Direct Access

B. Monien and R. Feldmann (Eds.): Euro-Par 2002, LNCS 2400, pp. 517–521. © Springer-Verlag Berlin Heidelberg 2002 ... Speeding Up Target Address Generation Using a Self-indexed FTB ... Juan C. Moure, Dolores I. Rexachs, and Emilio Luque1 ... Computer Architecture and ...

Economic growth in Sub-Saharan Africa (SSA) has averaged roughly 5 percent per year over the past decade, improving living standards and bolstering human development indicators across the continent. Stronger public institutions, a supportive, private sector–focused policy environment, responsible macroeconomic management, and a sustained commitment to structural reforms have greatly expanded opportunities for countries in SSA to participate in global markets. In recent years, many countries in the region have benefited from an increasingly favorable external environment, high commodity prices, and an especially strong demand for natural resources by emerging economies, particularly China. Over the longer term, leveraging Chinese investment to support broad-based growth will require policies designed to boost the competitiveness of sectors in which China’s economic rebalancing may create a comparative advantage for SSA. To date, few African countries have been able to benefit from la...

B. Monien and R. Feldmann (Eds.): Euro-Par 2002, LNCS 2400, pp. 517–521. © Springer-Verlag Berlin Heidelberg 2002 ... Speeding Up Target Address Generation Using a Self-indexed FTB ... Juan C. Moure, Dolores I. Rexachs, and Emilio Luque1 ... Computer Architecture and ...

The All-to-all broadcast collective operation is essential for many parallel scientific applications. This collective operation is called MPI_Allgather in the context of MPI. Contemporary MPI software stacks implement this collective on top of MPI point-to-point calls leading to several performance overheads. In this paper, we propose a design of All-to-All broadcast using the Remote Direct Memory Access (RDMA) feature offered by InfiniBand, an emerging high performance interconnect. Our RDMA based design eliminates the overheads associated with existing designs. Our results indicate that latency of the All-to-all Broadcast operation can be reduced by 30% for 32 processes and a message size of 32 KB. In addition, our design can improve the latency by a factor of 4.75 under no buffer reuse conditions for the same process count and message size. Further, our design can improve performance of a parallel matrix multiplication algorithm by 37% on eight processes, while multiplying a 256x256 matrix.

Economic growth will be unbalanced, but development can still be inclusive-that is the message of this year's World development report, the thirty-first in the series. As economies grow from low to high income, production becomes more concentrated spatially. Some places, cities, coastal areas, and connected countries-are favored by producers. As countries develop, the most successful ones also institute policies that make living standards of people more uniform across space. The principle for a successful spatial transformation, getting the immediate benefits of concentration of production, and the long term benefits of a convergence in living standards, is economic integration. The report first describes the spatial transformations needed for development. It analyzes these changes using the insights from economic history and recent research. Then it revisits the policy debates on urbanization, regional development, and international integration.

This paper describes the clinical outcomes from a novel direct access arrhythmia monitoring service. The study was carried out in the north of Scotland. Data was collected over a 29 month period between 18 June 2008 and 8 November 2010 from consecutive cases from two groups of patients, general practitioner (GP) direct access and 'redirected' consultant referrals. Monitor test results, frequency of arrhythmias requiring further care and clinic attendances were recorded. Statistical differences were analyzed using Χ(2), Fisher's and Student's t-test as appropriate with the significance taken at the 0.05 level. 239 patients were referred from 47 GP practices. There were 165 (69%) referrals through the 'direct' and 72 (31%) through the 'redirected' route. The average age was 55.5 ± 16.7 years with 84 (35.1%) males. 127 (53.1%) had a patient activated event recording and the remaining 112 (46.9%) had Holter monitoring. Of the 239 patients, only nine (3.8%...

The All-to-all broadcast collective operation is essential for many parallel scientific applications. This collective operation is called MPI_Allgather in the context of MPI. Contemporary MPI software stacks implement this collective on top of MPI point-to-point calls leading to several performance overheads. In this paper, we propose a design of All-to-All broadcast using the Remote Direct Memory Access (RDMA) feature offered by InfiniBand, an emerging high performance interconnect. Our RDMA based design eliminates the overheads associated with existing designs. Our results indicate that latency of the All-to-all Broadcast operation can be reduced by 30% for 32 processes and a message size of 32 KB. In addition, our design can improve the latency by a factor of 4.75 under no buffer reuse conditions for the same process count and message size. Further, our design can improve performance of a parallel matrix multiplication algorithm by 37% on eight processes, while multiplying a 256x256 matrix.

The Daytona™ data management system is used by AT&T to solve a wide spectrum of data management problems. For example, Daytona is managing a 4 terabyte data warehouse whose largest table contains over 10 billion rows. Daytona's architecture is based on translating its high-level query language Cymbal™ (which includes SQL as a subset) completely into C and then compiling that