February 06, 2013
SANTA CLARA, Calif., Feb. 5 – NVIDIA today announced a new report describing dramatic performance increases for math libraries accelerated using NVIDIA CUDA, the world's most pervasive parallel computing platform and programming model for accelerating scientific and engineering applications on NVIDIA GPUs.
Available as a free download on the NVIDIA Developer Zone website, the report details the ground-breaking application performance increases enabled by CUDA 5 GPU-accelerated math libraries running on the new NVIDIA Tesla K20 GPU accelerators. These include cuFFT, cuBLAS, cuSPARSE, cuRAND, the NVIDIA Performance Primitives (NPP) library, Thrust and a C99-compatible math library, all of which are available for free in the latest CUDA 5 Toolkit release.
Providing application developers with the easiest and fastest way to add GPU acceleration to their code, the CUDA libraries deliver dramatic performance advantages compared to Intel's Math Kernel Library (MKL) routines running on the latest Intel CPUs. Performance highlights from the report include:
NVIDIA is hosting a webinar on Wednesday, Feb. 6, from 10 – 11 a.m. PT to provide more information about CUDA 5 math library performance. Register for the webinar on the GTC Express website.
Hands-on Labs @ GTC
NVIDIA will also provide developers with a unique opportunity to experience the dramatic performance increases delivered by GPU-accelerated libraries in a series of hands-on labs at the fourth-annual GPU Technology Conference (GTC).
At GTC 2013, which will be held in San Jose, Calif., from March 18-21, developers will have the opportunity to expand their programming skills with CUDA C/C++ and OpenACC. They can also experience the performance advantages of the new CUDA 5 GPU-accelerated libraries using their own laptops to virtually access powerful GPU-based workstations hosted by Amazon in the cloud. To register for the conference, visit the GTC 2013 registration page. Once registered, attendees can log in to reserve a seat for the hands-on labs.
The CUDA 5 platform makes the development of GPU-accelerated applications faster and easier than ever, including support for dynamic parallelism, GPU-callable libraries, NVIDIA GPUDirect technology support for RDMA (remote direct memory access), and the NVIDIA Nsight Eclipse Edition integrated development environment (IDE). To learn more about CUDA or download the latest version, visit the CUDA website.
CUDA is a parallel computing platform and programming model developed by NVIDIA. It enables dramatic increases in computing performance by harnessing the power of GPUs. With more than 1.6 million downloads, supporting more than 180 leading engineering, scientific and commercial applications, the CUDA programming model is the most popular way for developers to take advantage of GPU-accelerated computing.
NVIDIA awakened the world to computer graphics when it invented the GPU in 1999. Today, its processors power a broad range of products from smartphones to supercomputers. NVIDIA's mobile processors are used in cell phones, tablets and auto infotainment systems. PC gamers rely on GPUs to enjoy spectacularly immersive worlds. Professionals use them to create 3D graphics and visual effects in movies and to design everything from golf clubs to jumbo jets. And researchers utilize GPUs to advance the frontiers of science with high performance computing. The company has more than 5,000 patents issued, allowed or filed, including ones covering ideas essential to modern computing. For more information, see www.nvidia.com.
Source: NVIDIA Corp.
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?