SC10 New Orleans Louisiana

Khronos Group demonstrates significant OpenCL momentum at SC10

Visit Khronos at SC10 Booth # 1132

OpenCL is the open, royalty-free standard for general purpose parallel programming across CPUs, GPUs and other processors. OpenCL provides software developers portable and efficient access to the full power of a wide range of systems including high-performance compute servers, desktop computer systems and handheld devices.

The Khronos Group is pleased to be participating at SC10 with a trade show booth, A BOF and a Tutorial. Visit the Khronos booth and sessions at SC10 and pick up your free copy of a laminated OpenCL Quick Reference card. This comprehensive six page quick reference covers all important aspects of the OpenCL API. You can also download a PDF copy of the OpenCL Quick Reference Card today.

BOF: OpenCL – Achievements and Next Steps

Date: Wednesday, 17 November 2010 5:30pm – 7:00pm
Room: 384-385

Abstract:

Heterogeneous computing with CPUs, GPUs and other processors has arrived. It seems everyone is experimenting with it. To move from “experimenting with” to “routine production computing platform” we need a rich software ecosystem. And that means we need a standard software platform so programmers can “write once to compile and run anywhere.” OpenCL is that standard. In this session we will talk about what’s new with OpenCL. We will also have a panel of experts discuss controversial and engaging issues pertaining to OpenCL. But the highlight of this BOF will be the annual OpenCL awards for the most surprising, most portable and best performing applications.

Session Leader Details: Tim Mattson (Primary Session Leader) - Khronos Group

OpenCL Sessions

Time	Type	Presentation	Speaker	Room
Monday November 15 8:30AM - 12:00PM	Tutorial	OpenCL: An Introduction to Heterogeneous Programming for HPC OpenCL is a standard for programming heterogeneous computers built from CPUs, GPUs and other processors. It includes a framework to define the platform in terms of a host (e.g. a CPU) and one or more compute devices (e.g. a GPU) plus a C-based programming language for writing programs for the compute devices. Using OpenCL, a programmer can write task-based and data-parallel programs that use all the resources of the heterogeneous computer. In this tutorial, we will introduce OpenCL. This will be a “programmer’s introduction” where we cover the ideas behind OpenCL but also show how these ideas are translated into source code. We will do this through a series of progressively more challenging examples … thereby providing a pedagogically useful examples that experienced HPC programmers will need to quickly become productive OpenCL programmers.	Benedict Gaster AMD Tim Mattson Intel Corp	391-392
Monday November 15 1:30PM - 5:00PM	Tutorial	Advanced Topics in Heterogeneous Programming with OpenCL OpenCL is a standard for programming heterogeneous computers built from CPUs, GPUs and other processors. In this tutorial, we will explore OpenCL in depth. We will briefly summarize the key features in OpenCL and then present a series of case studies showing how to use OpenCL. The focus will be on source code so we show not just want OpenCL can do, but how you can implement similar functionality in your own programs. We assume attendees have either worked with OpenCL before or have just completed our “Introduction to OpenCL” tutorial.	Benedict Gaster AMD Tim Mattson Intel Corp Ian Buck NVIDIA Peng Wang NVIDIA Mike Houston AMD	391-392
Tuesday November 16 1:30PM - 2:00PM	Exhibitor Forum	GPU Programming in OpenCL with CLIDE: Microway's Compute Language IDE Microway's CLIDE software makes it easy to develop and test algorithms on GPU and many-core/socket CPU Nodes using OpenCL. Microway’s GCS™ service enables customers to use CLIDE to share compute kernels, evaluate and track kernel performance history on Microway benchmark nodes. Features demonstrated include: 1) A large set of useful examples 2) Syntax highlighting and automatic statement completion 3) Integrated context-sensitive help and robust error handling GCS enables customers to benchmark OpenCL compute kernels against standard data sets to determine: 1) Which compute device or devices (CPU, GPU, APU) give the best performance running a kernel on data sets? 2) How do specific optimizations of kernels affect their performance on different devices? 3) How much power does the kernel consume running on different devices? 4) How much device memory is required to get optimal throughput on a large data size?	Michael B. Fried Microway	280-281

Time

Type

Presentation

Speaker

Room

Monday November 15
8:30AM - 12:00PM

Tutorial

OpenCL: An Introduction to Heterogeneous Programming for HPC
OpenCL is a standard for programming heterogeneous computers built from CPUs, GPUs and other processors. It includes a framework to define the platform in terms of a host (e.g. a CPU) and one or more compute devices (e.g. a GPU) plus a C-based programming language for writing programs for the compute devices. Using OpenCL, a programmer can write task-based and data-parallel programs that use all the resources of the heterogeneous computer. In this tutorial, we will introduce OpenCL. This will be a “programmer’s introduction” where we cover the ideas behind OpenCL but also show how these ideas are translated into source code. We will do this through a series of progressively more challenging examples … thereby providing a pedagogically useful examples that experienced HPC programmers will need to quickly become productive OpenCL programmers.

Benedict Gaster AMD
Tim Mattson Intel Corp

391-392

Monday November 15
1:30PM - 5:00PM

Tutorial

Advanced Topics in Heterogeneous Programming with OpenCL
OpenCL is a standard for programming heterogeneous computers built from CPUs, GPUs and other processors. In this tutorial, we will explore OpenCL in depth. We will briefly summarize the key features in OpenCL and then present a series of case studies showing how to use OpenCL. The focus will be on source code so we show not just want OpenCL can do, but how you can implement similar functionality in your own programs. We assume attendees have either worked with OpenCL before or have just completed our “Introduction to OpenCL” tutorial.

Benedict Gaster AMD
Tim Mattson Intel Corp
Ian Buck NVIDIA
Peng Wang NVIDIA
Mike Houston AMD

391-392

Tuesday November 16
1:30PM - 2:00PM

Exhibitor Forum

GPU Programming in OpenCL with CLIDE: Microway's Compute Language IDE
Microway's CLIDE software makes it easy to develop and test algorithms on GPU and many-core/socket CPU Nodes using OpenCL. Microway’s GCS™ service enables customers to use CLIDE to share compute kernels, evaluate and track kernel performance history on Microway benchmark nodes. Features demonstrated include: 1) A large set of useful examples 2) Syntax highlighting and automatic statement completion 3) Integrated context-sensitive help and robust error handling GCS enables customers to benchmark OpenCL compute kernels against standard data sets to determine: 1) Which compute device or devices (CPU, GPU, APU) give the best performance running a kernel on data sets? 2) How do specific optimizations of kernels affect their performance on different devices? 3) How much power does the kernel consume running on different devices? 4) How much device memory is required to get optimal throughput on a large data size?

Michael B. Fried Microway

280-281

Khronos members exhibiting at SC10

AMD Booth 3119 - Advanced Micro Devices (NYSE: AMD) is a leading global provider of innovative processing solutions in the computing and graphics markets. AMD is dedicated to driving open innovation, choice and industry growth by delivering superior customer-centric solutions that empower consumers and businesses worldwide.

Dell Booth 2412 - Dell is a premier provider of computer products and services required for customers worldwide to build their information-technology and Internet infrastructures. Dell designs, manufactures and customizes products and services to customer requirements, and offers an extensive selection of software and peripherals.

Fraunhofer Booth 1467 - The Fraunhofer Competence Center for HPC focuses on the development of highly parallel applications, the corresponding frameworks and tools to enable high scalability. The core technologies that are being developed at the Competence Center include the new Parallel File System, FhGFS, Cell-BE Solutions, RDMA-optimized Messaging Middleware, and Grid Computing Solutions.

Fujitsu Booth 3029 - Fujitsu has been leading HPC market over 30 years and is currently developing 10 Pflops class supercomputer with unique processor, interconnect and packaging technologies. We are also contributing to Japan's next generation supercomputer project led by RIKEN with these cutting-edge technologies. We provide a broad range of computing products such as x86-based cluster, SPARC-based SMP server, software and solutions to meet comprehensive HPC requirements.

IBM Booth 2539 - IBM is the world's leading provider of high performance computing systems and the largest information technology company. Building on decades of experience around the globe, IBM delivers a broad set of innovative, powerful solutions that help customers of any size address demands of intense computation, visualization and data manipulation. IBM strives to deliver greener products with leading energy efficiencies that help customers realize a competitive advantage.

Intel Booth 2813 - Intel, the world leader in silicon innovation, develops technologies, products and initiatives to continually advance how people work and live. Additional information about Intel is available at www.intel.com and blogs.intel.com.

Los Alamos National Laboratory Booth 4039 - Los Alamos National Laboratory, established in 1943, is a world leader in national security science. LANL directs its creativity toward solving the nation's most challenging scientific problems. LANL is home to Roadrunner, the first supercomputer to surpass a petaflop/s. LANL is operated by LANS, LLC for U.S. DOE/NNSA.

NEC Booth 2729 - NEC is the leading solutions provider for high performance computing. NEC builds highly efficient and powerful HPC systems with its advanced technology. The full range of HPC solutions includes SX series vector supercomputers, SMP scalar servers, cluster systems, high performance network and storage solutions. Teaming with partners and customers, NEC delivers world class large scale IT solutions.

NVIDIA Booth 1343 - NVIDIA® is the leader in visual computing technologies and the inventor of the graphics processing unit. NVIDIA serves the professional computing market with its Tesla™ solution with the world’s first teraflop many-core processor. NVIDIA Tesla computing systems deliver the necessary transition to energy efficient parallel computing power, solving the world’s most important computing challenges.

Sun Microsystems Booth 2529 -