Volume 2, Issue 4, April 2012)

International Journal of Emerging Technology and Advanced Engineering Website:   (ISSN , Volume 2, Issue 4, April 2012) A Technical Analysis Towards Digital Video Compression Rutika Joshi 1, Rajesh Rai 2, Rajesh Nema 3 1 Student, Electronics and Communication Department, NIIST College, Bhopal, 2,3 Prof., Electronics and

More information

IMR-Pathload: Robust Available Bandwidth Estimation under End-Host Interrupt Delay

IMR-Pathload: Robust Available Bandwidth Estimation under End-Host Interrupt Delay IMR-Pathload: Robust Available Bandwidth Estimation under End-Host Interrupt Delay Seong Kang Joint work with Dmitri Loguinov Internet Research Lab Department of Computer Science Texas A&M University,

More information

MAXIMIZING BANDWIDTH EFFICIENCY

MAXIMIZING BANDWIDTH EFFICIENCY MAXIMIZING BANDWIDTH EFFICIENCY Benefits of Mezzanine Encoding Rev PA1 Ericsson AB 2016 1 (19) 1 Motivation 1.1 Consumption of Available Bandwidth Pressure on available fiber bandwidth continues to outpace

More information

Video. Add / edit video

Video. Add / edit video Video Videos greatly support learning in an e-earning setting. It is a rather complex topic though, due to the variety of formats, codecs, containers and combination of operating systems and browsers.

More information

Request for Comments: 1787 T.J. Watson Research Center, IBM Corp. Category: Informational April 1995

Request for Comments: 1787 T.J. Watson Research Center, IBM Corp. Category: Informational April 1995 Network Working Group Y. Rekhter Request for Comments: 1787 T.J. Watson Research Center, IBM Corp. Category: Informational April 1995 Status of this Memo Routing in a Multi-provider Internet This memo

More information

Metaheuristic Development Methodology. Fall 2009 Instructor: Dr. Masoud Yaghini

Metaheuristic Development Methodology. Fall 2009 Instructor: Dr. Masoud Yaghini Metaheuristic Development Methodology Fall 2009 Instructor: Dr. Masoud Yaghini Phases and Steps Phases and Steps Phase 1: Understanding Problem Step 1: State the Problem Step 2: Review of Existing Solution

More information

NVIDIA GRID APPLICATION SIZING FOR AUTODESK REVIT 2016

NVIDIA GRID APPLICATION SIZING FOR AUTODESK REVIT 2016 NVIDIA GRID APPLICATION SIZING FOR AUTODESK REVIT 2016 BPG-08489-001 March 2017 Best Practices Guide TABLE OF CONTENTS Users Per Server (UPS)... 1 Technology Overview... 3 Autodesk Revit 2016 Application...

More information

Challenges to Embedding Computer Vision J. Scott Gardner General Manager and Editor-in-Chief Embedded Vision Alliance (www.embedded-vision.

Challenges to Embedding Computer Vision J. Scott Gardner General Manager and Editor-in-Chief Embedded Vision Alliance (www.embedded-vision. Challenges to Embedding Computer Vision J. Scott Gardner General Manager and Editor-in-Chief Embedded Vision Alliance (www.embedded-vision.com) May 16, 2011 Figure 1 HAL 9000 a machine that sees. Source:

More information

Quality versus Intelligibility: Evaluating the Coding Trade-offs for American Sign Language Video

Quality versus Intelligibility: Evaluating the Coding Trade-offs for American Sign Language Video Quality versus Intelligibility: Evaluating the Coding Trade-offs for American Sign Language Video Frank Ciaramello, Jung Ko, Sheila Hemami School of Electrical and Computer Engineering Cornell University,

More information

QUALITY OF EXPERIENCE IN INTERNET TELEVISION. Norwegian University of Science and Technology (NTNU), Trondheim, Norway

QUALITY OF EXPERIENCE IN INTERNET TELEVISION. Norwegian University of Science and Technology (NTNU), Trondheim, Norway QUALITY OF EXPERIENCE IN INTERNET TELEVISION Mathias Gjerstad Lervold (1), Liyuan Xing (2), Andrew Perkis (2) (1) Accenture (2) Centre for Quantifiable Quality of Service in Communication Systems (Q2S)

More information

Review and Implementation of DWT based Scalable Video Coding with Scalable Motion Coding.

Review and Implementation of DWT based Scalable Video Coding with Scalable Motion Coding. Project Title: Review and Implementation of DWT based Scalable Video Coding with Scalable Motion Coding. Midterm Report CS 584 Multimedia Communications Submitted by: Syed Jawwad Bukhari 2004-03-0028 About

More information

ECE 8823: GPU Architectures. Objectives

ECE 8823: GPU Architectures. Objectives ECE 8823: GPU Architectures Introduction 1 Objectives Distinguishing features of GPUs vs. CPUs Major drivers in the evolution of general purpose GPUs (GPGPUs) 2 1 Chapter 1 Chapter 2: 2.2, 2.3 Reading

More information

Experimental Calibration and Validation of a Speed Scaling Simulator

Experimental Calibration and Validation of a Speed Scaling Simulator IEEE MASCOTS 2016 Experimental Calibration and Validation of a Speed Scaling Simulator Arsham Skrenes Carey Williamson Department of Computer Science University of Calgary Speed Scaling: Inherent Tradeoffs

More information

ECE 637 Integrated VLSI Circuits. Introduction. Introduction EE141

ECE 637 Integrated VLSI Circuits. Introduction. Introduction EE141 ECE 637 Integrated VLSI Circuits Introduction EE141 1 Introduction Course Details Instructor Mohab Anis; manis@vlsi.uwaterloo.ca Text Digital Integrated Circuits, Jan Rabaey, Prentice Hall, 2 nd edition

More information

TR An Overview of NVIDIA Tegra K1 Architecture. Ang Li, Radu Serban, Dan Negrut

TR An Overview of NVIDIA Tegra K1 Architecture. Ang Li, Radu Serban, Dan Negrut TR-2014-17 An Overview of NVIDIA Tegra K1 Architecture Ang Li, Radu Serban, Dan Negrut November 20, 2014 Abstract This paperwork gives an overview of NVIDIA s Jetson TK1 Development Kit and its Tegra K1

More information

Why Real Testing Requires Emulation, Not Just Simulation for Layer 4-7

Why Real Testing Requires Emulation, Not Just Simulation for Layer 4-7 White Paper Why Real Testing Requires Emulation, Not Just Simulation for Layer 4-7 Testing for the real world is about emulating the entire network, not just individual switches and routers. This includes

More information

AP Physics 1 and 2 Summer Assignment

AP Physics 1 and 2 Summer Assignment AP Physics 1 and 2 Summer Assignment Due: First Day of Class Welcome to AP Physics! You are responsible for the material covered in the first three chapters of your textbook. The questions that follow

More information

Cursor Design Considerations For the Pointer-based Television

Cursor Design Considerations For the Pointer-based Television Hillcrest Labs Design Note Cursor Design Considerations For the Pointer-based Television Designing the cursor for a pointing-based television must consider factors that differ from the implementation of

More information

Lightweight Arithmetic for Mobile Multimedia Devices. IEEE Transactions on Multimedia

Lightweight Arithmetic for Mobile Multimedia Devices. IEEE Transactions on Multimedia Lightweight Arithmetic for Mobile Multimedia Devices Tsuhan Chen Carnegie Mellon University tsuhan@cmu.edu Thanks to Fang Fang and Rob Rutenbar IEEE Transactions on Multimedia EDICS Signal Processing for

More information

Configuration changes such as conversion from a single instance to RAC, ASM, etc.

Configuration changes such as conversion from a single instance to RAC, ASM, etc. Today, enterprises have to make sizeable investments in hardware and software to roll out infrastructure changes. For example, a data center may have an initiative to move databases to a low cost computing

More information

MARACAS: A Real-Time Multicore VCPU Scheduling Framework

MARACAS: A Real-Time Multicore VCPU Scheduling Framework : A Real-Time Framework Computer Science Department Boston University Overview 1 2 3 4 5 6 7 Motivation platforms are gaining popularity in embedded and real-time systems concurrent workload support less

More information

QoE-Driven Video Streaming and Video Content Caching

QoE-Driven Video Streaming and Video Content Caching CommNet2 & IcoreJoint Workshop on Content Caching & Distributed Storage for Future Communication Networks QoE-Driven Video Streaming and Video Content Caching Xiaohong Peng Adaptive Communications Networks

More information

Perceptual Quality Measurement and Control: Definition, Application and Performance

Perceptual Quality Measurement and Control: Definition, Application and Performance Perceptual Quality Measurement and Control: Definition, Application and Performance A. R. Prasad, R. Esmailzadeh, S. Winkler, T. Ihara, B. Rohani, B. Pinguet and M. Capel Genista Corporation Tokyo, Japan

More information

PERFORMANCE STUDY OCTOBER 2017 ORACLE MONSTER VIRTUAL MACHINE PERFORMANCE. VMware vsphere 6.5

PERFORMANCE STUDY OCTOBER 2017 ORACLE MONSTER VIRTUAL MACHINE PERFORMANCE. VMware vsphere 6.5 PERFORMANCE STUDY OCTOBER 2017 ORACLE MONSTER VIRTUAL MACHINE PERFORMANCE VMware vsphere 6.5 Table of Contents Executive Summary...3 Introduction...3 Test Environment... 4 Test Workload... 5 Virtual Machine

More information

Optimizing Tiered Storage Workloads with Precise for Storage Tiering

Optimizing Tiered Storage Workloads with Precise for Storage Tiering Applied Technology Abstract By introducing Enterprise Flash Drives to EMC Symmetrix and CLARiiON storage systems, EMC revitalized the importance of tiered storage. Optimizing the use of these tiers provides

More information

Race to Exascale: Opportunities and Challenges. Avinash Sodani, Ph.D. Chief Architect MIC Processor Intel Corporation

Race to Exascale: Opportunities and Challenges. Avinash Sodani, Ph.D. Chief Architect MIC Processor Intel Corporation Race to Exascale: Opportunities and Challenges Avinash Sodani, Ph.D. Chief Architect MIC Processor Intel Corporation Exascale Goal: 1-ExaFlops (10 18 ) within 20 MW by 2018 1 ZFlops 100 EFlops 10 EFlops

More information

Query-Sensitive Similarity Measure for Content-Based Image Retrieval

Query-Sensitive Similarity Measure for Content-Based Image Retrieval Query-Sensitive Similarity Measure for Content-Based Image Retrieval Zhi-Hua Zhou Hong-Bin Dai National Laboratory for Novel Software Technology Nanjing University, Nanjing 2193, China {zhouzh, daihb}@lamda.nju.edu.cn

More information

Google Workloads for Consumer Devices: Mitigating Data Movement Bottlenecks Amirali Boroumand

Google Workloads for Consumer Devices: Mitigating Data Movement Bottlenecks Amirali Boroumand Google Workloads for Consumer Devices: Mitigating Data Movement Bottlenecks Amirali Boroumand Saugata Ghose, Youngsok Kim, Rachata Ausavarungnirun, Eric Shiu, Rahul Thakur, Daehyun Kim, Aki Kuusela, Allan

More information

Business Objects Performance Testing

Business Objects Performance Testing Business Objects Performance Testing EXECUTIVE SUMMARY Performance testing informs decisions about the allocation of resources. Performance testing software like Apache JMeter and Micro Focus LoadRunner

More information

It s Time for Mass Scale VDI Adoption

It s Time for Mass Scale VDI Adoption It s Time for Mass Scale VDI Adoption Cost-of-Performance Matters Doug Rainbolt, Alacritech Santa Clara, CA 1 Agenda Intro to Alacritech Business Motivation for VDI Adoption Constraint: Performance and

More information

Understanding the performance of an X user environment

Understanding the performance of an X user environment Understanding the performance of an X550 11-user environment Overview NComputing s desktop virtualization technology enables significantly lower computing costs by letting multiple users share a single

More information

A Comparison of Still-Image Compression Standards Using Different Image Quality Metrics and Proposed Methods for Improving Lossy Image Quality

A Comparison of Still-Image Compression Standards Using Different Image Quality Metrics and Proposed Methods for Improving Lossy Image Quality A Comparison of Still-Image Compression Standards Using Different Image Quality Metrics and Proposed Methods for Improving Lossy Image Quality Multidimensional DSP Literature Survey Eric Heinen 3/21/08

More information

All Programmable: from Silicon to System

All Programmable: from Silicon to System All Programmable: from Silicon to System Ivo Bolsens, Senior Vice President & CTO Page 1 Moore s Law: The Technology Pipeline Page 2 Industry Debates Variability Page 3 Industry Debates on Cost Page 4

More information

Advanced Computer Architecture (CS620)

Advanced Computer Architecture (CS620) Advanced Computer Architecture (CS620) Background: Good understanding of computer organization (eg.cs220), basic computer architecture (eg.cs221) and knowledge of probability, statistics and modeling (eg.cs433).

More information

Recommended Readings

Recommended Readings Lecture 11: Media Adaptation Scalable Coding, Dealing with Errors Some slides, images were from http://ip.hhi.de/imagecom_g1/savce/index.htm and John G. Apostolopoulos http://www.mit.edu/~6.344/spring2004

More information

Background Heterogeneous Architectures Performance Modeling Single Core Performance Profiling Multicore Performance Estimation Test Cases Multicore

Background Heterogeneous Architectures Performance Modeling Single Core Performance Profiling Multicore Performance Estimation Test Cases Multicore By Dan Stafford Background Heterogeneous Architectures Performance Modeling Single Core Performance Profiling Multicore Performance Estimation Test Cases Multicore Design Space Results & Observations General

More information

by Mahender Reddy Concept To Reality / Summer 2006

by Mahender Reddy Concept To Reality / Summer 2006 by Mahender Reddy Demand for higher extrusion rates, increased product quality and lower energy consumption have prompted plants to use various methods to determine optimum process conditions and die designs.

More information

(ii) Why are we going to multi-core chips to find performance? Because we have to.

(ii) Why are we going to multi-core chips to find performance? Because we have to. CSE 30321 Computer Architecture I Fall 2009 Lab 06 Introduction to Multi-core Processors and Parallel Programming Assigned: November 3, 2009 Due: November 17, 2009 1. Introduction: This lab will introduce

More information

Native playback H264 Video freezes on high spec pro system Posted by Julia - 03 Jul :27

Native playback H264 Video freezes on high spec pro system Posted by Julia - 03 Jul :27 Native playback H264 Video freezes on high spec pro system Posted by Julia - 03 Jul 2012 07:27 I would like to be able to view 1080p video from my gopro Hero 2 within Lightworks without transcoding. After

More information

Mali GPU acceleration of HEVC and VP9 Decoder

Mali GPU acceleration of HEVC and VP9 Decoder Mali GPU acceleration of HEVC and VP9 Decoder 2 Web Video continues to grow!!! Video accounted for 50% of the mobile traffic in 2012 - Citrix ByteMobile's 4Q 2012 Analytics Report. Globally, IP video traffic

More information

WHITE PAPER ON2 TECHNOLOGIES, INC. TrueMotion VP7 Video Codec. January 10, 2005 Document Version: 1.0

WHITE PAPER ON2 TECHNOLOGIES, INC. TrueMotion VP7 Video Codec. January 10, 2005 Document Version: 1.0 ON2 TECHNOLOGIES, INC. WHITE PAPER TrueMotion VP7 Video Codec January 10, 2005 Document Version: 1.0 On2 Technologies, Inc. 21 Corporate Drive Suite 103 Clifton Park, NY 12065 www.on2.com Contents Introduction...

More information

Load Dynamix Enterprise 5.2

Load Dynamix Enterprise 5.2 DATASHEET Load Dynamix Enterprise 5.2 Storage performance analytics for comprehensive workload insight Load DynamiX Enterprise software is the industry s only automated workload acquisition, workload analysis,

More information

Adaptive Bit Rate (ABR) Video Detection and Control

Adaptive Bit Rate (ABR) Video Detection and Control OVERVIEW Adaptive Bit Rate (ABR) Video Detection and Control In recent years, Internet traffic has changed dramatically and this has impacted service providers and their ability to manage network traffic.

More information

SDR Forum Technical Conference 2007

SDR Forum Technical Conference 2007 THE APPLICATION OF A NOVEL ADAPTIVE DYNAMIC VOLTAGE SCALING SCHEME TO SOFTWARE DEFINED RADIO Craig Dolwin (Toshiba Research Europe Ltd, Bristol, UK, craig.dolwin@toshiba-trel.com) ABSTRACT This paper presents

More information