An MPI implementation supported by process migration and load balancing

This report describes an implementation of MPI-1 on the GENESIS cluster operating system and compares this implementation to a UNIX based MPI implementation. The changes that were made to the implementation are compared between the two, and the advantages of porting to GENESIS are detailed. This report demonstrates how GENESIS load balancing supported by process migration improves the execution performance of an MPI application. The significance of this report is in demonstrating how these services can enhance parallel programming tools to improve performance and how future parallel programming tool design could take advantage of these services.

Load balance algorithms for anycast

Increasingly, replicated anycast servers are being used to deliver network applications and service ever increasing user requests. Therefore, the strategies used to guarantee network bandwidth prerequisites and perform load balancing across the nodes of an anycast group are critical to the performance of online applications. In this paper, we model user requests, network congestion and latency, and server load using a combination of hydro-dynamics and queuing theory to develop an efficient job distribution strategy. Current, anycast research does not explicitly consider the system load of nodes within an anycast groups when distributing requests. Therefore, the performance of a heavily loaded anycast system can quickly become congested and uneven as jobs are routed to closely linked nodes which are already saturated with requests. In comparison, the nodes of further away systems remain relatively unused because of other issues such as network bandwidth and latency during these times. Our system redirects requests from busy systems to the idle, remotely linked nodes, to process requests faster in spite of slower network access. Using an empirical study, we show this technique can improve request performance, and throughput with minimal network probing overhead.

Adaptive load shedding for mining frequent patterns from data streams

Most algorithms that focus on discovering frequent patterns from data streams assumed that the machinery is capable of managing all the incoming transactions without any delay; or without the need to drop transactions. However, this assumption is often impractical due to the inherent characteristics of data stream environments. Especially under high load conditions, there is often a shortage of system resources to process the incoming transactions. This causes unwanted latencies that in turn, affects the applicability of the data mining models produced – which often has a small window of opportunity. We propose a load shedding algorithm to address this issue. The algorithm adaptively detects overload situations and drops transactions from data streams using a probabilistic model. We tested our algorithm on both synthetic and real-life datasets to verify the feasibility of our algorithm.

Varicella infection : evidence for peak activity in summer months

This study aimed to identify seasonal variations in the presentation of primary varicella infection in susceptible patients in Victoria (a temperate region in south-east Australia) by analysing Victorian hospital admissions data and medical locum service data from the years preceding the introduction of a universally offered vaccination, complemented by available surveillance data from the neighbouring state, of South Australia. Contrary to the conventional assumption, which is based on observations in temperate regions elsewhere in the world, we found no consistent evidence of seasonal peaks during late winter and early spring for varicella infection in Victoria. This finding may have implications for prevention in temperate regions elsewhere in the world wherever estimations of local seasonal trends have been based on international experiences.

Load balancing based concurrent execution of NAS parallel benchmarks with BYTE sequential benchmarks on a non-dedicated cluster

Dedicated clusters are becoming commonly used for high performance parallel processing. Computers of a non-dedicated cluster are often idle or lightly loaded. These under utilised computers can be employed to execute parallel applications. Thus, they have to be shared by parallel and sequential applications, which could lead to the improvement of their execution performance. There is a lack of experimental study showing the behaviour and performance of executing parallel and sequential applications concurrently on a non-dedicated cluster. We present the result of an experimental study into load balancing of a mixture of parallel and sequential applications on a non-dedicated cluster.

Effect of hydrogen/air flow rates and scan rate on the flame ionisation detection response of phopholipids, and their quantitative and qualitative analysis by Iatroscan (Mark‐6s) TLC‐FID

The combined effect of scan speed, hydrogen and air flow rates on the flame ionization detection (FID) peak response of phospholipid classes has been studied to determine the optimum levels of these parameters. The phospholipid composition of different types of commercial lecithins, as well as lecithins combined with fish oils, has been analyzed by Iatroscan TLC‐FID Mark‐6s under optimized conditions. An air flow rate of 2 L/min, a hydrogen flow rate of 150–160 mL/min, and a scan speed of 30 s/rod seem to be the ideal conditions for scanning phospholipids with complete pyrolysis in the flame in the Mark‐6 model. Increasing the scan speed rapidly decreased the FID response. A hydrogen flow rate as high as 170 mL/min could be used at relatively low air flow rates (&#x003C2 L/min) and the response declined when both air flow rate and hydrogen flow rate increased simultaneously. Both linear and curvilinear relationships had highly significant correlations (p&#x003C0.01) with the sample load. Time course reactions, including the hydrolysis of phosphatidylserine using enzymes, can be successfully monitored by the Iatroscan TLC‐FID Chromarod system.

A comparison of methods to calculate the optimal load for maximal power output in the power clean

The aim of this study was to compare three calculation methods to determine the load that maximises power output in the power clean. Five male athletes (height=179.8 10.5cms, weight 91 .8 8.8kg, power dean 1RM = 117.0 20.5kg) performed two per cleans at 10% increments from 50% to 100% of 1RM. Bar displacement data was collected using a Ballistic Measurement System (BMS) and vertical ground reaction force (VGRF) data was measured by a Kistler 9287B Force Plate. Power output was calculated for BMS (system mass), BMS (bar mass) and VGRF/BMS system mass. Optimal load was determined to be 70% for the BMS (system mass) and VGRF BMS (system mass) methods and 90% for the BMS (bar mass) method. Sports scientists should be aware of the technical issues underlying these findings due to the practical ramifications for athlete testing and training.

Is tendon pathology a continuum? A pathology model to explain the clinical presentation of load-induced tendinopathy

Overuse tendinopathy is problematic to manage clinically. People of different ages with tendons under diverse loads present with varying degrees of pain, irritability, and capacity to function. Recovery is similarly variable; some tendons recover with simple interventions, some remain resistant to all treatments.

The pathology of tendinopathy has been described as degenerative or failed healing. Neither of these descriptions fully explains the heterogeneity of presentation. This review proposes, and provides evidence for, a continuum of pathology. This model of pathology allows rational placement of treatments along the continuum.

A new model of tendinopathy and thoughtful treatment implementation may improve outcomes for those with tendinopathy. This model is presented for evaluation by clinicians and researchers.

Effect of load pattern on solar-boosted heat pump water heater performance

The performance of a solar-boosted heat pump water heater (HPWH) operating under full load and part load conditions was determined in an outdoor experimental study. The system utilised flat unglazed aluminium solar evaporator panels to absorb solar and ambient energy. Absorbed energy was transferred to the water tank by means of the heat pump and a wrap around condenser coil on the outside of the tank. The system COP was found to be in the range of 5–7 under clear daytime conditions and 3–5 under clear night-time conditions. Using part load testing of the HPWH system it was found that concentrating the coils in the lower portion of the tank could increase the efficiency of the condenser coil. It was also shown that there exists a generalised linear relationship that can be used to describe the system COP in terms of the temperature difference between the water in the storage tank and the ambient air.

Similar metabolic adaptations during exercise after low volume sprint interval and traditional endurance training in humans

Low-volume ‘sprint’ interval training (SIT) stimulates rapid improvements in muscle oxidative capacity that are comparable to levels reached following traditional endurance training (ET) but no study has examined metabolic adaptations during exercise after these different training strategies. We hypothesized that SIT and ET would induce similar adaptations in markers of skeletal muscle carbohydrate (CHO) and lipid metabolism and metabolic control during exercise despite large differences in training volume and time commitment. Active but untrained subjects (23 ± 1 years) performed a constant-load cycling challenge (1 h at 65% of peak oxygen uptake before and after 6 weeks of either SIT or ET (n = 5 men and 5 women per group). SIT consisted of four to six repeats of a 30 s ‘all out’ Wingate Test (mean power output ∼500 W) with 4.5 min recovery between repeats, 3 days per week. ET consisted of 40–60 min of continuous cycling at a workload that elicited ∼65% (mean power output ∼150 W) per day, 5 days per week. Weekly time commitment (∼1.5 versus ∼4.5 h) and total training volume (∼225 versus ∼2250 kJ week−1) were substantially lower in SIT versus ET. Despite these differences, both protocols induced similar increases (P < 0.05) in mitochondrial markers for skeletal muscle CHO (pyruvate dehydrogenase E1α protein content) and lipid oxidation (3-hydroxyacyl CoA dehydrogenase maximal activity) and protein content of peroxisome proliferator-activated receptor-γ coactivator-1α. Glycogen and phosphocreatine utilization during exercise were reduced after training, and calculated rates of whole-body CHO and lipid oxidation were decreased and increased, respectively, with no differences between groups (all main effects, P < 0.05). Given the markedly lower training volume in the SIT group, these data suggest that high-intensity interval training is a time-efficient strategy to increase skeletal muscle oxidative capacity and induce specific metabolic adaptations during exercise that are comparable to traditional ET.