Modeling, Simulation and Optimization of Complex Processes, 2012
Pricing and hedging of higher order derivatives such as multidimensional (up to 100 underlying as... more Pricing and hedging of higher order derivatives such as multidimensional (up to 100 underlying assets) European and first generation exotic options represent mathematically complex and computationally intensive problems. Grid computing promises to give the capability to handle such intense computations. With several Grid middleware solutions available for gridifying traditional applications, it is cumbersome to select an ideal candidate, to develop financial applications, that can cope up with time critical computational demand for complex pricing requests. In this paper we present SuperQuant Financial Benchmark Suite to evaluate and quantify the overhead imposed by a Grid middleware on throughput of the system and turnaround times for computation. This approach is a step towards producing a middleware independent, reproducible, comparable, self-sufficient and fair performance analysis of Grid middlewares. The result of such performance analysis can be used by middleware vendors to find the bottlenecks and problems in their design and implementation of the system and by financial application developers to verify implementation of their financial algorithms. In this paper we explain the motivation and the details of the proposed benchmark suite. As a proof of concept, we utilize the benchmarks in an International Grid Programming contest and demonstrate the result of initial experiments.
2006 Second IEEE International Conference on e-Science and Grid Computing (e-Science'06), 2006
This paper introduces a Grid software architecture offering fault tolerance, dynamic and aggressi... more This paper introduces a Grid software architecture offering fault tolerance, dynamic and aggressive load balancing and two complementary parallel programming paradigms. Experiments with financial applications on a real multi-site Grid assess this solution. This architecture has been designed to run industrial and financial applications, that are frequently time constrained and CPU consuming, feature both tightly and loosely coupled parallelism requiring generic programming paradigm, and adopt client-server business architecture.
Computing in financial services consists of a combination of time-critical computations completed... more Computing in financial services consists of a combination of time-critical computations completed during trading hours, such as Monte Carlo simulations for option pricing, and overnight calculations on massive data sets, such as those required for market risk measurement. To date, this has typically been done using traditional parallel or cluster computing techniques. The French National Research Agency (ANR), along with several banks and financial software companies have partnered with INRIA to explore the application of grid computing to this domain. The PicsouGrid project utilizes the ProActive Java distributed computing library to parallelize and distribute Monte Carlo option pricing simulations, concurrently utilizing 102-103 workers. PicsouGrid has been deployed on various grid systems to evaluate its scalability and performance. Issues arising from the heterogeneity and layering of grid infrastructures are addressed via an abstract process model which is applied at each layer. Timings of both the algorithms and the grid infrastructures are carefully measured to provide better insight into the behavior and utilization of computational grids for this important class of parallel simulations.
2008 Workshop on High Performance Computational Finance, 2008
Among derivative financial contracts, the widely traded in the financial markets are the Bermudan... more Among derivative financial contracts, the widely traded in the financial markets are the Bermudan-American options. However, pricing high-dimensional Bermudan-American options is quite computationally intensive and using traditional computing infrastructures may take up to hours for these computations. This can result in potential financial losses, further weakening the competitiveness of an organization. Several parallel approaches for pricing have been practiced utilizing
Modeling, Simulation and Optimization of Complex Processes, 2008
In recent years, PC-based cluster has become a mainstream branch in high performance computing (H... more In recent years, PC-based cluster has become a mainstream branch in high performance computing (HPC) systems. Like other systems supporting HPC, one of the most important concerns in PC-based cluster is how to improve response time, throughput, and utilization. Therefore, scheduling can have a significant impact on performance characteristics of the system. This paper focuses on building an adaptive scheduling algorithm for Supernode II, a PC-based cluster built at HCMC University of Technology. Based on characteristics of the system and running jobs, the adaptive space-sharing scheme has many advantages compare to other solutions. The proposed scheduling algorithm for Supernode II is an improvement over other existing adaptive space-sharing solutions. The algorithm includes two main parts: a function to calculate number of processors allocated for jobs (partitioning-function) and rules to select jobs to be executed (job-selecting rules). The partitioning-function creates partitions for jobs adapting to the current system status and satisfying the requirement of jobs. In combination with the partitioning-function, job-selecting rules will help improving system utilization, reducing resource fragmentation, preventing indefinite postponement and getting a fair-share between jobs. Although the algorithm is specially built for Supernode II, we prove that it is suitable for other PC-based clusters with the similar characteristics.
In this paper we present two parallel Monte Carlo based algorithms for pricing multi-dimensional ... more In this paper we present two parallel Monte Carlo based algorithms for pricing multi-dimensional Bermudan/American options. First approach relies on computation of the optimal exercise boundary while the second relies on classification of continuation and exercise values. We also evaluate the performance of both the algorithms in a desktop grid environment. We show the effectiveness of the proposed approaches in a heterogeneous computing environment, and identify scalability constraints due to the algorithmic structure.
- This paper aims to provide an overview and a performance comparison of some parallel and distri... more - This paper aims to provide an overview and a performance comparison of some parallel and distributed algorithms for Bermudian-American option pricing. We use two Monte Carlo based methods to ad-dress such pricing in the case of a high number of assets ...
- This paper aims to provide an overview and a comparison performance of parallel and dis-tribute... more - This paper aims to provide an overview and a comparison performance of parallel and dis-tributed Bermudian-American option pricing. We use two Monte Carlo methods to address this pricing in the case of an increasing number of assets (high-dimension). Our work is ...
As experience with independent and embarassingly par-allel computations in a grid environment mat... more As experience with independent and embarassingly par-allel computations in a grid environment mature, it has become possible to explore parallel computing on grids with higher levels of inter-task communication. The Pic-souGrid project applies grid computing concepts to ...
Thèmes COM et NUM Systèmes communicants et Systèmes numériques Projets Oasis and Tosca Rapport te... more Thèmes COM et NUM Systèmes communicants et Systèmes numériques Projets Oasis and Tosca Rapport technique n°365 Version 2-June 2009 30 pages
From birth to health, surfactants play an essential role in our lives. Due to the importance, the... more From birth to health, surfactants play an essential role in our lives. Due to the importance, their environmental impacts are well understood. One of the aspects that has been extensively studied is their impact on bacteria, particularly on their motility. Here, we uncover an alternate chemotactic strategy triggered by surfactantsdiff usiophoresis. We show that even a trace amount of ionic surfactants, down to a single ppm level, can promote the bacterial diffusiophoresis by boosting the surface charge of the cells. Because diffusiophoresis is driven by the surface−solute interactions, surfactant-enhanced diffusiophoresis is observed regardless of the types of bacteria. Whether Gram-positive or-negative, flagellated or nonflagellated, the surfactants enable fast migration of freely suspended bacteria, suggesting a ubiquitous locomotion mechanism that has been largely overlooked. We also demonstrate the implication of surfactant-enhanced bacterial diffusiophoresis on the rapid formation of biofilms in flow networks, suggesting environmental and biomedical implications.
Modeling, Simulation and Optimization of Complex Processes, 2012
Pricing and hedging of higher order derivatives such as multidimensional (up to 100 underlying as... more Pricing and hedging of higher order derivatives such as multidimensional (up to 100 underlying assets) European and first generation exotic options represent mathematically complex and computationally intensive problems. Grid computing promises to give the capability to handle such intense computations. With several Grid middleware solutions available for gridifying traditional applications, it is cumbersome to select an ideal candidate, to develop financial applications, that can cope up with time critical computational demand for complex pricing requests. In this paper we present SuperQuant Financial Benchmark Suite to evaluate and quantify the overhead imposed by a Grid middleware on throughput of the system and turnaround times for computation. This approach is a step towards producing a middleware independent, reproducible, comparable, self-sufficient and fair performance analysis of Grid middlewares. The result of such performance analysis can be used by middleware vendors to find the bottlenecks and problems in their design and implementation of the system and by financial application developers to verify implementation of their financial algorithms. In this paper we explain the motivation and the details of the proposed benchmark suite. As a proof of concept, we utilize the benchmarks in an International Grid Programming contest and demonstrate the result of initial experiments.
2006 Second IEEE International Conference on e-Science and Grid Computing (e-Science'06), 2006
This paper introduces a Grid software architecture offering fault tolerance, dynamic and aggressi... more This paper introduces a Grid software architecture offering fault tolerance, dynamic and aggressive load balancing and two complementary parallel programming paradigms. Experiments with financial applications on a real multi-site Grid assess this solution. This architecture has been designed to run industrial and financial applications, that are frequently time constrained and CPU consuming, feature both tightly and loosely coupled parallelism requiring generic programming paradigm, and adopt client-server business architecture.
Computing in financial services consists of a combination of time-critical computations completed... more Computing in financial services consists of a combination of time-critical computations completed during trading hours, such as Monte Carlo simulations for option pricing, and overnight calculations on massive data sets, such as those required for market risk measurement. To date, this has typically been done using traditional parallel or cluster computing techniques. The French National Research Agency (ANR), along with several banks and financial software companies have partnered with INRIA to explore the application of grid computing to this domain. The PicsouGrid project utilizes the ProActive Java distributed computing library to parallelize and distribute Monte Carlo option pricing simulations, concurrently utilizing 102-103 workers. PicsouGrid has been deployed on various grid systems to evaluate its scalability and performance. Issues arising from the heterogeneity and layering of grid infrastructures are addressed via an abstract process model which is applied at each layer. Timings of both the algorithms and the grid infrastructures are carefully measured to provide better insight into the behavior and utilization of computational grids for this important class of parallel simulations.
2008 Workshop on High Performance Computational Finance, 2008
Among derivative financial contracts, the widely traded in the financial markets are the Bermudan... more Among derivative financial contracts, the widely traded in the financial markets are the Bermudan-American options. However, pricing high-dimensional Bermudan-American options is quite computationally intensive and using traditional computing infrastructures may take up to hours for these computations. This can result in potential financial losses, further weakening the competitiveness of an organization. Several parallel approaches for pricing have been practiced utilizing
Modeling, Simulation and Optimization of Complex Processes, 2008
In recent years, PC-based cluster has become a mainstream branch in high performance computing (H... more In recent years, PC-based cluster has become a mainstream branch in high performance computing (HPC) systems. Like other systems supporting HPC, one of the most important concerns in PC-based cluster is how to improve response time, throughput, and utilization. Therefore, scheduling can have a significant impact on performance characteristics of the system. This paper focuses on building an adaptive scheduling algorithm for Supernode II, a PC-based cluster built at HCMC University of Technology. Based on characteristics of the system and running jobs, the adaptive space-sharing scheme has many advantages compare to other solutions. The proposed scheduling algorithm for Supernode II is an improvement over other existing adaptive space-sharing solutions. The algorithm includes two main parts: a function to calculate number of processors allocated for jobs (partitioning-function) and rules to select jobs to be executed (job-selecting rules). The partitioning-function creates partitions for jobs adapting to the current system status and satisfying the requirement of jobs. In combination with the partitioning-function, job-selecting rules will help improving system utilization, reducing resource fragmentation, preventing indefinite postponement and getting a fair-share between jobs. Although the algorithm is specially built for Supernode II, we prove that it is suitable for other PC-based clusters with the similar characteristics.
In this paper we present two parallel Monte Carlo based algorithms for pricing multi-dimensional ... more In this paper we present two parallel Monte Carlo based algorithms for pricing multi-dimensional Bermudan/American options. First approach relies on computation of the optimal exercise boundary while the second relies on classification of continuation and exercise values. We also evaluate the performance of both the algorithms in a desktop grid environment. We show the effectiveness of the proposed approaches in a heterogeneous computing environment, and identify scalability constraints due to the algorithmic structure.
- This paper aims to provide an overview and a performance comparison of some parallel and distri... more - This paper aims to provide an overview and a performance comparison of some parallel and distributed algorithms for Bermudian-American option pricing. We use two Monte Carlo based methods to ad-dress such pricing in the case of a high number of assets ...
- This paper aims to provide an overview and a comparison performance of parallel and dis-tribute... more - This paper aims to provide an overview and a comparison performance of parallel and dis-tributed Bermudian-American option pricing. We use two Monte Carlo methods to address this pricing in the case of an increasing number of assets (high-dimension). Our work is ...
As experience with independent and embarassingly par-allel computations in a grid environment mat... more As experience with independent and embarassingly par-allel computations in a grid environment mature, it has become possible to explore parallel computing on grids with higher levels of inter-task communication. The Pic-souGrid project applies grid computing concepts to ...
Thèmes COM et NUM Systèmes communicants et Systèmes numériques Projets Oasis and Tosca Rapport te... more Thèmes COM et NUM Systèmes communicants et Systèmes numériques Projets Oasis and Tosca Rapport technique n°365 Version 2-June 2009 30 pages
From birth to health, surfactants play an essential role in our lives. Due to the importance, the... more From birth to health, surfactants play an essential role in our lives. Due to the importance, their environmental impacts are well understood. One of the aspects that has been extensively studied is their impact on bacteria, particularly on their motility. Here, we uncover an alternate chemotactic strategy triggered by surfactantsdiff usiophoresis. We show that even a trace amount of ionic surfactants, down to a single ppm level, can promote the bacterial diffusiophoresis by boosting the surface charge of the cells. Because diffusiophoresis is driven by the surface−solute interactions, surfactant-enhanced diffusiophoresis is observed regardless of the types of bacteria. Whether Gram-positive or-negative, flagellated or nonflagellated, the surfactants enable fast migration of freely suspended bacteria, suggesting a ubiquitous locomotion mechanism that has been largely overlooked. We also demonstrate the implication of surfactant-enhanced bacterial diffusiophoresis on the rapid formation of biofilms in flow networks, suggesting environmental and biomedical implications.
Uploads
Papers by Viết Sang Đoàn