An innovative hybrid parallelization using multiple GPUs and MPI dramatically reduces solution time for structural analysis and sensitivity calculations. Offloading the intensive matrix computation on the GPU and using heterogeneous computing improves performance. Users also benefit from accelerated access to compute resources "in the cloud."
*Note: This session was not recorded.
Learn about NVIDIA Maximus 2 and the latest NVIDIA Kepler GPU architecture that is enabling design and creative professionals to realize groundbreaking performance and productivity benefits through a complete transformation of their traditional workflows.
Cameras are everywhere, including the phone in your pocket. We first discuss an enabler for Computational Photography called FCam, an API for precise control of the camera that allows quick capture of images with deterministic control parameters so you know exactly what you get and can combine the results to new output photographs. FCam is currently available on Tegra 3 development boards. We then discuss OpenCV, a de-facto standard library for computer vision, which has been optimized to utilize Tegra's various processing elements such as the 4 ARM cores, the NEON SIMD co-processors, and the mobile GPU embedded inside the Tegra SoC.
The Operational Based Vision Assessment (OBVA) simulator was designed and built by NASA and the United States Air Force (USAF) to provide the Air Force School of Aerospace Medicine (USAFSAM) with a scientific testing laboratory to study human vision and testing standards in an operationally relevant environment. This talk describes the general design objectives and implementation characteristics of the simulator visual system recently delivered to meet real-time, man-in-the-loop scenarios.
An overview of how animation projects at ad agencies may differ from studio work. With spontaneous timelines, unpredictable budgets, and multiple clients coming at you from every angle "one size fits all" approach to workflow management and hardware configurations just won't do. Because in today's fast-paced ad agency, it always pays to keep things moving!
NUKE is a high end image compositing system that is the industry standard across the film industry. Jon Wadelton from the Foundry will give an overview how the power of the GPU will be harnessed for image processing in the next major release of NUKE due later this year. The new release will include GPU accelerated versions of algorithmically expensive image processing operations such as retiming, image denoising, and post-process depth of field effects.
This talk introduces a new approach for compositing shaders together, using an API-agnostic description of effects for objects materials and scene management (post-processing, management of rendering passes). This approach builds on the original concepts of NVIDIA CgFX and expands it to new levels of flexibility and extensibility. Rather than creating a new shading language, we show how to supersede existing ones (GLSL, HLSL) to avoid complex parsing and yet deploy the effect in a variety of environments. Because nvFX will be open-sourced, developers will benefit from the engineering approach presented and can leverage its runtime or extend it to service other specific custom requirements.
Starting with CATIA V6 R2011x, CATIA integrates iRay rendering technology within modeling environment through the CATIA Live Rendering Workbench. CATIA Live Rendering is a photorealistic rendering solution with interactive ray-tracing & Global illumination that complements CATIAs real-time rendering solution.
Learn how creative professionals harness the power of AdobeÆ After EffectsÆ CS6 and NVIDIA GPUs to accelerate the motion graphics workflow with new 3D ray-traced rendering. Based on NVIDIAÆ Optixô technology, this simplifies the design of realistic geometric text and shapes in 3D space with up to 27x faster performance on NVIDIAÆ QuadroÆ GPUs.
Convincing virtual product presentations significantly depend on the ability to reproduce the elaborate appearance of the materials the products are fabricated from. This talk outlines RTT's MLC project as a powerful, next generation rendering technology. It provides an improved support of digital materials by widening the set of supported material classes, handles material libraries in an enterprise environment, adds the ability to specify own material shaders easily and extends the rendering consistency between rendering modes, hardware platforms and future versions
Global illumination is an important element for realistic image synthesis, but its computation is expensive and highly dependent on the complexity of the scene and of the BRDF of the involved surfaces. Last year, we presented a new approach to compute indirect lighting in real-time that is based on a discrete voxel representation of the scene geometry and reflected radiance. In this approach, we proposed a new volumetric cone-tracing algorithm that is used to efficiently gather indirect lighting and evaluate indirect visibility, and takes advantage of a hierarchical sparse voxel octree data structure. In this talk, we will first present an overview of the technique and then describe the latest advances and results of these researches related to the fast voxelization and octree construction for dynamic scenes, the management of multiple light bounces as well as the efficient implementation on Kepler hardware.
A technical presentation on how to animate and render thousands of skinned characters interactively using the GPU for real-time visualization. Autodesk introduced the ìproject Geppettoî plugin to 3ds Max 2013 allowing to generate crowd flows to a model in order to populate it. The generated crowd motion files can be replayed like any animation. We collaborated with the Geppetto team in order to develop a real-time GPU accelerated viewer of the crowd animation using the same asset as the plugin. The characters are animated and rendered on the GPU, all the character graphics resource as well as the repertoire of motion clips are loaded on video memory. The input to each character animation is a motion path describing the path and the blending of motion clips at each frame. The algorithm performs the animation skeleton from this motion path which update the skeleton pose of the character and then render it performing skinning. In order to gain performances with the GPU, the characters are instanced so the steps of the algorithm can be performed efficiently in large batch on the GPU. We will detail the architecture of the real-time viewer and demonstrate its performances.
The goal of this session is to explain the major methods a workstation developer would use to mix and adopt OpenGL and CUDA architecture to build (or improve) a high performance workstation application. This talk/demo would walk through the major concepts that utilize the Quadro (and Tesla) hardware (Multi-GPU programming, OpenGL/CUDA interop, CUDA streaming, and dual copy engines). At the end of this session, the developer will have a solid understanding of ï key concepts behind multi-GPU programming (including maximus configurations) ï features of OpenGL and CUDA that support multi-GPU programming and in which development environments (ie linux vs windows OpenGL vs directX) ï differentiating features of high-end Quadro/Tesla boards like Dual Copy Engines and when to take advantage of them. ï how to get started with GPU Accelerated software development
More applications are using CUDA on large datasets to calculate results that can best be appreciated with some sort of visualization. This presents new problems in how the graphics API interacts with CUDA and how to extract maximum performance from the combination. NVIDIA Nsight is a suite of tools to help you tackle just this type of setup. Some of the topics we will cover include: Debugging applications that include CUDA and graphics work running on the GPU Profiling your application to determine where bottlenecks hide Tracing the use of all system resources for a holistic approach to performance tuning
Attend this session to get the most out of OpenGL on NVIDIA Quadro and GeForce GPUs. Topics include the latest advances in the OpenGL Shading Language (GLSL) and Cg 3.1; programmable tessellation; interoperability with Direct3D conventions; integration with Direct3D and CUDA resources; bindless graphics; OpenGL-centric compute, and more. Learn how your application can benefit from NVIDIA's leadership driving OpenGL as a cross-platform, open industry standard.
How many ways can you program a GPU? This tutorial covers four practical methods of adding GPU acceleration to your applications. Take advantage of pre-packaged acceleration by dropping in a GPU-accelerated library to replace MKL, IPP, FFTW, or other libraries you are already using. Automatically parallelize ëforí loops in your C or Fortran code using OpenACC directives. Use powerful, cross-platform algorithms and data structures from the Thrust library of C++ to target CUDA, TBB and OpenMP. Develop your own parallel applications and libraries using a programming language you already know like C, C++, Fortran and more.
The future of GPU-based visual computing integrates the web, resolution-independent 2D graphics, and 3D to maximize interactivity and quality while minimizing consumed power. See what NVIDIA is doing today to accelerate resolution-independent 2D graphics for web content. This presentation explains NVIDIA's unique "stencil, then cover" approach to accelerating path rendering with OpenGL and demonstrates the wide variety of web content that can be accelerated with this approach.
The new NVIDIA VGX technology allows for true hardware virtualization of the GPU, enabling a true PC and Workstation experience in a virtual desktop environment. This session will cover a comparison of graphics virtualization technologies available in the industry (both SW and HW methods) as well as accelerated remoting solutions.
OpenSubdiv by Pixar is a new open source library that implements high performance subdivision surface drawing and evaluation of modern GPU and massively parallel CPU architectures. Pixar will show OpenSubdiv drawing the Catmull-Clark subdivision surface of a sculpted, creased, textured, and animated character at interactive frame rates on Kepler hardware. They will discuss how CUDA and GLSL are leveraged in the OpenSubdiv architecture for best performance on Kepler.
Learn the latest approaches in levering GPUs for the fastest possible ray tracing results from experts developing and leveraging the NVIDIA OptiX ray tracing engine and those making custom renderers. Multiple ray tracing techniques, out-of-core rendering, multi-GPU support, optimal hardware configurations, and new opportunities with Kepler GPU and will be covered in this up to date discussion of the fastest growing trend in advanced rendering.