Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Launch Supplement

Dalhousie University

$12,500.00

Irregular Parallelism in Symbolic and Scientific Computation

EF 2024-25

Résumé
Solving systems of polynomial equations (e.g. x2 + y2 = 4) is a fundamental problem in scientific computing. Symbolic methods for solving systems give exact, complete, and compact descriptions of the solutions. This contrasts with numerical methods which produce approximate solutions. Solutions produced by symbolic methods are preferred, or even required, in application domains such as cryptography, optimization, and theoretical physics. Modern workstation computers with powerful multicore processors and large quantities of main memory make these symbolic methods feasible. Yet, for these methods to be truly practical and high-performance, care is required to design software to explicitly and effectively exploit this modern hardware. Unfortunately, symbolic methods generally belong to the class of so-called irregular parallel applications. Irregular parallelism describes when the various sub-tasks of a problem require very different amounts of effort to be completed. Moreover, opportunities to process sub-tasks concurrently cannot be determined in advance, and load balancing and scheduling are a challenge. To make matters worse, symbolic computations are often limited by memory requirements and not just computational power. The research theme of this Discovery program is software performance engineering for irregular and symbolic computation. The long-term vision of this program is to produce a body of knowledge, and a collection of hardware-aware algorithms, models, and tools, to better execute irregular and symbolic applications on modern and near-future computer hardware. Over the next 5 years, our specific short-term objectives are: Objective 1. Advance algorithmic and implementation techniques for solving systems of polynomial equations. Objective 2. Establish a computational model and concurrency platform for irregular parallelism, focusing on dynamic memory requirements. Objective 3. Advances in automatic parallel program parameter optimization. Moore's law, which states that the number of transistors on an integrated circuit will double every two years, is coming to an end. One solution proposed by hardware manufacturers is to increase the number of cores on a multicore processor. But, effectively using those cores is a challenge. It has been observed that the limiting factor on computer performance will be software's ability to effectively make use of these cores and near-future hardware. The outcome of this research will be algorithms, techniques, and best practices which enable continued improvements in computing performance on highly parallel workstation computers despite the end of Moore's law. This research will directly support irregular and symbolic computation. It will also support applications in data processing, modelling and simulation, graphics and video games, and computer-aided design and manufacturing. Scientists, engineers, and content creators will be able to tackle larger and more complex workloads than ever before.
Programme
Discovery Launch Supplement
Bénéficiaire
Emplacement
Nova Scotia, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
DGECR-2024-00097
Name nom
Brandt, Alexander
Department département
Faculty of Computer Science
Organization id
79
Institution établissement
Dalhousie University
Province en
Nova Scotia
Province fr
Nouvelle-Écosse
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
12500
Program id
DGECR
Program name en
Discovery Launch Supplement
Program name fr
Supplément Tremplin vers la découverte
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1507
Committee name en
Computer Science
Committee name fr
Sciences informatiques
Area of application code
801
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Computer software
Area of application fr
Logiciel
Research subject code
2704
Research subject group en
Information technology
Research subject group fr
Technologie de l'information
Research subject en
Distributed and parallel processing
Research subject fr
Traitement réparti et simultané
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Irregular Parallelism in Symbolic and Scientific Computation
Keywords
symbolic computation; triangular decomposition; non-linear systems; irregular parallelism; parallelism and concurrency; parallel patterns; parallel models and algorithms; scientific computing; high-performance computing; performance portability
Application summary
Solving systems of polynomial equations (e.g. x2 + y2 = 4) is a fundamental problem in scientific computing. Symbolic methods for solving systems give exact, complete, and compact descriptions of the solutions. This contrasts with numerical methods which produce approximate solutions. Solutions produced by symbolic methods are preferred, or even required, in application domains such as cryptography, optimization, and theoretical physics. Modern workstation computers with powerful multicore processors and large quantities of main memory make these symbolic methods feasible. Yet, for these methods to be truly practical and high-performance, care is required to design software to explicitly and effectively exploit this modern hardware. Unfortunately, symbolic methods generally belong to the class of so-called irregular parallel applications. Irregular parallelism describes when the various sub-tasks of a problem require very different amounts of effort to be completed. Moreover, opportunities to process sub-tasks concurrently cannot be determined in advance, and load balancing and scheduling are a challenge. To make matters worse, symbolic computations are often limited by memory requirements and not just computational power. The research theme of this Discovery program is software performance engineering for irregular and symbolic computation. The long-term vision of this program is to produce a body of knowledge, and a collection of hardware-aware algorithms, models, and tools, to better execute irregular and symbolic applications on modern and near-future computer hardware. Over the next 5 years, our specific short-term objectives are: Objective 1. Advance algorithmic and implementation techniques for solving systems of polynomial equations. Objective 2. Establish a computational model and concurrency platform for irregular parallelism, focusing on dynamic memory requirements. Objective 3. Advances in automatic parallel program parameter optimization. Moore's law, which states that the number of transistors on an integrated circuit will double every two years, is coming to an end. One solution proposed by hardware manufacturers is to increase the number of cores on a multicore processor. But, effectively using those cores is a challenge. It has been observed that the limiting factor on computer performance will be software's ability to effectively make use of these cores and near-future hardware. The outcome of this research will be algorithms, techniques, and best practices which enable continued improvements in computing performance on highly parallel workstation computers despite the end of Moore's law. This research will directly support irregular and symbolic computation. It will also support applications in data processing, modelling and simulation, graphics and video games, and computer-aided design and manufacturing. Scientists, engineers, and content creators will be able to tackle larger and more complex workloads than ever before.