SIGGRAPH 2014

August 12-14
Vancouver, BC Convention Center
Booth 1001

Join Intel and our partners from Adobe and DreamWorks Animation in the West Hall, Room 208 for 3 days of talks, discussions, demos and how-to’s that show you how you can BE INNOVATION.

Double booked with another talk you want to see? Don’t worry; most of our talks will be delivered twice! See the schedule below:

TUESDAY | WEDNESDAY | THURSDAY

Intel Fast Forward

Various Speakers

Tuesday, 12:00-1:30

Join us for a lunchtime series of rapid-fire mini presentations. During the hosted lunch session we will highlight academic papers, all of our technical sessions, booth demos, and other technologies. Each presentation will be 5 minutes long, so hang on, sit tight, and let’s see what you learn when you Fast Forward.


Accelerating Computer Vision, and More, with Hardware-Accelerated OpenCV on Heterogeneous Platforms

PDF Download

Aaron Kunze - Intel

Tuesday, 2:00 - 3:00

OpenCV, the leading open source library for computer vision, is releasing OpenCV 3.0 with a new architecture that enables transparent use of an OpenCL-enabled GPU to accelerate computer vision workloads, providing unique benefits to SoC-integrated GPU designs.

This session shows how, and which, OpenCV algorithms Intel is developing and optimizing that are being contributed to the open source version of OpenCV library.

During this session, attendees will learn:

  • Specific optimization techniques that are effective for Intel designs
  • Performance and power efficiency improvements already being delivered to computer vision applications.
  • Future challenges that the industry will have to tackle in order to continue to improve the efficiency of computer vision applications on heterogeneous systems.

Compute Architecture of Intel® Processor Graphics


A Practical Implementation for Scalable High-Performance Volumetric Effect Rendering From Low-Power Tablets to High-End Graphics Processors

PDF Download

Doug McNabb - Intel

Tuesday, 3:15- 4:15

Volumetric Effects are an important part of visual effects for film and games. Graphics hardware has improved to the point where film-like volumetric effects are possible in realtime. Several developers have recently shown advances in volumetric effects rendering techniques. We demonstrate a framework for efficiently rendering volumetric effects like smoke, clouds, and fire suitable for use across a range of platform capabilities – from low-power tablets to high-end graphics processors.


Writing Efficient IA Vector Code using ISPC

PDF Download

James Brodman - Intel

Tuesday, 4:30 - 5:30

ISPC, the Intel SPMD Program Compiler, is an open-source language and compiler that produces high quality code for Intel Vector Extensions including SSE, AVX, and Xeon Phi. ISPC provides programmers with a familiar interface augmented by simple language extensions that enable programmers to write a single code that efficiently executes on IA hardware ranging from Atom all the way to Xeon Phi. ISPC greatly improves programmer productivity by providing higher-level abstractions that are more portable than manually writing intrinsics. This talk will provide an introduction to writing programs using ISPC by presenting an overview of the language and compiler. A number of graphics benchmarks will be used to explain key features of the language and demonstrate the significant additional performance that can be obtained with ISPC.

TUESDAY | WEDNESDAY | THURSDAY

Fast Forward Clustered Shading

PDF Download

Marc Fauconneau - Intel

Wednesday, 9:00-10:00 - 30 minute double headers

Clustered shading is a recent addition to the light culling arsenal. In this session, we apply it to forward rendering with a large number of lights. The performance obtained is superior to tile-based deferred rendering thanks to increased culling accuracy. Key techniques used to quickly build the acceleration structure on the CPU will be explained.


The Intel® Xeon E3 Processor with Iris Pro Graphics – Ready to Serve in the Cloud

PDF Download

David Cowperthwaite & Scott Jackson - Intel

Wednesday, 9:00-10:00 - 30 minute double headers

Intel® Processor Graphics have made tremendous generational improvements. While adding features and performance our process leadership has allowed us to simultaneously drive down TDP thus enabling incredibly dense form factors. Server and cloud suppliers are now finding innovative ways to leverage these features in new markets and usage models. This session explores different processor graphics technologies and the ecosystems able to take advantage of graphics advances in the cloud.

In this session we cover:

  • Overview of Intel® Processor Graphics features and capabilities in the cloud
  • Intel® Processor Graphics usages including: Cloud Transcode, Remote Graphics, and Workstation

Writing Efficient IA Vector Code using ISPC

PDF Download

James Brodman - Intel

Wednesday, 10:45 - 11:45

ISPC, the Intel SPMD Program Compiler, is an open-source language and compiler that produces high quality code for Intel Vector Extensions including SSE, AVX, and Xeon Phi. ISPC provides programmers with a familiar interface augmented by simple language extensions that enable programmers to write a single code that efficiently executes on IA hardware ranging from Atom all the way to Xeon Phi. ISPC greatly improves programmer productivity by providing higher-level abstractions that are more portable than manually writing intrinsics. This talk will provide an introduction to writing programs using ISPC by presenting an overview of the language and compiler. A number of graphics benchmarks will be used to explain key features of the language and demonstrate the significant additional performance that can be obtained with ISPC.


Embree Ray Tracing Kernels: Tutorial and Application at Dreamworks Animation and in Autodesk Maya

Part 1 - PDF Download

Part 2 - PDF Download

Part 3 - PDF Download

Sven Woop, Louis Feng, Charles Congdon - Intel & Evan Smyth - DreamWorks Animation

Wednesday, 2:00-4:15

This two hour session starts with a tutorial of the Intel Embree ray tracing kernels (optimized for Intel® Xeon® and Intel® Xeon Phi™), detailing the C++ and vectorized ISPC (Intel SPMD Compiler) versions of the Embree API, the most important Embree features, and how to use these features to efficiently handle dynamic scenes, instance geometry, reduce memory consumption, and render special effects such as hair, shadows of transparent objects, motion blur, and analytically defined geometries.

Following this, we will present practical integrations of the Embree ray tracing kernels into DreamWorks Animation's “Light Transport Simulator (LTS)” interactive lighting preview tool (including a prerecorded demo showing a lighting artist using the Light Transport Simulator), and an interactive Autodesk Maya* viewport plugin.


Creativity to the Power of Go with Adobe* Creative Cloud* and Intel® Core Processors with Iris™ Graphics

Betty Leong, Dave Helmly - Adobe Systems

Wednesday, 4:30 - 5:30

Come see how Adobe* Photoshop* CC 2014 and Premiere* Pro CC 2014 are enriched with the 'Power of Go' on compact systems and laptops based on Intel® Core™ processors with integrated Iris™ graphics. Through collaboration with Intel, the latest Creative Cloud updates bring 'Go Power' to Photoshop features such as Content-Aware technology, Focus Mask, Intelligent Upsampling and Blur Gallery. Premiere Pro delivers 'Go' on Intel Iris Graphics in over 80 OpenCL-enabled effects and transitions and RED® raw video debayering.

TUESDAY | WEDNESDAY | THURSDAY

Accelerating Computer Vision, and More, with Hardware-Accelerated OpenCV on Heterogeneous Platforms

PDF Download

Aaron Kunze - Intel

Thursday, 9:00 - 10:00

OpenCV, the leading open source library for computer vision, is releasing OpenCV 3.0 with a new architecture that enables transparent use of an OpenCL-enabled GPU to accelerate computer vision workloads, providing unique benefits to SoC-integrated GPU designs.

This session shows how, and which, OpenCV algorithms Intel is developing and optimizing that are being contributed to the open source version of OpenCV library.

During this session, attendees will learn:

  • Specific optimization techniques that are effective for Intel designs
  • Performance and power efficiency improvements already being delivered to computer vision applications.
  • Future challenges that the industry will have to tackle in order to continue to improve the efficiency of computer vision applications on heterogeneous systems.

Compute Architecture of Intel® Processor Graphics


Creativity to the Power of Go with Adobe* Creative Cloud* and Intel® Core Processors with Iris™ Graphics

Betty Leong, Dave Helmly - Adobe Systems

Thursday, 10:45 - 11:45

Come see how Adobe* Photoshop* CC 2014 and Premiere* Pro CC 2014 are enriched with the 'Power of Go' on compact systems and laptops based on Intel® Core™ processors with integrated Iris™ graphics. Through collaboration with Intel, the latest Creative Cloud updates bring 'Go Power' to Photoshop features such as Content-Aware technology, Focus Mask, Intelligent Upsampling and Blur Gallery. Premiere Pro delivers 'Go' on Intel Iris Graphics in over 80 OpenCL-enabled effects and transitions and RED® raw video debayering.


A Practical Implementation for Scalable High-Performance Volumetric Effect Rendering From Low-Power Tablets to High-End Graphics Processors

PDF Download

Doug McNabb - Intel

Thursday, 2:00 -3:00

Volumetric Effects are an important part of visual effects for film and games. Graphics hardware has improved to the point where film-like volumetric effects are possible in realtime. Several developers have recently shown advances in volumetric effects rendering techniques. We demonstrate a framework for efficiently rendering volumetric effects like smoke, clouds, and fire suitable for use across a range of platform capabilities – from low-power tablets to high-end graphics processors.


Fast Forward Clustered Shading

PDF Download

Marc Fauconneau - Intel

Thursday, 3:15- 4:15 - 30 minute double headers

Clustered shading is a recent addition to the light culling arsenal. In this session, we apply it to forward rendering with a large number of lights. The performance obtained is superior to tile-based deferred rendering thanks to increased culling accuracy. Key techniques used to quickly build the acceleration structure on the CPU will be explained.


The Intel® Xeon E3 Processor with Iris Pro Graphics – Ready to Serve in the Cloud

PDF Download

David Cowperthwaite & Scott Jackson - Intel

Thursday, 3:15- 4:15 - 30 minute double headers

Intel® Processor Graphics have made tremendous generational improvements. While adding features and performance our process leadership has allowed us to simultaneously drive down TDP thus enabling incredibly dense form factors. Server and cloud suppliers are now finding innovative ways to leverage these features in new markets and usage models. This session explores different processor graphics technologies and the ecosystems able to take advantage of graphics advances in the cloud.

In this session we cover:

  • Overview of Intel® Processor Graphics features and capabilities in the cloud
  • Intel® Processor Graphics usages including: Cloud Transcode, Remote Graphics, and Workstation