A5087200957- Cardiac electrophysiology and arrhythmias
- Model Reduction and Neural Networks
- Numerical methods for differential equations
- Fluid Dynamics and Turbulent Flows
- Cryospheric studies and observations
- Advanced Mathematical Modeling in Engineering
- Climate change and permafrost
- Advanced MRI Techniques and Applications
- Simulation Techniques and Applications
- Analog and Mixed-Signal Circuit Design
- Electron Spin Resonance Studies
- Scientific Computing and Data Management
- Flood Risk Assessment and Management
- ECG Monitoring and Analysis
- Software System Performance and Reliability
- Hydrology and Watershed Management Studies
- Rheology and Fluid Dynamics Studies
- Soil and Unsaturated Flow
- Distributed and Parallel Computing Systems
- Hydrological Forecasting Using AI
- Advanced Numerical Methods in Computational Mathematics
- Fault Detection and Control Systems
- Matrix Theory and Algorithms
University of Saskatchewan
2019-2023
Abstract The intent of this paper is to encourage improved numerical implementation land models. Our contributions in are two-fold. First, we present a unified framework formulate and implement model equations. We separate the representation physical processes from their solution, enabling use established robust methods solve Second, introduce set synthetic test cases (the laugh tests) evaluate include storage transmission water soils, lateral sub-surface flow, coupled hydrological...
The speed of mathematical function evaluations can significantly contribute to the overall performance numerical simulations. Two common approaches evaluate a are by direct evaluation or means lookup tables precomputed values. Direct is standard approach, and in graphics applications field-programmable gate array computing. We address usage versus for evaluating general univariate functions on general-purpose CPUs large-scale simulation. introduce small C++ library called FunC (for Function...
BACOLI is a Fortran software package for solving one-dimensional parabolic partial differential equations (PDEs) with separated boundary conditions by B-spline adaptive collocation methods. A distinguishing feature of its ability to estimate and control error correspondingly adapt meshes in both space time. Many models scientific interest, however, can be formulated as multiscale PDE systems, that is, couple system PDEs describing dynamics on global scale ordinary local scale. This article...
Implicit methods for the numerical solution of initial-value problems may admit multiple solutions at any given time step. Accordingly, their nonlinear solvers converge to these solutions. Below a critical timestep, exactly one (the consistent solution) occurs on branch principal branch) that can be continuously and monotonically continued back zero timestep. Standard step-size control promote convergence by adjusting timestep maintain an error estimate below tolerance. However, simulations...
Large-scale computations are a critical part of scientific discovery. The demand for such is generally met in two ways. Traditionally, the performed on high-performance computing clusters where hardware (compute nodes and networking) tightly controlled (the super-computing paradigm). More recently, however, heterogeneous collections consumer devices as desktop computers or mobile phones an ad hoc basis volunteer In this paper, we aim to establish middle ground between these paradigms by...
Large simulations with many sub-simulations are common in scientific computing. There a number of challenges, however, associated performing such shared computing environments. For example, may have completion times that vary by orders magnitude, or they not complete at all for given set parameters, leading to unpredictable runtimes and hence unbalanced inefficient use computational resources. In this study, we investigate the actor model concurrent improve both resource utilization fault...