VISUALIZATION

Today, Etnus announced the availability of TotalView at IBM's Deep Computing Capacity On Demand Center in Rochester, Minnesota running on the IBM Blue Gene platform. Etnus is also announcing a remarketing agreement with IBM in which IBM and Etnus will develop a "go to market plan" to promote the use of TotalView for a number of IBM platforms, including AMD, Intel and Power-based Linux clusters, Power-based AIX systems, and IBM's powerful Blue Gene solution. The TotalView debugger has had great success on the Lawrence Livermore National Laboratory Blue Gene/L, the world's fastest supercomputer where it has been used to debug ultra-scale jobs in which over 8,000 processors are orchestrated. Etnus and IBM have collaborated on the TotalView platforms since 1995. Today, TotalView is used by numerous IBM customers to quickly find elusive bugs and memory leaks for more productive software development. TotalView is used on three of the most powerful computers in the world - all IBM systems - according to TOP500 list of supercomputers, including both Blue Gene/L and ASC Purple systems at Lawrence Livermore National Laboratory and the eServer pSeries 655 at the Naval Oceanographic Office. "The proliferation of multi-core chips, multi-threaded code, multi-processor systems and cluster computing make fast, accurate debugging a considerable challenge," said Jim Chafel, Etnus vice president of business development. "We are pleased to help IBM offer its customers a software development productivity tool uniquely equipped to meet that challenge and designed to reduce the development cycle." "IBM is committed to provide the world's fastest, most powerful computers, to facilitate scientific discovery, business competitiveness and homeland security," said Herb Schultz, Deep Computing Marketing, IBM. "By marketing TotalView, as well as making it available at our Deep Computing Capacity on Demand centers, we have another capability that can help empower our customers to meet the rigorous debugging demands of these powerful systems." Etnus has numerous customers using TotalView to debug sophisticated programs developed on IBM systems. California Institute of Technology, for example, uses TotalView on 128 processors of an IBM Power 5 machine running AIX -- for a shock tube simulation that provides insight into Richtmyer-Meshkov (shock-driven) instabilities and turbulent mixing physics. Users there find the integrated, interactive memory debugger within TotalView to be a valuable development tool. "Our adaptive mesh refinement application involves dynamic memory consumption that, under extreme runtime conditions, can cause the application to perform very slowly or crash altogether," notes senior scientist Sharon Brunett. "TotalView enables us to detect and avoid these problems by monitoring memory usage at specified key stages in the application. In addition, it helps us understand differences in memory consumption when identical application sources and baseline conditions are run on different platforms." In another example, ABAQUS uses TotalView on IBM systems running AIX on Power to address the challenges of debugging in data-intensive and distributed computing environments. "ABAQUS uses TotalView to debug, large, complex programs on IBM systems running AIX on Power," said ABAQUS senior development engineer Robert Schweikert. "Additionally, we also use multiple programming languages and support multiple computing platforms, which add to our quality control requirements. TotalView has the power and flexibility to address these challenges and more. The software assists in making our quality checks more efficient and productive, which reduces our time-to-market."