Compiler/Runtime Framework for Dynamic Dataflow Parallelization of Tiled Programs

Task-parallel languages are increasingly popular. Many of them provide expressive mechanisms for intertask synchronization. For example, OpenMP 4.0 will integrate data-driven execution semantics derived from the StarSs research language. Compared to the more restrictive data-parallel and fork-join concurrency models, the advanced features being introduced into task-parallelmodels in turn enable improved scalability through load balancing, memory latency hiding, mitigation of the pressure on memory bandwidth, and, as a side effect, reduced power consumption. In this article, we develop a systematic approach to compile loop nests into concurrent, dynamically constructed graphs of dependent tasks. We propose a simple and effective heuristic that selects the most profitable parallelization idiom for every dependence type and communication pattern. This heuristic enables the extraction of interband parallelism (cross-barrier parallelism) in a number of numerical computations that range from linear algebra to structured grids and image processing. The proposed static analysis and code generation alleviates the burden of a full-blown dependence resolver to track the readiness of tasks at runtime. We evaluate our approach and algorithms in the PPCG compiler, targeting OpenStream, a representative dataflow task-parallel language with explicit intertask dependences and a lightweight runtime. Experimental results demonstrate the effectiveness of the approach.

La gestión del turismo en los programas de innovación rural de Castilla-La Mancha

[ES] Los programas de innovación rural han supuesto un verdadero revulsivo para determinadas regiones europeas, por cuanto han representado una cuantiosa inversión que ha necesitado de una adecuada gestión para dirigir el destino de unos fondos estructurales. En este sentido, el presente trabajo analiza la inversión de estos programas de desarrollo en medidas de apoyo al sector del turismo rural, que han supuesto un importante incremento de la oferta a diversos niveles: aumento de alojamientos rurales, del número de plazas disponibles para turistas, rehabilitación del patrimonio, puesta en valor de elementos históricos y recuperación, en definitiva, del patrimonio histórico, cultural, arquitectónico y natural del mundo rural.

Justicia victimal. Contribuciones y retos

[ES]La pena de cárcel, como única respuesta al delito, no constituye ninguna solución para el hecho delincuencial. No es solución para la víctima porque queda en el más profundo de los desamparos. No es solución para el infractor porque la cárcel no sólo no rehabilita sino que puede generar más delincuencia, como lo acredita el alto índice de reincidencia. Finalmente, no es una solución para la Comunidad por los altos costes, no sólo penitenciarios. Sólo integrada con otras respuestas no carcelarias, la respuesta prisional permite un abordaje sensato de la delincuencia. Se aboga, por ello, por una justicia que reconozca la existencia de otras instancias reparadoras como: la mediación, el arbitraje, el diálogo víctima - agresor, etc.

Rescue, rehabilitation, and release of marine mammals: An analysis of current views and practices.

Stranded marine mammals have long attracted public attention. Those that wash up dead are, for all their value to science, seldom seen by the public as more than curiosities. Animals that are sick, injured, orphaned or abandoned ignite a different response. Generally, public sentiment supports any effort to rescue, treat and return them to sea. Institutions displaying marine mammals showed an early interest in live-stranded animals as a source of specimens -- in 1948, Marine Studios in St. Augustine, Florida, rescued a young short-finned pilot whale (Globicephala macrorhynchus), the first ever in captivity (Kritzler 1952). Eventually, the public as well as government agencies looked to these institutions for their recognized expertise in marine mammal care and medicine. More recently, facilities have been established for the sole purpose of rehabilitating marine mammals and preparing them for return to the wild. Four such institutions are the Marine Mammal Center (Sausalito, CA), the Research Institute for Nature Management (Pieterburen, The Netherlands), the RSPCA, Norfolk Wildlife Hospital (Norfolk, United Kingdom) and the Institute for Wildlife Biology of Christian-Albrects University (Kiel, Germany).(PDF contains 68 pages.)

Lessons from established breeding programs: terrestrial and aquatic animals

Some relevant components of selection program theory and implementation are reviewed. This includes pedigree recording, genetic evaluation, balancing genetic gains and genetic diversity and tactical integration of key issues. Lessons learned are briefly described – illustrating how existing method and tools can be useful when launching a program in a novel species, and yet highlighting the importance of proper understanding and custom application according to the biology and environments of that species.

Breeding programs on Atlantic salmon in Norway: lessons learned

An early establishment of selective breeding programs on Atlantic salmon has been crucial for the success of developing efficient and sustainable salmon farming in Norway. A national selective breeding program was initiated by AKVAFORSK at the beginning of the 1970s, by collecting fertilized eggs from more than 40 Norwegian river populations. Several private selective breeding programs were also initiated in the 1970s and 1980s. While these private programs were initiated using individual selection (i.e. massselection) to genetically improve growth, the national program was designed to gradually include all economically important traits in the breeding objective (i.e. growth, age at sexual maturation, disease resistance and quality traits) using a combined family and within-family selection strategy. Independent of which selection strategy and program design used, it is important to secure and maintain a broad genetic variation in the breeding populations to maximize selection response. It has been documented that genetically improved salmon from the national selective breeding program grow twice as fast as wild Atlantic salmon and require 25 per cent less feed, while salmon representing the private breeding programs all show an intermediate growth performance. As a result of efficient dissemination of genetically improved Atlantic salmon, the Norwegian salmon farming industry has reduced its feed costs by more than US$ 230 million per year! The national selective breeding program on Atlantic salmon was commercialized into a breeding company (AquaGen) in 1992. Five years later, several private companies and the AKVAFORSK Genetics Center (AFGC) established a second breeding company (SalmoBreed) using breeding candidates from one of the private breeding programs. These two breeding companies have similar products, but different strategies on how to organize the breeding program and to disseminate the genetically improved seed to the Norwegian salmon industry. Greater competition has increased the necessity to document the genetic gain obtained from the different programs and to market the economic benefits of farming the genetically improved breeds. Both breeding companies have organized their dissemination to get a sufficient share of the economic benefits in order to sustain and improve their breeding programs.

Cooperative artificial propagation programs for salmon and steelhead, 1995-1996

Fifteen cooperative fish rearing and planting programs for salmon and steelhead were active from July 1, 1995 through June 30, 1996. For all programs, 134,213 steelhead trout,(Oncorhynchus mykiss), 7,742,577 chinook salmon,(~ tshawytscha),and 25,075 coho salmon(~ kisutch) were planted. (PDF contains 26 pages.)

Cooperative artificial propagation programs for salmon and steelhead, 1996-1997

Fourteen cooperative fish rearing and planting programs for salmon and steelhead were active from July 1, 1996 through June 30, 1997. For all programs, 208,922 steelhead trout, (Oncorhynchus mykiss), 10,334,457 chinook salmon,(O. tshawytscha),and 60,681 coho salmon(O. kisutch) were planted. (PDF contains 24 pages.)

Introduction to MLML_DBASE Programs

The MLML DBASE family of programs described here provides many of the. algorithms used in oceanographic data reduction, general data manipulation and line graphs. These programs provide a consistent file structure for serial data typically encountered in oceanography. This introduction should provide enough general knowledge to explain the scope of the program and to run the basic MLML_DBASE programs. It is not intended as a programmer's guide. (PDF contains 50 pages)