Birds of a Feather Sessions



NVIDIA Tesla K40 and CUDA 6.0 Launch in India

Programming GPUs with OpenACC and OpenACC 2.0 Updates

Speaker(s): Brent Leback, PGI

CUDA 6.0 Updates

Speaker(s): Pradeep Kumar Gupta, NVIDIA

This session will give overview of NVIDIA CUDA 6.0 features, the latest version of the world’s most pervasive parallel computing platform and programming model. The CUDA 6.0 provides programmers with a robust, easy-to-use platform to develop advanced science, engineering and high performance computing (HPC) applications on x86 CPU and ARM based systems.

NVIDIA events at HiPC 2013

click here



High Performance Cloud

Speaker(s): Ashrut Ambastha

Mellanox experts will present on virtualization technologies for both InfiniBand and high-performance RDMA-enabled Ethernet clusters. The session’s focus will be on Single-Root I/O Virtualization with Mellanox’s RDMA-based HCAs/NICs to enable unprecedented Storage I/O and virtual machine performance. We will discuss various virtualized cloud platforms and technologies that can run the complete range of standard applications for the most demanding HPC jobs without changing the system software stack. To conclude, we will open the forum for any InfiniBand/Ethernet/Interconnect related discussions.



Computational Challenges in the Energy Sector

Speaker(s): Murthy Udupa, Leonardo Spanu, Aarthi Thyagarajan, Prakalp Somawanshi and Suhas Phadke

The energy sector has been one of the largest consumers of high performance computing right from the early days. However, we have always been, and currently still are in a situation where the problems we are attempting to solve outpace the capabilities of current generation hardware and software systems. In this session, we will present a few of the computational challenges that we currently face and are actively trying to solve from a variety of domains. These include investigating how to use first principles molecular dynamics to simulate atomic structures, smoothed particle hydrodynamics to model fluid-structure interactions, and subsurface imaging for seismic migration. These areas span multiple orders of magnitude of length and time scales, but face common computational challenges such as parallelization schemes, hardware architectures, I/O bottlenecks, and scalable algorithms that we will discuss in this session.



Performance Productivity Challenges & Future Research

Speaker(s): Victor Lee (Research Scientist, Intel Labs)

Xeon Phi Enabling

Speaker(s): Paresh Pattani (Director, Data Center Computing, Software & Services Group)

Intel Events at HiPC 2013

click here