Tore Larsen

Anti-predator behaviour and mating systems in waders: aggressive nest defence selects for monogamy

Abstract Data collected from the literature on anti-predator behaviour, body weight and mating systems of 90 Holarctic waders are presented. Aggressive nest defence is correlated with body weight, heavy species being more likely to attack predators than are small species. Monogamy is common among large waders. Polygamy is generally restricted to small waders, and monogamous species are significantly more aggressive than polygamous species. Presumably, a certain body size is necessary to chase away predators, and being above this threshold size, two cooperating parents guard their nest and brood more efficiently than one parent. Small birds generally do not gain the same anti-predator advantage from monogamy as large birds, and this is presumed to explain the high frequency of polygamous mating systems in small waders.

Past-Set - A Distributed Structured Shared Memory System

The architecture and performance of a structured distributed shared memory system, PastSet, is described. The PastSet abstraction allows programmers to write applications that run efficiently on different architectures from four-way SMP nodes to larger clusters. PastSet is a tuple-based three-dimensional structured distributed shared memory system, which provides the programmer with operations to perform causally ordered reads and writes of tuples to a virtual structured memory space called the PastSet. PastSet is specially implemented to utilize physical shared memory where available and distributed memory otherwise. The contribution of the PastSet model is good performance combined with ease of programming new and porting existing applications. It has been show in [1] that a shared memory version of PastSet is able to outperform System V IPC on the same platform, both multiprocessors and uni-processor systems. We have also previously shown that running on a cluster of multiprocessors, PastSet was able to outperform MPI and PVM when running real applications.

Using overdecomposition to overlap communication latencies with computation and take advantage of SMT processors

Parallel programs running on clusters are typically decomposed and mapped to run with one thread per processor each working on its disjoint subset of the data. We evaluate performance improvements and limitations for a micro-benchmark and the NAS benchmarks, by using overdecomposition to map multiple threads to each processor to overlap computation with communication. The experiment platform is a cluster with Pentium 4 symmetric multithreading (SMT) processor nodes interconnected through gigabit Ethernet. Micro-benchmark results demonstrate execution time improvements up to 1.8. However, for the NAS benchmarks overdecomposition and SMT provides only slight performance gains, and sometimes significant performance loss. We evaluated improvement and limitation sensitivity to problem size, communication structure and whether SMT is enabled or not. We found that performance improvements are limited by applications having communication dependencies that limit thread-level parallelism, increase in cache misses, or increased systems activity. Our study contributes a better understanding of these limitations

MultiStream A Cross-Platform Display Sharing System Using Multiple Video Streams

For collaboration, cross-platform sharing of display content amongst desktop, laptop, handheld computers and smart phones is needed. Due to architectural and performance differences, support for sharing of display content is complex and the performance is low. By using standard media players and video stream formats we reduce or avoid several of these complexities and performance bottlenecks. We do continuous video encoding of display content, and a video server processes and distributes customized or fixed rate videos to viewers. We report on a configuration encoding a video of the display content of a laptop, streaming it to the video server, and playing it back on up to 560 clients distributed over 28 computers. The video displays on a wall-sized high-resolution tiled display. FPS is reduced by 2/3 on the laptop, but the bandwidth usage is reduced by two orders of magnitude. The CPU load on the server is below 50%.

Systems support for remote visualization of genomics applications over wide area networks

Microarray experiments can provide molecular-level insight into a variety of biological processes, from yeast cell cycle to tumorogenesis. However, analysis of both genomic and protein microarray data requires interactive collaborative investigation by biology and bioinformatics researchers. To assist collaborative analysis, remote collaboration tools for integrative analysis and visualization of microarray data are necessary. Such tools should: (i) provide fast response times when used with visualization-intensive genomics applications over a low-bandwidth wide area network, (ii) eliminate transfer of large and often sensitive datasets, (iii) work with any analysis software, and (iv) be platform-independent. Existing visualization systems do not satisfy all requirements. We have developed a remote visualization system called Varg that extends the platform-independent remote desktop system VNC with a novel global compression method. Our evaluations show that the Varg system can support interactive visualization-intensive genomic applications in a remote environment by reducing bandwidth requirements from 30:1 to 289:1.

Extending collective operations with application semantics for improving multi-cluster performance

We identify two ways of increasing the performance of allreduce-style of collective operations in a multi-cluster with large WAN latencies: (i) hiding latency in system noise, and (ii) conditional-allreduce where knowledge about the application is used to reduce the number of WAN messages. In our multicluster, system noise was not large enough to hide the WAN latency. But, the latency could be hidden using conditional-allreduce, since on many iterations only cluster-local values were needed, and many of the values needed from other clusters were prefetched. A speedup of 2.4 was achieved for a microbenchmark. Prefetching introduced a small overhead in the cluster with the slowest hosts.

PastSet-an efficient high level inter process communication mechanism

A new high level IPC mechanism, PastSet, is presented. PastSet supports partial causal ordered logging of synchronization and communication events. Communicated data are represented as tuples and stored in a common repository accessible by all processes. In the repository, tuples of the same template are causally ordered. While being equivalent to semaphores, messages and pipes, PastSet also offer a powerful abstraction suitable for knapsack type parallel applications as well as applications that require state logging. An implementation on uniprocessors and four-way multiprocessors running Linux. The mechanism is integrated into the Linux kernel alongside existing Sys V mechanisms. PastSet gives up to 40% faster synchronization and up to 56% better small package bandwidths than achieved with Linux semaphores, messages, and pipes. Although being a higher abstraction level mechanism, PastSet proves to consistently outperform existing Linux Sys V interprocess communication mechanisms on multiprocessors and to a large extend also on uniprocessors.

A resource request model for microcode compaction

A new approach to resource conflict analysis in microcode compaction has been developed. This paper begins with a description of the resource binding problem in microcode compaction and an analysis of earlier solutions. Then a new specification technique is presented and related to a standard problem in combinatorial theory. This new technique allows microoperations to be specified with resource choices, the actual binding to specific resources being delayed until compaction time. A polynomial-time algorithm for analyzing resource conflicts under this model is explained. Further extensions to the algorithm to support more complex machine models are suggested.