As HPC capability and software adaptability continues to expand, the interest to perform complex system-wide simulations involving multiple interacting components grows. In this paper, we present a novel integrative software platform -- the Distributive Interoperable Executive Library (DIEL) - to facilitate the collaboration, exploration, and execution of multiphysics modeling projects suited for a diversified research community on emergent large-scale parallel computing platforms. It does so by providing a managing executive, a layer of numerical libraries, a number of commonly used physics modules, and two set of native communication protocols. DIEL allows users to plug in their individual modules, prescribe the interactions between those modules, and schedule communications between them. The DIEL framework is designed to be applicable for preliminary concept design, sensitivity prototyping, and productive simulation of a complex system.

Decision support (DSS) workloads generally contain multiple large concurrent scan operations. These are often executed as relational table scans which can take up a lot of I/O bandwidth. This is especially true for ad-hoc queries where the workload is not known in advance. Common database management systems have only limited ability to reuse memory buffer content across multiple running queries due to their treatment of queries in isolation. Previous attempts to coordinate scans for better buffer reuse were less than satisfactory due to drifting between scans and the required radical DBMS architecture changes. In this paper, we describe a new mechanism to keep similar table scans closer together during scanning. This is achieved via dynamic grouping and regrouping of scans based on their runtime behavior and via adaptive throttling of scan speeds based on scan group characteristics. The required memory footprint is very small and the effort required to extend existing database management systems is minimal, as shown in our DB2 UDB prototype. Our experiments show significant gains in end-to-end response times as well as average response times for TPC-H workloads.

Database queries often take the form of correlated SQL queries. Correlation refers to the use of values from the outer query block to compute the inner subquery. This is a convenient paradigm for SQL programmers and closely mimics a function invocation paradigm in a typical computer programming language. Queries with correlated subqueries are also often created by SQL generators that translate queries from application domain-specific languages into SQL. Another significant class of queries that use this correlated subquery form is that involving temporal databases using SQL. Performance of these queries is an important consideration particularly in large databases. Several proposals to improve the performance of SQL queries containing correlated subqueries can be found in database literature. One of the main ideas in many of these proposals is to suitably decorrelate the subquery internally to avoid a tuple-at-a-time invocation of the subquery. Magic decorrelation is one method that has been successfully used. Another proposal is to cache the portion of the subquery that is invariant with the changing values of the outer query block. What we propose here is a new technique to handle some typical correlated queries. We go a step further than to simply decorrelate the subquery. By making use of extended window aggregation capabilities, we eliminate redundant access to common tables referenced in the outer query block and the subquery. This technique can be exploited even for non-correlated subqueries. It is possible to get a huge boost in performance for queries that can exploit this technique, which we call WinMagic. This technique was implemented in IBM® DB2® Universal Database" Version 7 and Version 8. In addition to improving DB2 customer queries that contain aggregation subqueries, it has provided significant improvements in a number of TPCH benchmarks that IBM has published since late in 2001.