Automatic and effective multi-dimensional parallelisation of structured mesh based codes

The most common parallelisation strategy for many Computational Mechanics (CM) (typified by Computational Fluid Dynamics (CFD) applications) which use structured meshes, involves a 1D partition based upon slabs of cells. However, many CFD codes employ pipeline operations in their solution procedure. For parallelised versions of such codes to scale well they must employ two (or more) dimensional partitions. This paper describes an algorithmic approach to the multi-dimensional mesh partitioning in code parallelisation, its implementation in a toolkit for almost automatically transforming scalar codes to parallel form, and its testing on a range of ‘real-world’ FORTRAN codes. The concept of multi-dimensional partitioning is straightforward, but non-trivial to represent as a sufficiently generic algorithm so that it can be embedded in a code transformation tool. The results of the tests on fine real-world codes demonstrate clear improvements in parallel performance and scalability (over a 1D partition). This is matched by a huge reduction in the time required to develop the parallel versions when hand coded – from weeks/months down to hours/days.

Parallel performance in multi-physics simulation

A comprehensive simulation of solidification/melting processes requires the simultaneous representation of free surface fluid flow, heat transfer, phase change, non-linear solid mechanics and, possibly, electromagnetics together with their interactions in what is now referred to as "multi-physics" simulation. A 3D computational procedure and software tool, PHYSICA, embedding the above multi-physics models using finite volume methods on unstructured meshes (FV-UM) has been developed. Multi-physics simulations are extremely compute intensive and a strategy to parallelise such codes has, therefore, been developed. This strategy has been applied to PHYSICA and evaluated on a range of challenging multi-physics problems drawn from actual industrial cases.

Multi-function oxidases are responsible for the synergistic interactions occurring between repellents and insecticides in mosquitoes

Background: With the spread of pyrethroid resistance in mosquitoes, the combination of an insecticide (carbamate or organophosphate) with a repellent (DEET) is considered as a promising alternative strategy for the treatment of mosquito nets and other relevant materials. The efficacy of these mixtures comes from the fact that they reproduce pyrethroid features and that positive interactions occur between insecticides and repellent. To better understand the mechanisms involved and assess the impact of detoxifying enzymes (oxidases and esterases) in these interactions, bioassays were carried out in the laboratory against the main dengue vector Aedes aegypti. Methods: Topical applications of DEET and propoxur (carbamate), used alone or as a mixture, were carried out on female mosquitoes, using inhibitors of the two main detoxification pathways in the insect. PBO, an inhibitor of multi-function oxidases, and DEF, an inhibitor of esterases, were applied one hour prior to the main treatment. Results: Results showed that synergism between DEET and propoxur disappeared in the presence of PBO but not with DEF. This suggests that oxidases, contrary to esterases, play a key role in the interactions occurring between DEET and cholinesterase inhibitors in mosquitoes. Conclusion: These findings are of great interest for the implementation of "combination nets" in the field. They support the need to combine insecticide with repellent to overcome insecticide resistance in mosquitoes of public health importance.

Numerical and experimental studies of a substructural identification strategy

For structural health monitoring it is impractical to identify a large structure with complete measurement due to limited number of sensors and difficulty in field instrumentation. Furthermore, it is not desirable to identify a large number of unknown parameters in a full system because of numerical difficulty in convergence. A novel substructural strategy was presented for identification of stiffness matrices and damage assessment with incomplete measurement. The substructural approach was employed to identify large systems in a divide-and-conquer manner. In addition, the concept of model condensation was invoked to avoid the need for complete measurement, and the recovery process to obtain the full set of parameters was formulated. The efficiency of the proposed method is demonstrated numerically through multi-storey shear buildings subjected to random force. A fairly large structural system with 50 DOFs was identified with good results, taking into consideration the effects of noisy signals and the limited number of sensors. Two variations of the method were applied, depending on whether the sensor could be repositioned. The proposed strategy was further substantiated experimentally using an eight-storey steel plane frame model subjected to shaker and impulse hammer excitations. Both numerical and experimental results have shown that the proposed substructural strategy gave reasonably accurate identification in terms of locating and quantifying structural damage.