CEA and Bull select Allinea Debugging and Profiling solutions for global research into sustainable energy sources

Both the CEA and Bull have reaffirmed their longstanding relationships with Allinea Software (www.allinea.com) by choosing the company’s industry leading DDT and OPT debugging and Optimising / Profiling solutions for the International Fusion Energy Research Center (IFERC).
Earlier this year the CEA (on behalf of F4E) selected Bull to provide, maintain and operate the computer that is to be installed at Rokkasho (Japan). This computer is intended to allow the most advanced modeling and simulation in the field of plasmas and controlled fusion equipment. It will be available to European and Japanese researchers for a period of five years from January 2012.

With computational power from Bull exceeding that of 1 Petaflop, it ranks as one of the most powerful systems in the world. The supercomputer is an important milestone of Europe’s contribution to the Broader Approach (BA), an Agreement signed between Europe and Japan to complement the ITER project through various R&D activities which are developed in the field of nuclear fusion.

François Robin, leader of the CEA’s contribution in the project, explains: “As with the introduction of Bull, the selection of Allinea Software is another positive step for this project.  Allinea DDT is an impressive and complete toolset for parallel debugging at any scale which will ensure that the aims of the project are attainable.”

Olivier David, ISV Alliances Director at Bull, comments: “The bullx supercomputer dedicated to ITER is one of the largest systems in the world, and we needed a debugging solution that can rise up to the challenges of such a petascale system. Allinea is able to address scale and GPU challenges, and has considerable expertise in the parallel field.   Moving forward, the DDT debugger will be available to run on the full size of the system in Japan.”

Also offered as part of the bullx supercomputer suite, Allinea is able to offer their suite of interoperable debugging, optimization and profiling tools for parallel and high performance computing (HPC) on the bullx server ranges.

Jacques Philouze, Worldwide Vice President Sales & Marketing, Allinea Software, comments: “We’re honoured to have been selected by both CEA and Bull to deliver on such an important global project and one which reinforces our presence in the APAC market. Alternative energy research has now become a major and urgent challenge that is driving the quest for speeding up time to results. It’s therefore very rewarding that intuitive tools like ours have been recognised as key components for global researchers, helping to speed up developers’ adoption and time to debug which has a direct impact on the timescale for achieving results.”

Argonne National Labs selects Allinea DDT for Blue Gene and Linux
Debugging tools evolution to support petascale and exascale 

The US Department of Energy's Argonne National Laboratory (www.anl.gov) has selected Allinea DDT as the default tool for parallel debugging on its Blue Gene (BG) and Linux Systems. 
Allinea DDT provides advanced parallel debugging tools for scalar, multi-threaded and parallel applications, and is acclaimed for its intuitiveness and scalability.

Kalyan Kumaran, Manager, Performance Engineering, at Argonne National Labs, explains: “Our on-going collaboration with Allinea will now result in improved tools for our Blue Gene petascale class, as well as for future Exascale systems. We continue to work with Allinea DDT because it enables us to scale to petascale with current commercially released products, as well as providing ease of use, an intuitive user interface, and support for multiple architectures.  Allinea are very focused on continuing to grow their tools capabilities.”

Mike Fish, CEO of Allinea Software, adds: “As scale increases, using a debugger becomes even more essential, yet the performance and interface complexity challenges become greater.  This collaboration with one of the USA’s leading Labs, Argonne, confirms both Allinea DDT’s scalability in support of current, and future, BG systems, as well as the greater confidence in Allinea to deploy on Linux and BG petascale systems. In addition to driving science, the collaborations with DoE labs like Argonne have radically advanced debugging technology, created reusable capabilities shared by multiple agencies and customers, and helped Allinea validate its technology with one of the world’s most demanding supercomputer users.”

Earlier this year Allinea announced the release of DDT 3.0 – the first debugger capable of handling applications running across hundreds of thousands of cores – needed to deliver Petaflops performance. DDT 3.0 enables programmers to increase their efficiency when working at scale. It provides an interface that simplifies the presentation of many processes by merging similarities and highlighting differences, and is an easy to use tool on 100,000 cores as well as on just 100 cores.

Blue Gene is a computer architecture project designed to produce several supercomputers, designed to reach operating speeds in the PFLOPS (petaFLOPS) range, and currently reaching sustained speeds of nearly 500 TFLOPS (teraFLOPS). It is a cooperative project between IBM (particularly IBM Rochester and the Thomas J. Watson Research Center), the Lawrence Livermore National Laboratory, the United States Department of Energy (which is partially funding the project), and academia. There are four Blue Gene projects in development: Blue Gene/L, Blue Gene/C, Blue Gene/P, and Blue Gene/Q.

European supercomputing centre LRZ selects Allinea to debug IBM SuperMUC software

Germany’s Leibniz Supercomputing Centre (LRZ, http://www.lrz.de) has chosen AllineaSoftware (www.allinea.com) to support debugging for the targeted large-scale parallel applications on the IBM SuperMUC system next year.

LRZ, which is the computer centre for Munich's universities and for the Bavarian Academy of Sciences and Humanities, provides high-end computing facilities for the scientific community in Germany and beyond.

The Partnership for Advanced Computing in Europe (PRACE) announced the implementation of the IBM system, which will comprise of more than a hundred thousand cores capable of delivering an impressive three peak petaflops/s of processing power.  When SuperMUC is installed at LRZ in the middle of 2012, it will most likely be a top 10 system.  SuperMUC will follow the 1.0 petaflop "JUGENE" (Jülich Blue Gene" Blue Gene/P supercomputer), already in service at Forschungszentrum Juelich (FZJ), and the 1.25 petaflop Bull-built "Curie" system at the Commissariat a l'Energie Atomique (CEA) in France.

These machines currently occupy 9th and 6th position respectively on the TOP500 list. It will be one of the Tier-0 systems available in the PRACE infrastructure which provides large-scale computational resources and infrastructure for scientific applications throughout Europe.

Dr. Reinhold Bader, deputy group leader of HPC support at LRZ, comments: “Our choice of Allinea’s debugging solutions was determined by a need for both ease of use and scalability which would hold no barriers in using the development tools and could offer researchers superior processing times for their results.”

It will enable LRZ's scientific community to test theories, design experiments and predict outcomes as never before. The supercomputer will be jointly funded by the German federal government and the state of Bavaria.

Mike Fish, CEO of Allinea Software, explains: “We are delighted that LRZ has chosen Allinea to handle its software debugging for its IBM SuperMUC system.  Our software solutions are engineered to empower developers with the tools to fix bugs both quickly and easily, being powerful enough for the most advanced parallel programmers, but also accessible for beginners without training.  In effect, we’re providing target scale debugging whether programming for 8 cores or 200,000”.

This partnership with LRZ represents yet another significant deal for Allinea in the European market.

Allinea Software signs further collaboration agreement with CEA
Addressing large scale debugging with MPC and CUDA

Allinea Software - www.allinea.com - a leading supplier of software development tools for high performance computing (HPC), has signed a further collaboration agreement with CEA, the French government-funded technological research organisation www.cea.fr/english_portal.    

The aim of the collaboration to date has been to develop enhancements to Allinea Software’s market-leading Allinea DDT (Distributed Debugging Tool) for next generation hybrid and “many-core” computer systems.  Based on the work carried out in phase one, the two organisations are seeking to address a major challenge and to offer a solution for large scale systems.

Dr. David Lecomber, CTO at Allinea Software, explains: “Research organisations require ever more powerful computational resources, and so their software needs to run on ever larger numbers of processors. This increases the complexity and risks of errors in software development; Allinea is one of the only suppliers of tools that can handle this complexity and identify errors before they hold up important research.

Based on all of the knowledge we acquired from the first project with CEA, this new collaboration intends to address the concerns and support the aims of all HPC developers worldwide. A key aim is clearly to provide a solution now and in the future for MPC unified parallel runtime (see: http://mpc.sourceforge.net/) large scale debugging.”

The focus of phase two is to make debugging tools both portable and easy to use for large scale debugging well over 100,000 cores. Allinea DDT's interface is intuitive at every scale of parallelism, and its architecture has already been shown to scale well to existing large systems.  Another important aim is to improve the previous work carried out on CUDA debugging by adding more features to address the CUDA architecture and fully exploit the enhancements which NVidia now offers.

Pierre Leca, head of Simulation and Information Sciences Department, CEA, comments: “We’re very pleased with the results of our collaboration with Allinea Software over the past year and we want this to continue.  They are clearly an active, technology leader, with the right product already available, and unrivalled expertise in parallel performance. 

Working together, we can address the growing numbers of cores in clusters and within individual nodes to take advantage of optimizations that are feasible for these systems. This will ensure that developers can use a tool that scales well on large many-core clusters both in terms of performance and in ease of use.”

This next phase of collaboration with CEA is a further illustration of Allinea Software’s leadership in High Performance Computing (HPC) software with the most demanding  HPC sites in both Europe and the USA. 

Allinea announces the availability of DDT on the NVIDIA CUDA architecture

Warwick, United Kingdom – Allinea Software today announced that that it was making a pre-release version of the Distributed Debugging Tool (DDT) for the CUDA architecture available to early-adopter customers. This coincides with the release of Nvidia’s CUDA Toolkit version 3.0.

Over the last year, Allinea has been collaborating with the Commissariat à l’Énergie Atomique (CEA) t develop CUDA-specific feature within its market-leading DDT product. The preliminary results of this close collaboration were shown at Supercomputing ’09, when native debugging of CUDA-enabled applications running on NVIDIA Tesla GPUs and x86 host processors was demonstrated. Following the full release of Nvidia’s CUDA Toolkit 3.0, Allinea is pleased to announce that it will be making a pre-release version of DDT for CUDA available to its customers.

“We are delighted to see the results of our collaboration with the CEA making it into our mainstream DDT product,” said David Lecomber, CTO of Allinea Software. We are now able to provide application developers with a single tool that can debug hybrid MPI, OpenMP and CUDA applications on a single workstation or GPU cluster. Features such as the detection of invalid memory accesses, the visualization of GPU data, and GPU thread control have been designed to help our customers find their GPU porting bugs simply and efficiently."

“Developing code for highly parallel GPU architectures is far from straightforward,” said Pierre LECA, Head of Simulation and Information Sciences Department at CEA, “but the problem is exacerbated when we are faced with parallelism on multiple levels. The CEA have been working closely with Allinea so as to ensure that our application developers will be able to debug applications on hybrid x86/GPU systems. We are very happy with the first results of this collaboration, specifically aimed at the Nvidia CUDA architecture, and look forward to a continued collaboration in order to further extend these capabilities.”

CUDA-specific debugging features will be available in the up-coming product release of DDT, although potential customers are invited to contact Allinea if they are interested in joining the pre-release program;