Multicore-enabling the MPJ express messaging library

With the transition to multicore processors almost complete, the parallel processing community is seeking efficient ways to port legacy message passing applications on shared memory and multicore processors. MPJ Express is our reference implementation of Message Passing Interface (MPI)-like bindings for the Java language. Starting with the current release, the MPJ Express software can be configured in two modes: the multicore and the cluster mode. In the multicore mode, parallel Java applications execute on shared memory or multicore processors. In the cluster mode, Java applications parallelized using MPJ Express can be executed on distributed memory platforms like compute clusters and clouds. The multicore device has been implemented using Java threads in order to satisfy two main design goals of portability and performance. We also discuss the challenges of integrating the multicore device in the MPJ Express software. This turned out to be a challenging task because the parallel application executes in a single JVM in the multicore mode. On the contrary in the cluster mode, the parallel user application executes in multiple JVMs. Due to these inherent architectural differences between the two modes, the MPJ Express runtime is modified to ensure correct semantics of the parallel program. Towards the end, we compare performance of MPJ Express (multicore mode) with other C and Java message passing libraries---including mpiJava, MPJ/Ibis, MPICH2, MPJ Express (cluster mode)---on shared memory and multicore processors. We found out that MPJ Express performs signicantly better in the multicore mode than in the cluster mode. Not only this but the MPJ Express software also performs better in comparison to other Java messaging libraries including mpiJava and MPJ/Ibis when used in the multicore mode on shared memory or multicore processors. We also demonstrate effectiveness of the MPJ Express multicore device in Gadget-2, which is a massively parallel astrophysics N-body siimulation code.

Data-mining chess databases

This is a report on the data-mining of two chess databases, the objective being to compare their sub-7-man content with perfect play as documented in Nalimov endgame tables. Van der Heijden’s ENDGAME STUDY DATABASE IV is a definitive collection of 76,132 studies in which White should have an essentially unique route to the stipulated goal. Chessbase’s BIG DATABASE 2010 holds some 4.5 million games. Insight gained into both database content and data-mining has led to some delightful surprises and created a further agenda.

Implications for DNA identification arising from an analysis of Australian forensic databases

Previous analyses of Australian samples have suggested that populations of the same broad racial group (Caucasian, Asian, Aboriginal) tend to be genetically similar across states. This suggests that a single national Australian database for each such group may be feasible, which would greatly facilitate casework. We have investigated samples drawn from each of these groups in different Australian states, and have quantified the genetic homogeneity across states within each racial group in terms of the "coancestry coefficient" F(ST). In accord with earlier results, we find that F(ST) values, as estimated from these data, are very small for Caucasians and Asians, usually <0.5%. We find that "declared" Aborigines (which includes many with partly Aboriginal genetic heritage) are also genetically similar across states, although they display some differentiation from a "pure" Aboriginal population (almost entirely of Aboriginal genetic heritage).

Baculovirus superinfection: a probable restriction factor on the surface display of proteins for library screening

In addition to the expression of recombinant proteins, baculoviruses have been developed as a platform for the display of complex eukaryotic proteins on the surface of virus particles or infected insect cells. Surface display has been used extensively for antigen presentation and targeted gene delivery but is also a candidate for the display of protein libraries for molecular screening. However, although baculovirus gene libraries can be efficiently expressed and displayed on the surface of insect cells, target gene selection is inefficient probably due to super-infection which gives rise to cells expressing more than one protein. In this report baculovirus superinfection of Sf9 cells has been investigated by the use of two recombinant multiple nucleopolyhedrovirus carrying green or red fluorescent proteins under the control of both early and late promoters (vAcBacGFP and vAcBacDsRed). The reporter gene expression was detected 8 hours after the infection of vAcBacGFP and cells in early and late phases of infection could be distinguished by the fluorescence intensity of the expressed protein. Simultaneous infection with vAcBacGFP and vAcBacDsRed viruses each at 0.5 MOI resulted in 80% of infected cells coexpressing the two fluorescent proteins at 48 hours post infection (hpi), and subsequent infection with the two viruses resulted in similar co-infection rate. Most Sf9 cells were re-infectable within the first several hours post infection, but the reinfection rate then decreased to a very low level by 16 hpi. Our data demonstrate that Sf9 cells were easily super-infectable during baculovirus infection, and super-infection could occur simultaneously at the time of the primary infection or subsequently during secondary infection by progeny viruses. The efficiency of super-infection may explain the difficulties of baculovirus display library screening but would benefit the production of complex proteins requiring co-expression of multiple polypeptides.