Current generations of NUMA node clusters feature multicore or manycore processors. Programming such architectures eciently is a challenge because numerous hardware characteristics have to be taken into account, especially the memory hierarchy. One appealing idea to improve the performance of parallel applications is to decrease their communication costs by matching the communication pattern to the underlying hardware architecture. In this report, we detail the algorithm and techniques proposed to achieve such a result: rst, we gather both the communication pattern information and the hardware details. Then we compute a relevant reordering of the various process ranks of the application. Finally, those new ranks are used to reduce the communication costs of the application.

ObjectiveTo assess whether bed bug infestation was linked to sleep disturbances and symptoms of anxiety and depression.DesignExploratory cross-sectional study.SettingConvenience sample of tenants recruited in apartment complexes from Montreal, Canada.Participants39 bed bug-exposed tenants were compared with 52 unexposed tenants.Main outcome measuresThe effect of bed bug-exposed tenants on sleep disturbances, anxiety and depression symptoms measured using the Pittsburgh Sleep Quality Index, 5th subscale, Generalised Anxiety Disorder 7-item scale and Patient Health Questionnaire, 9-item, respectively.ResultsIn adjusted models, bed bug infestation was strongly associated with measured anxiety symptoms (OR (95% CI)=4.8 (1.5 to 14.7)) and sleep disturbance (OR (95% CI)=5.0 (1.3–18.8)). There was a trend to report more symptoms of depression in the bed bug-infested group, although this finding was not statistically significant ((OR (95% CI)=2.5(0.8 to 7.3)).ConclusionsThese results suggest that individuals exposed to bed bug infestations are at risk of experiencing sleep disturbance and of developing symptoms of anxiety and possibly depression. Greater clinical awareness of this problem is needed in order for patients to receive appropriate mental healthcare. These findings highlight the need for undertaking of deeper inquiry, as well as greater collaboration between medical professionals, public health and community stakeholders.

Residential wood burning can be a significant wintertime source of ambient fine particles in urban and suburban areas. We developed a statistical model to predict minute (min) levels of particles with median diameter of o1 mm (PM1) from mobile monitoring on evenings of winter weekends at different residential locations in Quebec, Canada, considering wood burning emissions. The 6 s PM1 levels were concurrently measured on 10 preselected routes travelled 3 to 24 times during the winters of 2008--2009 and 2009--2010 by vehicles equipped with a GRIMM or a dataRAM sampler and a Global Positioning System device. Route-specific and global land-use regression (LUR) models were developed using the following spatial and temporal covariates to predict 1-min-averaged PM1 levels: chimney density from property assessment data at sampling locations, PM2.5 ''regional background'' levels of particles with median diameter of o2.5 mm (PM2.5) and temperature and wind speed at hour of sampling, elevation at sampling locations and day of the week. In the various routes travelled, between 49% and 94% of the variability in PM1 levels was explained by the selected covariates. The effect of chimney density was not negligible in ''cottage areas.'' The R 2 for the global model including all routes was 0.40. This LUR is the first to predict PM1 levels in both space and time with consideration of the effects of wood burning emissions. We show that the influence of chimney density, a proxy for wood burning emissions, varies by regions and that a global model cannot be used to predict PM in regions that were not measured. Future work should consider using both survey data on wood burning intensity and information from numerical air quality forecast models, in LUR models, to improve the generalisation of the prediction of fine particulate levels.

Reading and writing data efficiently from storage systems is critical for high performance data-centric applications. These I/O systems are being increasingly characterized by complex topologies and deeper memory hierarchies. Effective parallel I/O solutions are needed to scale applications on current and future supercomputers. Data aggregation is an efficient approach consisting of electing some processes in charge of aggregating data from a set of neighbors and writing the aggregated data into storage. Thus, the bandwidth use can be optimized while the contention is reduced. In this work, we take into account the network topology for mapping aggregators and we propose an optimized buffering system in order to reduce the aggregation cost. We validate our approach using micro-benchmarks and the I/O kernel of a large-scale cosmology simulation. We show improvements up to 15× faster for I/O operations compared to a standard implementation of MPI I/O.