25 26 27 Salmonella genomic island 1 (SGI1) and variants (SGI1-I and the new variant SGI128 O) were mapped in five strains of Proteus mirabilis isolated from humans and food 29 in China. Sequencing showed SGI1 and variants were integrated at the 3’-end of the 30 chromosomal thdF gene as previously described in Salmonella strains. 31

This paper presents an MPI-parallelized version of a matching algorithm for synthetic aperture radar (SAR) images. First the necessity for automated matching and the chosen algorithm are outlined briefly, and the algorithmic structure and the applied parallelization strategy are then presented and discussed. Performance results on a Cray T3D and a PowerChallenge from SGI are given, showing unex...

We describe a simulator-parallel method for parallelization of a large industrial Compositional Simulator. Our approach is based on domain-decomposition, with a coarse grid operator, generated by a Galerkin technique. The technique is highly object-oriented, which gives a high level of parallelization. Also inclusion of local mesh reenement can be done in a eecient and exible way. The code is r...

In preparation of upcoming procurements in Germany a comprehensive performance evaluation for a selection of applications and kernels has been performed. The main focus of our study was on codes from computational fluid dynamics, which are known to be memory-intensive. In the light of our results we want to comment on the discussion about the need for tailored HPC systems such as SGI Altix or v...

We present a new formulation of fast Fourier transformation (FFT) kernels for radix 2, 3, 4, and 5, which have a perfect balance of multiplies and adds. These kernels give higher performance on machines that have a single multiply–add (mult–add) instruction. We demonstrate the superiority of this new kernel on IBM and SGI workstations. Key word. FFT kernels AMS subject classifications. 65-04, 4...

In this paper we present some computational strategies tailored for the Finite Element solution of large-scale ow problems in high performance computers. Reduced Integration techniques and edge-based data structures are studied. We also introduce new precon-ditioners for implicit schemes based on superedges clustering. The performance of these new computational strategies is evaluated on the Cr...