All Resources

TotalView for HPC

Faster fault isolation, improved memory optimization, and dynamic visualization for your high performance computing apps. TotalView for HPC is a scalable and intuitive debugger for parallel applications written in C, C++, and Fortran that simplifies and shortens the process of developing, debugging, and optimizing complex applications. Purpose-built for multicore and parallel computing, TotalView for HPC delivers a set of tools providing unprecedented control over processes and thread execution, along with deep visibility into program states and data.

Continue Reading Here

Rogue Wave tools and libraries for big data

Big data applications are among the fastest-growing and demanding in the business and research communities, encompassing a range of workloads in real-time analytics, data mining, complex event processing, MapReduce applications, and visualization. Besides the need for high performance, big data applications must also handle data complexity, varying data properties, and optimizing storage capacity.

In this report, senior analyst Michael Feldman from Intersect360 Research discusses these challenges in the big data developer market and presents an analysis of how unique tools and libraries from Rogue Wave Software increase developer productivity and reduce deployment and maintenance costs.

Continue Reading Here

Rogue Wave tools and libraries for financial services

Financial services is the second largest vertical market of high performance computing, including high-frequency trading, risk management, securities and derivatives pricing, and economic analytics. With increased competition, a growing skills gap, and stronger regulations, it's more challenging than ever to deliver applications that meet demanding performance requirements.

In this report, senior analyst Michael Feldman from Intersect360 Research discusses the current financial services developer market and presents an analysis of how unique tools and libraries from Rogue Wave Software increase developer productivity and reduce deployment and maintenance costs.

Continue Reading Here

Introducing TotalView 8.14

Join Rogue Wave VP Scott Lasica as he provides an overview of TotalView 8.14, a graphical debugger specializing in HPC, supercomputing, and highly parallel debugging. Learn what’s new in the latest version, including support for CUDA 6.0, OpenACC, and Intel Xeon Phi.

Continue Reading Here

Tags: TotalView

TotalView CUDA

CUDA introduces developers to a number of new concepts (such as kernels, streams, warps, and explicitly multilevel memory) that are not encountered in serial or other parallel programming paradigms. Visibility into these elements is critical for troubleshooting and tuning applications that make use of CUDA. This paper will highlight CUDA concepts implemented in CUDA 3.0 - 4.0, the impact of those concepts for troubleshooting CUDA, and how TotalView helps users deal with these new CUDA-specific constructs. CUDA is frequently used alongside MPI parallelism and host-side multicore and multithread parallelism. The TotalView parallel debugger provides developers with an integrated view of all three levels of parallelism within a single debugging session.

Continue Reading Here

Transitioning to multicore: Part II

Multicore systems are ubiquitous; it’s virtually impossible to buy even commodity computers without a dual, quad, or hex-core processor. It won’t be long before many-core processors start to be prevalent as well. Each core in a multicore processor is capable of executing a program, so a quad-core processor can run four separate programs at the same time. That’s great if you have many different programs you need to run at one time, but can become a problem when you need performance from a single program. Those four cores can also potentially run one program faster than a single core processor would, but only if the program is written correctly. If you run a sequential (or serial) program written for single core architectures on a multicore platform, it will generally only be able to leverage a single core. Serial programs don’t run any faster, and may even run slightly slower, on multicore processors.

Continue Reading Here

Many integrated core debugging

Intel® Xeon® Phi™ coprocessors present an exciting opportunity for HPC developers to take advantage of many-core processor technology. Since the Intel Xeon Phi coprocessor shares many architectural features and much of the development tool chain with multicore Intel Xeon processors, it is generally fairly easy to migrate a program to the Intel Xeon Phi coprocessor. However, to fully leverage the Intel Xeon Phi coprocessor, a new level of parallelism needs to be expressed which may require significantly rethinking algorithms. Scientists need tools that support debugging and optimizing hybrid MPI/OpenMP parallel applications that may have dozens or even hundreds of threads per node.

Continue Reading Here

Validating C++ translations of MATLAB models with TotalView

This paper describes one of the many ways that developers have been able to use TotalView to enable the development of cutting edge applications. Frequently, applications in HPC are developed in stages. Problem domains such as signal processing and data analysis, applied physics modeling (geophysics, meteorology, astrophysics, computational chemistry), digital content creation, and financial analysis each have highly specialized computational challenges that call on the skills of domain specialists with deep experience in the subject matter. The algorithms, computational kernels, and other modules written by these domain specialists often are integrated together by software engineers and computational scientists who are specialists in areas like modern software component architectures, parallelism, and embedded development. This division of labor sets the stage for the creation of sophisticated applications that no single developer could have written.

Continue Reading Here

Displaying results 1-10 (of 45)
 |<  < 1 - 2 - 3 - 4 - 5  >  >|