Verified tensor-program optimization via high-level scheduling rewrites

We present a lightweight Coq framework for optimizing tensor kernels written in a pure, functional array language. Optimizations rely on user scheduling using series of verified, semantics-preserving rewrites. Unusually for compilation targeting imperative code with arrays and nested loops, all rewrites are source-to-source within a purely functional language. Our language comprises a set of core constructs for expressing high-level computation detail and a set of what we call reshape operators, which can be derived from core constructs but trigger low-level decisions about storage patterns and ordering. We demonstrate that not only is this system capable of deriving the optimizations of existing state-of-the-art languages like Halide and generating comparably performant code, it is also able to schedule a family of useful program transformations beyond what is reachable in Halide.

Quadrant Mapping Artificial Lift Concept

This paper outlines a concept for monitoring performance of artificial lift performance such as electrical submersible pump (ESP), hydraulic pumping unit (HPU), sucker rod pump (SRP) and progressive cavity pump (PCP), for a large number of wells. The objective is to generate simplified monitoring performance of artificial lift with a huge number of wells on one page by creating quadrant mapping consisting of two coordinates with x axis representing pump efficiency and y axis showing pump submergence. We made a four-quadrant limit by pump efficiency (50%) and submergence (200 m). Optimum wells will show on range pump efficiency above 50% and submergence below 200 m, and 3 other quadrants are classified as artificial lift problems, well potential and sizing/design problems. By using the quadrant mapping concept, we can generate performance of artificial lift for 1500++ wells in one page, and this mapping consists of four quadrants (quadrant 1, quadrant 2, quadrant 3 and quadrant 4), quadrant 1 (Submergence above 200 meter and lifting efficiency below 50%) showing wells which have artificial lift problem, quadrant 2 (Submergence is above 200 meters and efficiency is above 50%) showing well which have potential to increased production, quadrant 3 (Submergence is below 200 meters and efficiency is above 50%) showing the optimum wells operation and quadrant 4 (Submergence is below 200 meters and efficiency is below 50%) showing the wells which required to re-sizing/re-design artificial lift. This quadrant mapping can be shown to Engineers, manager's and shareholder to show overall performance and classification detailed problems to create a troubleshooting, optimization program to increased oil production, run life artificial and result in better production performance. This mapping also helps petroleum engineers to get a better field view and create priorities and program optimization based on the quadrant mapping result and classification.

Data Pattern Aware Reliability Enhancement Scheme for 3D Solid-State Drives

3D charge-trap (CT) NAND flash-based SSD has been used widely for its large capacity, low cost per bit, and high endurance. One-shot program (OSP) scheme, as a variation of incremental step pulse programming (ISPP) scheme, has been employed to program data for CT flash, whose program unit is the Word-Line (WL) instead of the page. The existing program optimization schemes either make trade-offs among program latency and reliability by adjusting the program step voltage on demand; or remap the most error-prone cell states to others by re-encoding programmed data. However, the data pattern, which represents the ratio of 1s in data values, has not been thoroughly studied. In this paper, we observe that most small files do not contain uniform 1s and 0s among these common file types (i.e., image, audio, text, executable file), leading to programming WL cells in different states unevenly. Some cell states dominate over the WL, while others are not. Based on this observation, we propose a flexible reliability enhancement scheme based on the OSP scheme. This scheme programs the cells into different states with varied , i.e., these cells in one state, whose number is the largest in one WL, are programmed with a fine-grained (namely slow write). In contrast, the minority are programmed with a coarse-grained (namely fast write). So the reliability is improved due to averaging the major enhanced cells with the minor degraded cells without program latency overhead. A series of experiments have been conducted, and the results indicate that the proposed scheme achieves 34% read performance improvement and 16% lifetime elongation on average.

Impact of Transitioning to Virtual Delivery of a Cardiovascular Health Improvement Program for Latinos During the COVID-19 Pandemic

Background: Community Heart Health Actions for Latinos at Risk (CHARLAR) is a promotora-led cardiovascular disease (CVD) risk-reduction program for socio-demographically disadvantaged Latinos and consists of 11 skill-building sessions. The COVID-19 pandemic has led to worsening health status in U.S. adults and necessitated transition to virtual implementation of the CHARLAR program. Methods: A mixed-methods approach was used to evaluate virtual delivery of CHARLAR. Changes in health behaviors were assessed through a pre/post program survey. Results from virtual and historical (in-person delivery) were compared. Key informant interviews were conducted with promotoras and randomly selected participants and then coded and analyzed using a thematic approach.Results: An increase in days of exercise per week (+1.52), daily servings of fruit (+0.60) and vegetables (+0.56), and self-reported general health (+0.38), were observed in the virtual cohort [all p<0.05]. A numeric decrease in PHQ-8 (-1.07 p= .067) was also noted. The historical cohort showed similar improvements from baseline in days of exercise per week (+0.91), daily servings of fruit (+0.244) and vegetables (+0.282), and PHQ-8 (-1.89) [all p<0.05]. Qualitative interviews revealed that the online format provided valuable tools supporting positive behavior change. Despite initial discomfort and technical challenges, promotoras and participants adapted and deepened valued relationships through additional virtual support. Conclusion: Improved health behaviors and CVD risk factors were successfully maintained through virtual delivery of the CHARLAR program. Optimization of virtual health programs like CHARLAR has the potential to increase reach and improve CVD risk among vulnerable Latinos.

EvoStencils: a grammar-based genetic programming approach for constructing efficient geometric multigrid methods

For many systems of linear equations that arise from the discretization of partial differential equations, the construction of an efficient multigrid solver is challenging. Here we present EvoStencils, a novel approach for optimizing geometric multigrid methods with grammar-guided genetic programming, a stochastic program optimization technique inspired by the principle of natural evolution. A multigrid solver is represented as a tree of mathematical expressions that we generate based on a formal grammar. The quality of each solver is evaluated in terms of convergence and compute performance by automatically generating an optimized implementation using code generation that is then executed on the target platform to measure all relevant performance metrics. Based on this, a multi-objective optimization is performed using a non-dominated sorting-based selection. To evaluate a large number of solvers in parallel, they are distributed to multiple compute nodes. We demonstrate the effectiveness of our implementation by constructing geometric multigrid solvers that are able to outperform hand-crafted methods for Poisson’s equation and a linear elastic boundary value problem with up to 16 million unknowns on multi-core processors with Ivy Bridge and Broadwell microarchitecture.

Development of method for road network management program optimization

The object of research is the processes of road network management. The main hypothesis of the study is the application of optimization methods to the road asset management program, which allows to implement a road maintenance strategy, which includes a set of measures involved in ensuring and maintaining roads at the appropriate operational level. In general, the sequence of asset management begins with the definition of objectives, policy of road administrations and the available budget. According to this starting point, data collection, performance monitoring, analysis of options and optimization programs, selection and implementation of a set of projects are performed. The goals and objectives of road project management depending on the category are considered. The model of strategic planning of the program on management of a network of highways is offered and the modern architecture of model is constructed. The main approaches to the optimization of the road asset management program are identified. The target function for the optimization model is proposed, which is to minimize the cost of maintenance of the road network and defines the goals to be achieved in the optimization process. The optimization model is a mathematical description designed to compare alternative strategies and identify the relative benefits of each strategy according to the assigned decision criteria, such as safety, cost, environment and others. The main idea and goal of road network management projects is to improve maintenance practices and policies and thus improve the condition of the network, or maintain it at the current level, taking into account resource constraints. The results of program optimization have practical value and can be used to manage road repair and operation projects depending on: – defined project goals and plans, including scheduling, budgeting, planning, established efficiency requirements and selection of project participants; – maximizing the efficient use of resources; – implementation of functions of coordination and control, planning, design, evaluation, conclusion of contracts and implementation of projects under contracts.

The Model for Supporting Decision-Making During Designing Programs of Innovation Development at Enterprises of Electro-Technical Industry of Machine-Building

Electro-technical industry takes a specific place in economy of the Russian Federation. It is connected with its serious contribution to building-up favorable conditions for developing actually all industries of economy, social and household fields, comfortable living and working conditions of people. Provision of sustainability of electro-technical industry functioning is an essential foundation for power security of the country. At the same time sustainability and competitiveness of these enterprises depend on the quality of programs for their innovation development. The goal of the research is to work out the mathematic model for supporting decision-making during designing programs of innovation development at electro-technical enterprises. Cost minimization is a criterion of decision optimality in the model. One specific feature of the model is taking into account the uncertainty in estimation of possible costs connected with development and implementation of projects, included in the program. To resolve the problem the author put forward the iteration algorithm of successive shaping of expedient version of the program that takes into consideration specific features of the model. The model and algorithm can serve as a foundation to design concrete methods of innovation development program optimization. Introduction of such methodology could provide an opportunity to raise the quality of innovation development management at enterprises of electrotechnical industry of machine-building.