May 28, 2010
UOMS will be presented at the International Supercomputing Conference 2010, in Hamburg, Germany
SANTIAGO DE COMPOSTELA, Spain, May 28 -- A team of researchers from the Galicia Supercomputing Center (CESGA), the Computer Architecture Group of the University of A Coruña (UDC) and the Computer Architecture Group of the University of Santiago de Compostela (USC) has developed the first microbenchmarking suite for the parallel programming language Unified Parallel C (UPC). The software, named UPC Microbenchmarking Operations Suite (UOMS) will be released for its use by the research community and will serve as a test to check the performance of the UPC implementations. The new tool allows evaluators to test the execution performance of a series of operations for a wide range of problem sizes and configurations in current and future systems. This code is a part of the project "Improving UPC Usability and Performance in Constellation Systems: Implementation/Extensions of UPC Libraries," developed by CESGA, Hewlett Packard (HP), and the Computer Architecture groups of the UDC and USC.
The work introduces a suite that intends to cover the lack of evaluation software and tools for UPC compilers. PGAS (Partitioned Global Address Space) languages, as UPC, due to its youth, lack some necessary tools and libraries. UOMS addresses the lack of evaluation tools to assess the performance of UPC memory transfer functions, as well as computational collectives and synchronization operations. With the UOMS, UPC is ready to tackle the next step in its evolution.
Languages Increasingly Important
Traditional paradigms are becoming less effective for programming current HPC platforms. The introduction of manycore systems represent a significant architecture shift where message passing approaches do not naturally fit. Thus, the interest in new languages that overcome the complexity and programmability issues of current parallel libraries and languages is rising. For that reason PGAS languages, UPC in particular, are becoming increasingly important for nowadays parallel programming. HPC vendors are catching up and have their own implementations, and academia provided several open source compilers and runtimes to worldwide developers.
The code will be presented at the International Supercomputing Conference 2010, the most prestigious conference in Europe in its field, which will be held this June in Hamburg. It is available now on the Web at upc.cesga.es. Galicia Supercomputing Center (CESGA) is a Foundation taken part for the Galician Government (Xunta de Galicia) and for the Spanish National Research Council (CSIC). From its creation in 1993, CESGA has the mission of contributing to the advance of science and the technology through research and application of high performance computing and communications, in collaboration with other institutions, for the welfare of society.
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?