- For Attendees
- For Submitters
Industrial Talks Track
IND1 [Imagination Technologies]: PowerVR: Masterclass in Graphics Technology and Optimization
Tuesday, May 15th, 11:00-12:00 in Room B
Presenters: M. Weber (Developer Technology Engineer, Imagination Technologies)
Abstract: Imagination's PowerVR graphics IP is a crucial part of today's consumer electronics market, integrated into a wide variety of the latest smartphone, mobile console, computing and other products. Get the latest information on PowerVR graphics acceleration with practical tips, rules and guidelines to extract the best graphics performance. Discover the new PowerVR Series 6 Rogue architecture, PowerVR's unique Tile Based Deferred Rendering approach, the components of the cores, how they implement APIs like OpenGL ES 2.0 and the kind of techniques and 'gotchas' that take graphics from good to great.
IND2 [IBM and VARGROUP]: IBM solutions for High Performance Computing Cloud for EDA and CAE
Tuesday, May 15th, 14:00-15:00 in Room A
Presenters: Walter Bernocchi (Client Technical Architect– IBM STG Technical Computing Italy)
Abstract: The relevant technology progresses occurred in the last decade on microprocessors, networking, storage, cluster integration and middleware, qualify high performance computing capabilities a strategic asset for product designing at affordable costs. IBM provides advanced architectures to efficaciously integrate servers, storage and middleware, to address new needs on simulations for product development. Sizing, flexibility, reliability, resilience, availability, virtualization, TCO: criteria to be used to identify the best solution for an industrial organization which intends to invest in high performance computing. The exploitation of parallel computing for engineering applications in addition of an efficient data handling and 3D visualization, efficaciously managed by an user friendly resource scheduler, determines an optimal methodology for innovative product designing. Concepts as Cloud Computing and the specific IBM Engineering Computing framework implement the computing-on-demand service to a private, hybrid or public cloud infrastructure with the aim to optimize costs due to infrastructure itself and system management. This talk intends to give an overview of the architecture and technology that made the IBM cloud for High Performance Computing with example of implementations in the 2D and 3D spaces.
IND3 [NVIDIA]: From Raw Triangles to Beautiful Pixels
Wednesday, May 16th, 9:40-10:40 in Room A
Presenters: J. Pantaleoni (Senior Architect, NVIDIA)
Abstract: In this talk the presenter will give an overview of some of the technologies recently investigated by the graphics division of NVIDIA Research targeting all aspects of the graphics pipeline(s), from the out of core management of massive geometry sets to the the high performance conversion of raw triangles to screen samples, and from there to the final reconstruction of "beautiful pixels".
IND4 [Siemens Corporate Research]: State of the Art in GPU-Accelerated Techniques for Medical Imaging
Wednesday, May 16th, 15:20-16:20 in Room B
Presenters: G. Paladini (Program Manager, Imaging Architectures, Siemens Corporate Research)
Abstract: This talk will provide an overview of world-class medical imaging algorithms exploiting the immense computational power of modern GPUs. The range of techniques where GPUs have proven invaluable range from advanced visualization techniques such as real-time rendering of fused multi-modal datasets, time-varying volumetric ultrasound, GPU-based reconstruction, non-linear filtering, non-rigid registration, segmentation and flow simulation. Siemens Corporate Research conducts R&D for all Siemens Healthcare business units and has a large world-wide network of clinical and academic collaborators. We will present examples of a broad spectrum of clinical applications, both for post-processing as well as real-time interventional procedures, which have recently become feasible only thanks to advances in programmable GPUs.
IND5 [AMD]: AMD FirePro opens the Heterogeneous Systems Era for 3D Graphics and Compute
Thursday, May 17th, 11:00-12:00 in Room A
Presenters: B. Stefanizzi (AMD)
Abstract: AMD with the FirePro products is moving to a new Era of programming and processor performance for 3D graphics and compute. We will present the heterogeneous Systems era that new AMD processors are bringing with the new GCN based GPUs and the new APUs. We will show how together they represent a huge step forward for programming flexibility and performance efficiently for 3D graphics and compute.
IND6 [Siemens Corporate Research]: Volumetric processing and visualization on heterogeneous architectures
Thursday, May 17th, 14:00-15:00 in Room A
Presenters: G. Paladini (Program Manager, Imaging Architectures, Siemens Corporate Research), W. Li (Siemens Corporate Research), W. Hong (Siemens Corporate Research)
Abstract: Volumetric data is typically very large and requires intensive computation for processing and visualization. We have developed an OpenCL-based framework that can utilize all available resources in a system or a cluster of systems. The framework manages one or more OpenCL devices. A large volume is partitioned into bricks. Each OpenCL device is associated with a set of brick producers that generates the contents of bricks while optionally utilizing other bricks as input. The framework provides common brick producers and users, such as brick-loader, brick encoder/decoder, brick-deformer, brick-raycaster, and additional brick processors can be added. Each brick processing operation is then converted to a virtual instruction with the involved bricks as its operands. The framework is composed of a scheduler that distributes virtual brick instructions to different devices and as well as chooses an optimized processing order aiming at certain criteria. The scheduler customizes compiler optimizations such as register allocation and instruction scheduling, according to the characteristics of volumetric processing. Example application of the framework includes large volume rendering with advanced lighting effects, deformable and 4D volume visualization, integrated volume segmentation/registration with visualization, lattice-based flow simulation and visualization, etc.
IND7 [NICE]: Running 3D applications remotely in the "3D Cloud"
Thursday, May 17th, 16:00-17:00 in Room B
Presenters: A. Rodolico (CTO, NICE)
Abstract: Visualization of large volumes of data, stored or generated in remote data centers or cloud infrastructures, is an increasingly common requirement. Moving this data to the workstation is unpractical and time consuming and implies security risks and management headaches. In the last years IT departments took advantage of virtualization and remote access to overcome these problems, but such technologies are not broadly applied to 3D applications because the GPU cannot be effectively virtualized and accessed remotely with the proper performance. NICE Desktop Cloud Visualization (DCV) addresses these limitations leveraging the power of server-side GPUs, and delivering the hardware-accelerated 3D graphics over a standard TCP/IP network using dynamically adjustable image quality to allow users to experience fluid 3D interaction over LAN, WiFi and WAN, from standard PCs or mobile devices. The combination of virtualization and remote access allows off-the-shelf 3D applications to be consolidated in a Public or Private "3D Cloud", rendering the most complex models and effects with negligible overhead compared to native GPU performance, while leveraging the high speed network of the datacenter to work faster, leaving the data where it's best managed and performing.