CacheFlow: Cache Optimizations for Data Driven Multithreading

2006 ◽  
Vol 16 (02) ◽  
pp. 229-244 ◽  
Author(s):  
Costas Kyriacou ◽  
Paraskevas Evripidou ◽  
Pedro Trancoso
Keyword(s):  
Data Driven ◽  
Management Policy ◽  
Conflict Avoidance ◽  
Cache Management ◽  
Short Term ◽  
Simulation Experiments ◽  
Long Latency ◽  
Latency Tolerance ◽  
Management Policies ◽  
Access Patterns

Data-Driven Multithreading is a non-blocking multithreading model of execution that provides effective latency tolerance by allowing the computation processor do useful work, while a long latency event is in progress. With the Data-Driven Multithreading model, a thread is scheduled for execution only if all of its inputs have been produced and placed in the processor's local memory. Data-driven sequencing leads to irregular memory access patterns that could affect negatively cache performance. Nevertheless, it enables the implementation of short-term optimal cache management policies. This paper presents the implementation of CacheFlow, an optimized cache management policy which eliminates the side effects due to the loss of locality caused by the data-driven sequencing, and reduces further cache misses. CacheFlow employs thread-based prefetching to preload data blocks of threads deemed executable. Simulation results, for nine scientific applications, on a 32-node Data-Driven Multithreaded machine show an average speedup improvement from 19.8 to 22.6. Two techniques to further improve the performance of CacheFlow, conflict avoidance and thread reordering, are proposed and tested. Simulation experiments have shown a speedup improvement of 24% and 32%, respectively. The average speedup for all applications on a 32-node machine with both optimizations is 26.1.

scholarly journals Science Informed Policies for Managing Water

Hydrology ◽  
2021 ◽  
Vol 8 (2) ◽  
pp. 66
Author(s):  
Daniel P. Loucks
Keyword(s):  
Water Management ◽  
Water Resource ◽  
Policy Making ◽  
Water Resource Management ◽  
Information Policy ◽  
Management Policy ◽  
Multiple Impacts ◽  
Policy Makers ◽  
Science Policy Interface ◽  
Management Policies

Water resource management policies impact how water supplies are protected, collected, stored, treated, distributed, and allocated among multiple users and purposes. Water resource policies influence the decisions made regarding the siting, design, and operation of infrastructure needed to achieve the underlying goals of these policies. Water management policies vary by region depending on particular hydrologic, economic, environmental, and social conditions, but in all cases they will have multiple impacts affecting these conditions. Science can provide estimates of various economic, ecologic, environmental, and even social impacts of alternative policies, impacts that determine how effective any particular policy may be. These impact estimates can be used to compare and evaluate alternative policies in the search for identifying the best ones to implement. Among all scientists providing inputs to policy making processes are analysts who develop and apply models that provide these estimated impacts and, possibly, their probabilities of occurrence. However, just producing them is not a guarantee that they will be considered by policy makers. This paper reviews various aspects of the science-policy interface and factors that can influence what information policy makers need from scientists. This paper suggests some ways scientists and analysts can contribute to and inform those making water management policy decisions. Brief descriptions of some water management policy making examples illustrate some successes and failures of science informing and influencing policy.

scholarly journals Data-Driven Study on the Sustainable Log Movements: Impact of Rail Car Fleet Size on Freight Storage and Car Idling

2020 ◽  
Vol 12 (11) ◽  
pp. 4563
Author(s):  
Sangpil Ko ◽  
Pasi Lautala ◽  
Kuilin Zhang
Keyword(s):  
Service Life ◽  
Seasonal Dynamics ◽  
Lake Superior ◽  
Data Driven ◽  
Short Term ◽  
Fleet Size ◽  
Interesting Observation ◽  
The Us ◽  
Term Storage ◽  
Scale Data

Rail car availability and the challenges associated with the seasonal dynamics of log movements have received growing attentions in the Lake Superior region of the US, as a portion of rail car fleet is close to reaching the end of its service life. This paper proposes a data-driven study on the rail car peaking issue to explore the fleet of rail cars dedicated to being used for log movements in the region, and to evaluate how the number of cars affects both the storage need at the sidings and the time the cars are idled. This study is based on the actual log scale data collected from a group of forest companies in cooperation with the Lake State Shippers Association (LSSA). The results of our analysis revealed that moving the current log volumes in the region would require approximately 400–600 dedicated and shared log cars in ideal conditions, depending on the specific month. While the higher fleet size could move the logs as they arrive to the siding, the lower end would nearly eliminate the idling of rail cars and enable stable volumes throughout the year. However, this would require short-term storage and additional handling of logs at the siding, both elements that increase the costs for shippers. Another interesting observation was the fact that the reduction of a single day in the loading/unloading process (2.5 to 1.5 days) would eliminate almost 100 cars (20%) of the fleet without reduction in throughput.

Short Term Injection Re-Distribution STIR: Real-Time Waterflood Optimization Technique Using Advanced Data Analytics

2021 ◽  
Author(s):  
Gaurav Modi ◽  
Manu Ujjwal ◽  
Srungeer Simha
Keyword(s):  
Real Time ◽  
Data Analytics ◽  
Optimization Technique ◽  
Material Balance ◽  
Oil Production ◽  
Data Driven ◽  
Petroleum Engineering ◽  
Short Term ◽  
Injection Water ◽  
Facility Level

Abstract Short Term Injection Re-distribution (STIR) is a python based real-time WaterFlood optimization technique for brownfield assets that uses advanced data analytics. The objective of this technique is to generate recommendations for injection water re-distribution to maximize oil production at the facility level. Even though this is a data driven technique, it is tightly bounded by Petroleum Engineering principles such as material balance etc. The workflow integrates and analyse short term data (last 3-6 months) at reservoir, wells and facility level. STIR workflow is divided into three modules: Injector-producer connectivity Injector efficiency Injection water optimization First module uses four major data types to estimate the connectivity between each injector-producer pair in the reservoir: Producers data (pressure, WC, GOR, salinity) Faults presence Subsurface distance Perforation similarity – layers and kh Second module uses connectivity and watercut data to establish the injector efficiency. Higher efficiency injectors contribute most to production while poor efficiency injectors contribute to water recycling. Third module has a mathematical optimizer to maximize the oil production by re-distributing the injection water amongst injectors while honoring the constraints at each node (well, facility etc.) of the production system. The STIR workflow has been applied to 6 reservoirs across different assets and an annual increase of 3-7% in oil production is predicted. Each recommendation is verified using an independent source of data and hence, the generated recommendations align very well with the reservoir understanding. The benefits of this technique can be seen in 3-6 months of implementation in terms of increased oil production and better support (pressure increase) to low watercut producers. The inherent flexibility in the workflow allows for easy replication in any Waterflooded Reservoir and works best when the injector well count in the reservoir is relatively high. Geological features are well represented in the workflow which is one of the unique functionalities of this technique. This method also generates producers bean-up and injector stimulation candidates opportunities. This low cost (no CAPEX) technique offers the advantages of conventional petroleum engineering techniques and Data driven approach. This technique provides a great alternative for WaterFlood management in brownfield where performing a reliable conventional analysis is challenging or at times impossible. STIR can be implemented in a reservoir from scratch in 3-6 weeks timeframe.

scholarly journals Fishing location choice and risk preferences among small fishers – Implications for fisheries management policies

2020 ◽  
Vol 15 (2) ◽  
pp. 140-156
Author(s):  
Riad Sultan ◽  
Keyword(s):  
Fisheries Management ◽  
Risk Preferences ◽  
Random Parameter ◽  
Management Policy ◽  
Linear Utility ◽  
South West Coast ◽  
Random Parameter Logit ◽  
Management Policies ◽  
Expected Revenue ◽  
Random Parameter Logit Model

The study provides evidence for how risk preferences determine fishing location choices by artisanal fishers on the south-west coast of the island of Mauritius. Risk preference is modelled using a random linear utility framework defined over mean-standard deviation space. The study estimates expected revenue and revenue risk from the Just and Pope production function and applies the random parameter logit model to account for fisher-specific and location-specific characteristics. The findings are consistent with utility-maximising fishers, whereby the likelihood to choose a fishing location is positively associated with expected revenue and negatively related to revenue risk. Distance from fishing station to fishing grounds affects the choice of fishing location negatively. The estimated model allows heterogeneity in risk preferences and concludes that 51% of fishers can be classified as risk averse, 31% as risk seekers and the remaining as risk neutral. The study also estimates the degree of substitutability and complementarity between fishing locations based on the risk preferences of fishers and discusses the relevance of this for fisheries management policy.

scholarly journals The Forecasting of Groundwater Fluctuations using Time Series Analysis and Combination of Data-Driven Models

2021 ◽  
Author(s):  
Amirhossein Najafabadipour ◽  
Gholamreza Kamali ◽  
Hossein Nezamabadi-pour
Keyword(s):  
Iron Ore ◽  
Least Squares Method ◽  
Moving Average ◽  
Groundwater Resources ◽  
Mining Area ◽  
Data Driven ◽  
Short Term ◽  
Iron Ore Mine ◽  
Groundwater Fluctuations ◽  
Groundwater Drainage

The Forecasting of Groundwater Fluctuations is a useful tool for managing groundwater resources in the mining area. Water resources management requires identifying potential periods for groundwater drainage to prevent groundwater from entering the mine pit and imposing high costs. In this research, Auto-Regressive Integrated Moving Average (ARIMA) and Holt-Winters Exponential Smoothing (HWES) data-driven models were used for short-term modeling of the groundwater fluctuations in a piezometer around the Gohar Zamin Iron Ore Mine. For this purpose, 250 non-seasonal groundwater fluctuations data in the period 22-Nov-2018 to 29-Jul-2019, 200 data for modeling, and 50 data for prediction were used. To take advantage of all the features of the two developed models, the predictions are combined with different methods and specific weights. The results show better accuracy for the ARIMA method between the two short-term forecasts, while the HWES method requires less time for modeling. Also, among all the predictions made, the highest accuracy for the combined least-squares method is for forecasting the groundwater fluctuations in the short-term. All the forecasts show a decrease in the groundwater fluctuations, indicating pumping wells around the Gohar Zamin Iron Ore Mine area.

scholarly journals Data-driven predictions of a multiscale Lorenz 96 chaotic system using machine-learning methods: reservoir computing, artificial neural network, and long short-term memory network

2020 ◽  
Vol 27 (3) ◽  
pp. 373-389 ◽  
Author(s):  
Ashesh Chattopadhyay ◽  
Pedram Hassanzadeh ◽  
Devika Subramanian
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Short Term Memory ◽  
Data Driven ◽  
Reservoir Computing ◽  
Short Term ◽  
Term Memory ◽  
Machine Learning Methods ◽  
Long Short Term Memory ◽  
Lorenz 96

Abstract. In this paper, the performance of three machine-learning methods for predicting short-term evolution and for reproducing the long-term statistics of a multiscale spatiotemporal Lorenz 96 system is examined. The methods are an echo state network (ESN, which is a type of reservoir computing; hereafter RC–ESN), a deep feed-forward artificial neural network (ANN), and a recurrent neural network (RNN) with long short-term memory (LSTM; hereafter RNN–LSTM). This Lorenz 96 system has three tiers of nonlinearly interacting variables representing slow/large-scale (X), intermediate (Y), and fast/small-scale (Z) processes. For training or testing, only X is available; Y and Z are never known or used. We show that RC–ESN substantially outperforms ANN and RNN–LSTM for short-term predictions, e.g., accurately forecasting the chaotic trajectories for hundreds of numerical solver's time steps equivalent to several Lyapunov timescales. The RNN–LSTM outperforms ANN, and both methods show some prediction skills too. Furthermore, even after losing the trajectory, data predicted by RC–ESN and RNN–LSTM have probability density functions (pdf's) that closely match the true pdf – even at the tails. The pdf of the data predicted using ANN, however, deviates from the true pdf. Implications, caveats, and applications to data-driven and data-assisted surrogate modeling of complex nonlinear dynamical systems, such as weather and climate, are discussed.

Sign in / Sign up

Export Citation Format

Share Document