WO2007067338A1 - Methods and systems for computing platform - Google Patents

Methods and systems for computing platform Download PDF

Info

Publication number
WO2007067338A1
WO2007067338A1 PCT/US2006/044902 US2006044902W WO2007067338A1 WO 2007067338 A1 WO2007067338 A1 WO 2007067338A1 US 2006044902 W US2006044902 W US 2006044902W WO 2007067338 A1 WO2007067338 A1 WO 2007067338A1
Authority
WO
WIPO (PCT)
Prior art keywords
analytical
operations
matrix
computer
processing units
Prior art date
Application number
PCT/US2006/044902
Other languages
French (fr)
Inventor
Kumar Metlapalli
Original Assignee
Kumar Metlapalli
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Kumar Metlapalli filed Critical Kumar Metlapalli
Priority to CA002631517A priority Critical patent/CA2631517A1/en
Priority to EP06838067A priority patent/EP1955191A1/en
Publication of WO2007067338A1 publication Critical patent/WO2007067338A1/en

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F15/00Digital computers in general; Data processing equipment in general
    • G06F15/76Architectures of general purpose stored program computers
    • G06F15/78Architectures of general purpose stored program computers comprising a single central processing unit
    • G06F15/7867Architectures of general purpose stored program computers comprising a single central processing unit with reconfigurable architecture
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/03Credit; Loans; Processing thereof

Definitions

  • the present invention relates generally to methods and systems for providing a unique and alternative computing platform operable to address the needs of high performance computing areas.
  • Financial industry is just one example of such a field where there exists such a need for high performance computing.
  • high performance problems are solved through computer clusters, scalable grids etc. whose costs increase exponentially as the number of units involved in the processing increase.
  • VaR Value at Risk
  • the present invention addresses the foregoing problems of the prior art.
  • the present invention provides a computer method and apparatus for analytical financial computing (computational processing) and more generally for a hardware accelerated numerical system and analysis.
  • the invention analytical computing apparatus comprises (a) one or more analytical processing units and (b) an application interface enabling use of respective desired ones of the analytical processing units by one or more computer application programs.
  • Different computer application programs utilize different ones of the analytical processing units, and each computer application program performs a respective analytical computation, hi accordance with one aspect of the present invention, each processing unit is formed of a respective field programmable gate (FPGA) array and performs a respective analytical operation.
  • FPGA field programmable gate
  • a computer system comprises a mother board hosting a multiplicity of processing cards.
  • Each card carries respective analytical processing units (FPGA's).
  • FPGA's analytical processing units
  • One subset of the FPGA's provides numerical algebra operations (e.g., matrix mathematics, eigen values, eigen vectors, etc.) and linear equations.
  • Another subset of the FPGA's provides statistical operations (e.g., random number generator, variance-covariance matrix construction, maximum likelihood estimator, etc.)
  • a third subset of the FPGA's provides real time matrix operations for data stream (continuous) and time rate data feeds (e.g., foreign exchange rate estimator, interest rate estimator, etc.)
  • the present invention facilitates high performance computing (ranging from ten to two hundred times performance improvement and approximately one-third of the "hard” costs for an equivalent grid computing solution) in three fundamental areas:
  • Fig. 1 is a schematic diagram showing an overview of a system according to the present invention.
  • Fig. 2 is a schematic diagram of a computer network in which embodiments of the present invention operate.
  • Fig. 3 is a block diagram of a computer node in the network of Fig. 2.
  • Fig. 4 is a block diagram of a computer system embodying the present invention. DETAILED DESCRIPTION OF THE INVENTION
  • the present invention overcomes the disadvantages of prior systems by using FPGA' s (Field-Programmable Gate Arrays) to solve financial industry analytical problems and related computational processing.
  • FPGA' s Field-Programmable Gate Arrays
  • An FPGA is a semiconductor device containing programmable logic components and programmable
  • the programmable logic components can be programmed to duplicate the functionality of basic logic gates (such as AND, OR, XOR, INVERT) or more complex combinatorial functions such as decoders or simple math functions. In most FPGA's these programmable logic components (or logic blocks, in FPGA parlance) also include memory elements, which may be simple flip-flops or more complete blocks of memories.
  • a hierarchy of programmable interconnects allows the logic blocks of an FPGA to be interconnected as needed by the system designer, somewhat like a one- chip programmable breadboard. These logic blocks and interconnects can be programmed after the manufacturing process by the customer/designer (hence the term "field-programmable") so that the FPGA can perform whatever logical function is needed.
  • FPGA's are commercially available, for example, from Altera Corporation of San Jose, California.
  • FPGA's are employed in a financial industry analytical computing or processing system 10 as illustrated in Fig. 1.
  • the illustrated system 10 is formed of (a) a PCI card or board 12 hosting one or more analytical processing units (or building blocks) 15; and (b) an application interface 14 built using any common programming language such as Java, C++, Jscript, etc., that interacts between industry applications 11 desired and the PCI card 12.
  • the analytical processing units or building blocks 15 are FPGA' s programmed to solve different financial industry problems of interest (e.g., analytical financial problems). In one embodiment, there are three categories or sets of analytical processing units/building blocks 15 implemented on the FPGA's using known techniques, processor routines and/or algorithms. One category or subset of FPGA building blocks 15 performs or otherwise implements linear algebra
  • building blocks/FPGA's 15 provide statistical operations such as random number generators, variance-covariance matrix construction and correlation matrix construction, maximum likelihood estimation and non-stationary correlator among a time series.
  • a third category or subset of building blocks/FPGA's 15 implements real-time matrix building.
  • linear algebra processing units 15 include FPGA's that respectively implement matrix addition, matrix subtraction, matrix transpose, matrix trace, matrix extraction (as to form or compute a minor, adjoin and cofactor), matrix multiplication (including inner product and direct/outer product), matrix determinant (indicating if singular), matrix inverse (if not singular), matrix decomposition (including solving eigen-values and eigen- vectors), matrix inertial and linear equation solvers.
  • a scalar is a 1 by 1 matrix (0-dimension tensor), and a vector is a 1 by N or N by 1 matrix (1 -dimension tensor) and a planar matrix has 2-dimension tensor.
  • the matrix multiplication preferably follows the Winograd or similar suitable method. In matrix decomposition, use of Jacobian matrixes is preferred and output includes the diagonal, an eigen value matrix and a matrix of the corresponding eigen vectors.
  • the decomposition may be performed using SVD (singular value decomposition), Cholesky's algorithm or the like.
  • SVD singular value decomposition
  • linear equation solvers preferably include implementation of
  • Gauss-Jordan algorithm Gauss elimination, Jordan elimination and various approximation algorithms (iterations, Seidal, relaxation and the like).
  • the random number generators generate an output in a matrix format with variables as rows and realizations as columns, or vice versa.
  • the exponential- weighted moving average algorithm is preferably used as follows.
  • the maximum likelihood estimator (processing unit 15) first estimates four parameters k, ⁇ , ⁇ , ⁇ using the general equilibrium model detailed below. Second, when all four parameters are decided, the maximum likelihood estimator starts iterations day by day to simulate future interest rate paths using a random matrix.
  • Input includes: (a) Historical interest rates (grid-level: 1 month, 3 month, 2 year, 5 year, 10 year) is a G x N matrix where G is the number of grid levels and N is the number of historical rates. This is for purposes of illustration, not limitation; G could be larger.
  • r l+1 ⁇ + k( ⁇ - ⁇ )At + ⁇ ifz, yf ⁇ l, (2) where one needs to evaluate the four parameters k, ⁇ , ⁇ , ⁇ , using historical data and maximum likelihood estimation.
  • the correlated random number matrix z is a linear combination of standard normal variables, when considering the correlations among different term interest rates. That is, r should be a vector when representing a term structure, and each term has its own four parameters of k, ⁇ , ⁇ , ⁇ .
  • Z is calculated as U ⁇ * SQRT(B) * Normal distribution, i.e., correlation matrix decomposed from SVD decomposition multiplied by normal random number matrix.
  • G X G co variance matrix needs building. If one has a 5 x 10,000 matrix for day 1, for the next day create a new random matrix of 5 x 10,000 (or reuse a previous random matrix in some way such as by shuffling) and use the simulated previous day grid values (G r values) to get the next day simulated results.
  • the goal is to have, for each day, for each grid, all the 10,000 simulated results, arranged in matrix format—a list of matrices. For example, assuming 30 days, one should see 30 x 5 x 100,000 double precision numbers.
  • spot exchange rate uses the simulated results from equation (3) above for both domestic and foreign interest rates, so can arrive at future arbitrage-free exchange rates. If one cannot simulate the foreign interest rates because of data availability issues, then one has to reasonably estimate it.
  • Black-Scholes model to calculate the expected implied volatility. If not, one uses historical underlying return data and a standard deviation measure on it as the estimate of volatility as above in Equation (1).
  • building blocks 15 include multi-variant regression analysis and other analytical financial programs or processing routines.
  • An example building block 15 is a pure random number generation model that can be used in a Monte Carlo simulation for example.
  • Application 11 may utilize the random number output in conjunction with some matrix algebra that may vary from application to application in solving a specific problem. Initially, by way of non-limiting example, the following focuses on one specific problem, Value at Risk (VaR) for a number of portfolios, based on a covariance matrix.
  • VaR Value at Risk
  • Random number generation in accordance with the present invention has two parts:
  • a C program that generates these random numbers it may be useful to generate these random numbers based on different algorithms (e.g., the
  • a C program uses data from Step 1, Step 2 and Step 4 to compute and generate portfolio analytics. In batch mode, analytics is done for multiple portfolios in parallel.
  • Steps 1 through 4 are essentially a definition of a specific problem that the present invention addresses through an FPGA (building blocks 15) solution. Though Steps 2 through 4 may change from applied problem to problem, Step 1 can be used as the common denominator for many such problems. Any design should consider flexibility in changes to C programs in all of the above steps in general and specifically to Step 1 in particular.
  • KBSFPGAInterface fpgalnterface (KBSFPGAInterface) new KBSFPGAInterfacelmplementation () ;
  • KBSFPGAInterface fpgalnterface (KBSFPGAInterface) new KBSFPGAInterfacelmplementation () ;
  • DataSet dataset fpgalnterface.UploadSecurityData (securities, secReturns);
  • This covariance matrix is used as a lookup by the FPGA.
  • the rows is determined by the numOfRiskFactors and the columns is determined by the numOfScenarios .
  • the covariance matrix is decomposed using Singular Value Decomposition (SVD) .
  • Matrix KBSFPGA_GetReturnMatrix (void* decomposedCovarianceMatrix, void* randomMatrix) ;
  • void* GaussJordanSolver (Matrix coefficients, Vector solutions); void GaussJordanSolver (Matrix coefficients, Vector solutions, Matrix workMatrix) ;
  • KBSFPGA_GenerateRandomMatrix (portSecurities .count, 100000) ;
  • KBSFPGA_GetReturnMatrix (decomposedCovarianceMatrix, zMatrix) ;
  • Fig. 3 is a diagram of the internal structure of a computer (e.g., client processor/device 50 or server computers 60) in the computer system of Fig. 2.
  • Each computer 50, 60 contains system bus 79, where a bus is a set of hardware lines used for data transfer among the components of a computer or processing system.
  • Bus 19 is essentially a shared conduit that connects different elements of a computer system (e.g., processor, disk storage, memory, input/output ports, network ports, etc.) that enables the transfer of information between the elements.
  • Attached to system bus 19 is VO device interface 82 for connecting various input and output devices (e.g., keyboard, mouse, displays, printers, speakers, etc.) to the computer 50, 60.
  • Network interface 86 allows the computer to connect to various other devices attached to a network (e.g., network 70 of Fig. 2).
  • Memory 90 provides volatile storage for computer software instructions 92 and data 94 used to implement an embodiment of the present invention (e.g., FPGA building blocks 15, API's 14 and applications 11 detailed above).
  • Disk storage 95 provides non-volatile storage for computer software instructions 92 and data 94 used to implement an embodiment of the present invention.
  • Central processor unit (CPU) 84 is also attached to system bus 79 and provides for the execution of computer instructions.
  • the processor routines 92 and data 94 are a computer program product (generally referenced 92), including a computer readable medium (e.g., a removable storage medium such as one or more DVD-ROM's, CD-ROM's, diskettes, tapes, etc.) that provides at least a portion of the software instructions for the invention system.
  • Computer program product 92 can be installed by any suitable software installation procedure, as is well known in the art.
  • at least a portion of the software instructions may also be downloaded over a cable, communication and/or wireless connection.
  • the invention programs are a computer program propagated signal product 107 embodied on a propagated signal on a propagation medium (e.g., a radio wave, an infrared wave, a laser wave, a sound wave, or an electrical wave propagated over a global network such as the Internet, or other network(s)).
  • a propagation medium e.g., a radio wave, an infrared wave, a laser wave, a sound wave, or an electrical wave propagated over a global network such as the Internet, or other network(s).
  • Such carrier medium or signals provide at least a portion of the software instructions for the present invention routines/program 92.
  • the propagated signal is an analog carrier wave or digital signal carried on the propagated medium.
  • the propagated signal may be a digitized signal propagated over a global network (e.g., the Internet), a telecommunications network, or other network.
  • the propagated signal is a signal that is transmitted over the propagation medium over a period of time, such as the instructions for a software application sent in packets over a network over a period of milliseconds, seconds, minutes, or longer.
  • the computer readable medium of computer program product 92 is a propagation medium that the computer system 50 may receive and read, such as by receiving the propagation medium and identifying a propagated signal embodied in the propagation medium, as described above for computer program propagated signal product.
  • carrier medium or transient carrier encompasses the foregoing transient signals, propagated signals, propagated medium, storage medium and the like.
  • building blocks 15 may be distributed on one or more computers 50, 60.
  • a computer node 50, 60 may be dedicated to a respective building block 15.
  • one or more PCI cards/boards 12 may be employed to carry the FPGA building blocks 15. Accordingly, a computer 50, 60 may be customized having certain PCI cards/boards 12 with desired ones of the FPGA building blocks 15.
  • a computer system 40 having a mother board 42 carrying a CPU 30, typical SCSI interface 32 (for I/O and network communications) and working or storage memory 46. Also coupled to or otherwise carried on mother board 42 is a plurality of PCI boards 12 holding invention FPGA's/processing units 15.
  • each PCI board 12 holds four FPGA's 15.
  • the FPGA's are programmed as described above to respectively implement fundamental algebraic operations, statistical operations and real time matrix building operations employed by desired applications for financial or other computing, quantitative analyses and the like.
  • computer system 40 employs a WindowsTM
  • Linux (Microsoft) operating system and in other embodiments a Linux operating system.
  • Other operating systems and board configurations are suitable.
  • Rpt ypt-x*q*x'ypt
  • beta - estimate of the cointegrating vector a beta - estimate of the cointegrating vector a.
  • proc (4) ci_vec(mat,r); local e,n,beta,alpha;
  • n rows(mat);
  • alpha - mat[.,2*n+l:3*n]*beta;

Abstract

Hardware accelerated analytical computing apparatus, method and system includes one or more analytical processing units. Each processing unit is formed of a respective field programmable gate array (FPGA) and performs a respective analytical operation. An application interface enables use of respective desired ones of the analytical processing units by one or more computer application programs. Different computer application programs utilize different ones of the analytical processing units. Each computer application program performs a respective analytical computation. Alternatively, an application program may be programmed into one of the FPGA's.

Description

METHODS AND SYSTEMS FOR COMPUTING PLATFORM
RELATED APPLICATIONS
This application is a continuation of and claims priority to U.S. Patent
Application No. , filed November 15, 2006, entitled "Methods and Systems for Computing Platform" (Attorney Docket No. 4040.1000-001). This application also claims the benefit of U.S. Provisional Application No. 60/741,680, filed on December 2, 2005. The entire teachings of the above applications are incorporated herein by reference.
FIELD OF THE INVENTION
The present invention relates generally to methods and systems for providing a unique and alternative computing platform operable to address the needs of high performance computing areas. Financial industry is just one example of such a field where there exists such a need for high performance computing. Currently, high performance problems are solved through computer clusters, scalable grids etc. whose costs increase exponentially as the number of units involved in the processing increase.
BACKGROUND OF THE INVENTION
At present, all industries utilize a variety of extremely costly computer platforms to address certain computationally-intensive solutions such as numerical analysis, monte-carlo simulations and related problems. For example Value at Risk (VaR) problems are addressed in the finance industry through deployment of very powerful (and expensive) services employing multiple CPU's and a number of multithreaded programs that utilize these CPU's. The disadvantage to this process is the fact that as the number of CPU's are increased beyond a critical number (e.g., eight), the costs of such a server and associated services increase exponentially. At the same time, the need for "number crunching" continues to increase due to (a) an increasing number of investment managers; (b) an increasing number of portfolios per investment manager; (c) the kind and the volumes of complex derivatives and (c) an increase in available historical data. It is desirable to provide methods and systems that overcome this disadvantage, as well as a solution that can be applicable to any industry and can be utilized by accessing the computing power through widely known programming languages such as C, C++, Java, and Ruby, with potential gains of 10-200 times performance improvement for l/3rd the cost. The details in the following sections consider Financial Industry as an example to explain the benefits of the present invention.
SUMMARY OF THE INVENTION
The present invention addresses the foregoing problems of the prior art. In particular, the present invention provides a computer method and apparatus for analytical financial computing (computational processing) and more generally for a hardware accelerated numerical system and analysis.
In a preferred embodiment, the invention analytical computing apparatus comprises (a) one or more analytical processing units and (b) an application interface enabling use of respective desired ones of the analytical processing units by one or more computer application programs. Different computer application programs utilize different ones of the analytical processing units, and each computer application program performs a respective analytical computation, hi accordance with one aspect of the present invention, each processing unit is formed of a respective field programmable gate (FPGA) array and performs a respective analytical operation.
In other embodiments, a computer system comprises a mother board hosting a multiplicity of processing cards. Each card carries respective analytical processing units (FPGA's). One subset of the FPGA's provides numerical algebra operations (e.g., matrix mathematics, eigen values, eigen vectors, etc.) and linear equations. Another subset of the FPGA's provides statistical operations (e.g., random number generator, variance-covariance matrix construction, maximum likelihood estimator, etc.) A third subset of the FPGA's provides real time matrix operations for data stream (continuous) and time rate data feeds (e.g., foreign exchange rate estimator, interest rate estimator, etc.)
In accordance with one aspect, the present invention facilitates high performance computing (ranging from ten to two hundred times performance improvement and approximately one-third of the "hard" costs for an equivalent grid computing solution) in three fundamental areas:
1) Numerical algebra comprising entire matrix operations including eigen value decomposition and a linear equation solver. This portion unto itself can be considered as "numerical recipes in hardware".
2) Monte Carlo simulations and general equilibrium model
3) Real time high frequency data (trades and quotes, TAQ, data in financial industry) normalization and associated trigger generation.
Applicant applies the above three fundamental technology blocks to the following functional areas within the financial industry (as a sample set):
1) Identification of correlation and cointegration among different instruments (pairs, groups) and subsequent trigger generation based on real time data
2) Instantaneous capability of performing complex calculations such as marginal and incremental value at risk
3) Interpolation results of TAQ data for further analysis
4) Prediction of interest rate paths and correspondingly, currency exchange rates
5) Algorithmic trading
6) Complex derivative pricing
The above functional areas are of interest to and part of the responsibilities of business units such as
Quantitive research;
High frequency trading;
Trading desks; and
Risk management
in traditional asset management firms, hedge funds, banks (credit risk) and insurance companies and the like.
BRIEF DESCRIPTION OF THE DRAWINGS
The foregoing will be apparent from the following more particular description of example embodiments of the invention, as illustrated in the accompanying drawings in which like reference characters refer to the same parts throughout the different views. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating embodiments of the present invention.
Fig. 1 is a schematic diagram showing an overview of a system according to the present invention.
Fig. 2 is a schematic diagram of a computer network in which embodiments of the present invention operate.
Fig. 3 is a block diagram of a computer node in the network of Fig. 2.
Fig. 4 is a block diagram of a computer system embodying the present invention. DETAILED DESCRIPTION OF THE INVENTION
A description of example embodiments of the invention follows.
The present invention overcomes the disadvantages of prior systems by using FPGA' s (Field-Programmable Gate Arrays) to solve financial industry analytical problems and related computational processing. An FPGA is a semiconductor device containing programmable logic components and programmable
interconnects. The programmable logic components can be programmed to duplicate the functionality of basic logic gates (such as AND, OR, XOR, INVERT) or more complex combinatorial functions such as decoders or simple math functions. In most FPGA's these programmable logic components (or logic blocks, in FPGA parlance) also include memory elements, which may be simple flip-flops or more complete blocks of memories.
A hierarchy of programmable interconnects allows the logic blocks of an FPGA to be interconnected as needed by the system designer, somewhat like a one- chip programmable breadboard. These logic blocks and interconnects can be programmed after the manufacturing process by the customer/designer (hence the term "field-programmable") so that the FPGA can perform whatever logical function is needed.
FPGA's are commercially available, for example, from Altera Corporation of San Jose, California.
In accordance with the present invention, FPGA's are employed in a financial industry analytical computing or processing system 10 as illustrated in Fig. 1. The illustrated system 10 is formed of (a) a PCI card or board 12 hosting one or more analytical processing units (or building blocks) 15; and (b) an application interface 14 built using any common programming language such as Java, C++, Jscript, etc., that interacts between industry applications 11 desired and the PCI card 12.
The analytical processing units or building blocks 15 are FPGA' s programmed to solve different financial industry problems of interest (e.g., analytical financial problems). In one embodiment, there are three categories or sets of analytical processing units/building blocks 15 implemented on the FPGA's using known techniques, processor routines and/or algorithms. One category or subset of FPGA building blocks 15 performs or otherwise implements linear algebra
(algebraic) operations such as functional matrix operations and linear equation solving. Another category or subset of building blocks/FPGA's 15 provides statistical operations such as random number generators, variance-covariance matrix construction and correlation matrix construction, maximum likelihood estimation and non-stationary correlator among a time series. Preferably a third category or subset of building blocks/FPGA's 15 implements real-time matrix building.
With regard to the linear algebra processing units 15 included are FPGA's that respectively implement matrix addition, matrix subtraction, matrix transpose, matrix trace, matrix extraction (as to form or compute a minor, adjoin and cofactor), matrix multiplication (including inner product and direct/outer product), matrix determinant (indicating if singular), matrix inverse (if not singular), matrix decomposition (including solving eigen-values and eigen- vectors), matrix inertial and linear equation solvers.
In the preferred embodiment, a scalar is a 1 by 1 matrix (0-dimension tensor), and a vector is a 1 by N or N by 1 matrix (1 -dimension tensor) and a planar matrix has 2-dimension tensor. Further the matrix inertial processing unit 15 outputs three numbers, namely the number of eigen values >0, the number of eigen values =0 and those <0. The matrix multiplication preferably follows the Winograd or similar suitable method. In matrix decomposition, use of Jacobian matrixes is preferred and output includes the diagonal, an eigen value matrix and a matrix of the corresponding eigen vectors. The decomposition may be performed using SVD (singular value decomposition), Cholesky's algorithm or the like. Using SVD the output is U, B, V where A=U1BV . Assuming a covariance matrix is already built, there will be situations whereby a new row needs to be added to the covariance matrix that corresponds to a new instrument and subsequently SVD decomposition is applied again.
Further the linear equation solvers preferably include implementation of
Gauss-Jordan algorithm, Gauss elimination, Jordan elimination and various approximation algorithms (iterations, Seidal, relaxation and the like).
With regard to the statistical operations implemented by respective FPGA 's 15, preferably the random number generators generate an output in a matrix format with variables as rows and realizations as columns, or vice versa. There is a uniform distribution within (0,1) and a standard normal distribution (mean = 0, standard deviation = 1, skewness = 0, kurtosis = 3). In the variance-covariance and
correlation (paired correlation, not the one through SVD) matrix construction, the exponential- weighted moving average algorithm is preferably used as follows.
To arrive at a variance-covariance matrix Σ, the element of the matrix is written as:
1 X 1 N
σ UX = TT≡ ^ λ r--τ->rj,τ-, Q)
When the decay factor λ is set to 0.9, one needs roughly 20-day historical data; for λ = 0.99 roughly 250-days historical data is needed, depending upon resolution desired.
Use the SVD decomposition to get the correlation matrix C, i.e., Σ = CTC, where C is from SVD (theoretically the same as principal component
decomposition, for Σ is real and symmetric matrix). C = Λ1/2U, whereΛ is the eigenvalue matrix and U is the eigen-vector matrix. Note: Here, returns are log returns, i.e., r = In(I + rtotai)-
In the preferred embodiment, the maximum likelihood estimator (processing unit 15) first estimates four parameters k, θ, σ, γ using the general equilibrium model detailed below. Second, when all four parameters are decided, the maximum likelihood estimator starts iterations day by day to simulate future interest rate paths using a random matrix.
Input includes: (a) Historical interest rates (grid-level: 1 month, 3 month, 2 year, 5 year, 10 year) is a G x N matrix where G is the number of grid levels and N is the number of historical rates. This is for purposes of illustration, not limitation; G could be larger.
(b) Range and increments for the four parameters k, θ, σ, γ is (4*3 = 12)
(c) Number of forecast horizon days is D
(d) Number of simulations is S
(e) Default value for Δt is 1/365, i.e., year as the unit
Output is then:
(i) A G x Dl matrix, each row representing the grid level, and Dl representing the mean of expected value from simulations. ,
(ii) A G x D2 matrix, each row representing the grid level, and D2 representing the standard deviation of expected value from simulations.
(iii) Optionally a list (of K elements) G x S matrices.
General Equilibrium Model for interest rates forecast modeling
For a single term interest rate,
rl+1 =η + k(θ -η)At + σifz, yfλl, (2) where one needs to evaluate the four parameters k, θ, σ, γ, using historical data and maximum likelihood estimation. Note the correlated random number matrix z is a linear combination of standard normal variables, when considering the correlations among different term interest rates. That is, r should be a vector when representing a term structure, and each term has its own four parameters of k, θ, σ, γ.
Z is calculated as Uτ * SQRT(B) * Normal distribution, i.e., correlation matrix decomposed from SVD decomposition multiplied by normal random number matrix.
Maximum likelihood estimation algorithm:
From equation (2), one knows:
The expectation:
E[rM] = η + k(θ-rt)At (3)
The standard deviation: StdDev [rl+1 ] = σrj VΔ7 (4)
Then the probability
Figure imgf000009_0001
Trying N historical days, one gets a joint probability
JP - p [rt_N ] p [rt_N+l ]...p [rM ] , then select the combination of k, θ, σ, γ, which maximize JP.
It is noted that the above processing needs different seeds for each of the G grids, and a G X G co variance matrix needs building. If one has a 5 x 10,000 matrix for day 1, for the next day create a new random matrix of 5 x 10,000 (or reuse a previous random matrix in some way such as by shuffling) and use the simulated previous day grid values (G r values) to get the next day simulated results.
The goal is to have, for each day, for each grid, all the 10,000 simulated results, arranged in matrix format—a list of matrices. For example, assuming 30 days, one should see 30 x 5 x 100,000 double precision numbers.
Further, an example embodiment of the non-stationary correlation among time series processing unit 15 follows the algorithm (code) in Appendix I.
With regard to real time matrix building operations, respective building blocks FPGA' s 15 implement Black-Scholes algorithm (routine or calculation) for option pricing, an algorithm for pricing mortgage backed securities and an algorithm for predicting interest rates. In the preferred embodiment, these FPGA' s 15 receive as input real time TAQ (Trades and Quotes) data streams, time rate data feeds (e.g., from Reuters Company) and similar continuous data. Other input includes historical interest rates and/or historical foreign exchange rates. Output then includes a matrix of future foreign exchange rates and/or simulated future interest rates. Preferably future exchange rates are estimated through interest rates as follows.
Exchange Rates Modeling '
1 + r
forwardφC \ FC) = spotφC \ FC){^^-) (6) l + rF DC/FC means the ratio of domestic currency to foreign currency, and ro and rp mean the domestic and foreign term interest rates, respectively.
One knows spot exchange rate and uses the simulated results from equation (3) above for both domestic and foreign interest rates, so can arrive at future arbitrage-free exchange rates. If one cannot simulate the foreign interest rates because of data availability issues, then one has to reasonably estimate it.
Implied Volatility Modeling
If one has sufficient option pricing data, one can use Black-Scholes model to calculate the expected implied volatility. If not, one uses historical underlying return data and a standard deviation measure on it as the estimate of volatility as above in Equation (1).
Returning to Fig. 1, using the processing units/building blocks 15, different application programs 11 of interest (e.g., financial applications) are built. In particular, known algorithms are used in conjunction with the present invention FPGA building blocks 15 to implement the following applications 11:
• Portfolio VaR (Value at Risk) calculations
• Portfolio Optimization
• Alpha Back testing
• Algorithm trading
• Monte Carlo simulation
Other example financial applications utilizing building blocks 15 include multi-variant regression analysis and other analytical financial programs or processing routines.
Restated, each financial application 11 needs different ones of the FPGA building blocks 15. API layer 14 enables a programmer to utilize the appropriate FPGA building blocks 15 to build a desired application 11. Alternatively, the application 11 of interest may be programmed or "burned into" one or more of the FPGA's 15 on the PCI card/board 12.
An example building block 15 is a pure random number generation model that can be used in a Monte Carlo simulation for example. Application 11 may utilize the random number output in conjunction with some matrix algebra that may vary from application to application in solving a specific problem. Initially, by way of non-limiting example, the following focuses on one specific problem, Value at Risk (VaR) for a number of portfolios, based on a covariance matrix.
Step One - Random Number Generation
Random number generation in accordance with the present invention has two parts:
1. A C program that generates these random numbers—it may be useful to generate these random numbers based on different algorithms (e.g., the
KCM algorithm) that are reflected in different C programs. (See, e.g., Chapman, "Fast integer multipliers fit in FPGAs," EDN Magazine, May 1994; and Boullis, et al., "Some Optimizations of Hardware Multiplication by Constant Matrices," both of which are incorporated herein by reference as if set forth in their entireties) .
and
2. A number of random sequences—close to 14,000 sequences, with each sequence being around 100,000 observations. There is no data involved in this process. All that is required is a simple C program that can be modified from application to application. A single program can be written with different functions that can be called, based on the argument (variable values, coefficients, etc.) that is passed. Step Two - Covariance Matrix Generation - Once a Day
Data Size is:
6 GB ~ worst case (2.5 MB (each day for 80,000 securities) x 10 (number of years) x 252 (number of days));
1 GB— best case (universe of securities is limited to about 14,000, which is more of the norm and more practical).
Program: Assuming that the universe of securities is limited to 14,000, the end result is a 14,000 x 14,000 matrix with each cell representing a double. Each cell is again created through a C program, and the process can be run in parallel since computation of a single cell in the above matrix is independent and all computations of cells use the above data set. Step Three— Covariance Matrix Decomposition— Once a Day
A minimum of 10 runs of a C program that decomposes the above matrix- resulting in a 14,000 x 14,000 matrix each time. Each run also stores the decomposed matrices in a suitable memory, data store or the like. Step Four— Portfolio Analysis
For portfolio analysis there are two possible scenarios:
1. batch mode;
2. interactive.
In both cases, inputs is a list of securities and their corresponding active weights. A C program uses data from Step 1, Step 2 and Step 4 to compute and generate portfolio analytics. In batch mode, analytics is done for multiple portfolios in parallel.
The above Steps 1 through 4 are essentially a definition of a specific problem that the present invention addresses through an FPGA (building blocks 15) solution. Though Steps 2 through 4 may change from applied problem to problem, Step 1 can be used as the common denominator for many such problems. Any design should consider flexibility in changes to C programs in all of the above steps in general and specifically to Step 1 in particular.
Pseudo code further illustrating the API layer 14 interaction with FPGA building blocks 15 and example financial applications 11 (specifically a
multivariate attribution program and a VaR program calculation in Java) follow.
KBSFPGAInterface. java
package com. kbs . fpga;
public interface KBSFPGAInterface
{
public DataSet UploadSecurityData (Vector securities, Matrix secReturns) ;
public Matrix GenerateCovarianceMatrix (Vector securities, DataSet dataset) ; l public Matrix GenerateRandottiMatrix(int numOfRiskFactors, int numOfScenarios) ;
public Matrix DecomposeCovarianceMatrix (Matrix
covarianceMatrix) ;
public Matrix GetReturnMatrix (Matrix
decomposedCovarianceMatrix, Matrix randomMatrix) ,- public Matrix MultiplyMatrix (Matrix Ihs, Matrix rhs) ;
public Vector MultiVariateRegression (Matrix observations, Vector targets) ;
public Matrix GaussJordanSolver (Matrix coefficients, Matrix solutions) ;
} mvattribution. Java
import com.kbs . fpga. *;
public class mvattribution
{
public static void main (String args [] )
{
/*
* Assume we have the observations for the various factors (independent variables) and the target (dependent variable) available
* Also, assume there is an implementation for the KBSFPGAInterface called KBSFPGAInterfacelmplementation
*/
KBSFPGAInterface fpgalnterface = (KBSFPGAInterface) new KBSFPGAInterfacelmplementation () ;
Matrix observations =
Vision. getAlphaFactorObservations () ;
Vector targets = Vision.getRegressionTargets ();
/* Multivariate Regression involves Matrix inversion routines */
Vector coefficients =
fpgalnterface.MultiVariateRegression (observations, targets) ;
}
} var. java import com.kbs . fpga. * ;
public class var i
{
public static void main (String args [] )
{
/*
* Assume security data and return data for securities are available from some repository, say - Vision
* Also, assume there is an implementation for the KBSFPGAInterface called KBSFPGAInterfacelmplementation
*/
KBSFPGAInterface fpgalnterface = (KBSFPGAInterface) new KBSFPGAInterfacelmplementation () ;
Vector securities = Vision. getSecurities (); Matrix secReturns = Vision.getSecReturnDataFor
(securities) ;
DataSet dataset = fpgalnterface.UploadSecurityData (securities, secReturns);
Vector portSecurities =
Vision.getSecuritiesInPortfolio("MY_PORTFOLIO") ;
Matrix covarianceMatrix =
fpgalnterface.GenerateCovarianceMatrix (portSecurities, dataset) ;
Matrix zMatrix =
fpgalnterface.GenerateRandomMatrix (portSecurities .length, 100000) ;
Matrix decomposedCovarianceMatrix =
fpgalnterface .DecomposeCovarianceMatrix (covarianceMatrix) ;
Matrix returnMatrix =
fpgalnterface .GetReturnMatrix (decomposedCovarianceMatrix, zMatrix) ;
}
}
Further examples of pseudo code in other programming languages for building blocks 15, API layer 14 and applications 11 are:
kbsfpga.h
#ifndef _kbsfpga_h_included_
#define _kbsfpga_h_included_
/*
* Used to upload historical return data for a set of securities.
* Returns a handle to this security dataset.
*/
void* KBSFPGAJUploadSecurityData (Vector securities, Matrix
secReturns) ;
/*
* Used to generate the covariance matrix for securities in the dataset .
* This covariance matrix is used as a lookup by the FPGA.
* Returns the 'sub-matrix' of covariances for securities.
*/
void* KBSFPGA_GenerateCovarianceMatix (Vector securities, void* dataset) ;
void KBSFPGA_GenerateCovarianceMatix (Vector securities, void* dataset, Matrix* covarianceMatrix) ;
/*
* Used to generate a z-matrix i.e. a matrix of independent standard normal distribution random numbers.
* The rows is determined by the numOfRiskFactors and the columns is determined by the numOfScenarios .
*/
void* KBSFPGA_GenerateRandomMatrix(int numOfRiskFactors, int numOfScenarios) ;
void KBSFPGA_GenerateRandomMatrix(int numOfRiskFactors, int numOfScenarios, Matrix* randomMatrix) ;
/*
* Used to decompose the covariance matrix. * The covariance matrix is decomposed using Singular Value Decomposition (SVD) .
* Returns the decomposed form - sqrt(s) transpose (u) of the covariance matrix
*/
void* KBSFPGA_DecomposeCovarianceMatrix(void* covarianceMatrix) ; void KBSFPGA_DecomposeCovarianceMatrix(void* covarianceMatrix, Matrix* decomposedCovarianceMatrix) , /*
* Used to generate the 1-day return vector for the risk factors.
* Returns the return vector for the different risk factors .
*/
Matrix KBSFPGA_GetReturnMatrix (void* decomposedCovarianceMatrix, void* randomMatrix) ;
/*
* Used to multiply two matrices.
* Returns the result of multiplying the two matrices.
*/
void* KBSFPGA_MultiplyMatrix(void* lhs, void* rhs) ;
void KBSFPGA_MultiplyMatrix(void* lhs, void* rhs, Matrix* result);
/*
* Used to perform multivariate regression.
* Returns the coefficients of the regression.
*/
void* KBSFPGA_MultiVariateRegression (Matrix observations, Vector targets) ;
void KBSFPGA_MultiVariateRegression (Matrix observations, Vector targets, Vector coefficients) ,
/*
* Used to solve a system of linear equations by Gauss-Jordan elimination.
* We can also provide other methods that can solve a system of linear equations.
* Here Gauss-Jordan elimination method is provided as an example.
* Returns the work matrix. */
void* GaussJordanSolver (Matrix coefficients, Vector solutions); void GaussJordanSolver (Matrix coefficients, Vector solutions, Matrix workMatrix) ;
#endif mvattribution. c
#include <kbsfpga.h> int main()
{
/*
* Assume we have the observations for the various factors (independent variables) and the target (dependent variable) available
*/
Matrix observations = getAlphaFactorObservations () ;
Vector targets = getRegressionTargets () ; void* coefficients =
KBSFPGA_MultiVariateRegression (observations, targets) ;
} var.c
#include <kbsfpga.h> int main()
{
/*
* Assume security data and return data for securities are available from some repository, say - Vision
*/
Vector securities = getSecurities () ;
Matrix secReturns = getSecReturnDataFor (securities) ; void* dataset = KBSFPGA_UploadSecurityData (securities, secReturns) ; Vector portSecurities =
getSecuritiesInPortfolio ( "MY_PORTFOLIO" ) ;
void* covarianceMatrix =
KBSFPGA_GenerateCovarianceMatrix (portSecurities, dataset) ; void* zMatrix =
KBSFPGA_GenerateRandomMatrix (portSecurities .count, 100000) ;
void* decomposedCovarianceMatrix =
KBSFPGA_DecomposeCovarianceMatrix (covarianceMatrix) ; .
Matrix returnMatrix =
KBSFPGA_GetReturnMatrix (decomposedCovarianceMatrix, zMatrix) ;
}
Referring back to Fig. 1, system 10 communicates to host computer through an I/O interface 13. The host computer may be a computer node 60, 50 in a network of computers as illustrated in Fig. 2.
Fig. 2 illustrates a generalized computer network or similar digital processing environment in which the present invention may be implemented. Client computer(s)/devices 50 and server computer(s) 60 provide processing, storage, and input/output devices executing application programs and the like. Client computer(s)/devices 50 can also be linked through communications network 70 to other computing devices, including other client devices/processes 50 and server computer(s) 60. Communications network 70 can be part of a remote access network, a global network (e.g., the Internet), a worldwide collection of computers, Local area or Wide area networks, and gateways that currently use respective protocols (TCP/IP, Bluetooth, etc.) to communicate with one another. Other electronic device/computer network architectures are suitable.
Fig. 3 is a diagram of the internal structure of a computer (e.g., client processor/device 50 or server computers 60) in the computer system of Fig. 2. Each computer 50, 60 contains system bus 79, where a bus is a set of hardware lines used for data transfer among the components of a computer or processing system. Bus 19 is essentially a shared conduit that connects different elements of a computer system (e.g., processor, disk storage, memory, input/output ports, network ports, etc.) that enables the transfer of information between the elements. Attached to system bus 19 is VO device interface 82 for connecting various input and output devices (e.g., keyboard, mouse, displays, printers, speakers, etc.) to the computer 50, 60. Network interface 86 allows the computer to connect to various other devices attached to a network (e.g., network 70 of Fig. 2). Memory 90 provides volatile storage for computer software instructions 92 and data 94 used to implement an embodiment of the present invention (e.g., FPGA building blocks 15, API's 14 and applications 11 detailed above). Disk storage 95 provides non-volatile storage for computer software instructions 92 and data 94 used to implement an embodiment of the present invention. Central processor unit (CPU) 84 is also attached to system bus 79 and provides for the execution of computer instructions.
In one embodiment, the processor routines 92 and data 94 are a computer program product (generally referenced 92), including a computer readable medium (e.g., a removable storage medium such as one or more DVD-ROM's, CD-ROM's, diskettes, tapes, etc.) that provides at least a portion of the software instructions for the invention system. Computer program product 92 can be installed by any suitable software installation procedure, as is well known in the art. In another embodiment, at least a portion of the software instructions may also be downloaded over a cable, communication and/or wireless connection. In other embodiments, the invention programs are a computer program propagated signal product 107 embodied on a propagated signal on a propagation medium (e.g., a radio wave, an infrared wave, a laser wave, a sound wave, or an electrical wave propagated over a global network such as the Internet, or other network(s)). Such carrier medium or signals provide at least a portion of the software instructions for the present invention routines/program 92.
In alternate embodiments, the propagated signal is an analog carrier wave or digital signal carried on the propagated medium. For example, the propagated signal may be a digitized signal propagated over a global network (e.g., the Internet), a telecommunications network, or other network. In one embodiment, the propagated signal is a signal that is transmitted over the propagation medium over a period of time, such as the instructions for a software application sent in packets over a network over a period of milliseconds, seconds, minutes, or longer. In another embodiment, the computer readable medium of computer program product 92 is a propagation medium that the computer system 50 may receive and read, such as by receiving the propagation medium and identifying a propagated signal embodied in the propagation medium, as described above for computer program propagated signal product.
Generally speaking, the term "carrier medium" or transient carrier encompasses the foregoing transient signals, propagated signals, propagated medium, storage medium and the like.
In other embodiments, building blocks 15 may be distributed on one or more computers 50, 60. Alternatively, a computer node 50, 60 may be dedicated to a respective building block 15. Further, one or more PCI cards/boards 12 may be employed to carry the FPGA building blocks 15. Accordingly, a computer 50, 60 may be customized having certain PCI cards/boards 12 with desired ones of the FPGA building blocks 15.
Illustrated in Fig. 4 is yet another embodiment of the present invention. Presented is a computer system 40 having a mother board 42 carrying a CPU 30, typical SCSI interface 32 (for I/O and network communications) and working or storage memory 46. Also coupled to or otherwise carried on mother board 42 is a plurality of PCI boards 12 holding invention FPGA's/processing units 15.
Preferably there are sixteen PCI boards 12 and each PCI board 12 holds four FPGA's 15. The FPGA's are programmed as described above to respectively implement fundamental algebraic operations, statistical operations and real time matrix building operations employed by desired applications for financial or other computing, quantitative analyses and the like.
In some embodiments computer system 40 employs a Windows™
(Microsoft) operating system and in other embodiments a Linux operating system. Other operating systems and board configurations are suitable.
While this invention has been particularly shown and described with references to example embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the invention encompassed by the appended claims. Other configurations, combinations and computer architectures (stand alone or networked) are suitable and in the purview of one skilled in the art given the foregoing description.
APPENDIX I
/*PROC JOHANSEN
Pedro J.F. de Lima, May 1993
Department of Economics, The Johns Hopkins University
This code is written and submitted for public, non-commercial use.
There are no performance guarantees.
Please acknowledge this code (and its author)
if you find it useful in your own work FOPvMAT
{tr, l_max, z} = johansen (levels, nlags);
INPUT
levels - levels of N series to test for cointegration
nlags - number of lags in the VAR specification
OUTPUT
tr - Trace Statistic, for the hypothesis that there are at most q=0, 1, 2, ..., n-1 cointegrating vectors against the completely unrestricted VAR(p) model.
That is Ho: r<=q Hl: 0<=q<n.
1-max - 1 max statistic for the null hypothesis that there are
at most r<=q cointegrating vectors (q=0, 1, ..., n-1) against the alternative that only one additional cointegrating vector exists.
z - N x N* 3 matrix with the following three NxN matrices:
- E = NxN matrix of normalized eigenvectors (E'Spp E=I), the first column of E giving the eigenvector associated with the largest eigenvalue and so on.
- SOO = NxN matrix with the product moment matrix SOO
- SOp = NxN matrix with the product moment matrix SOp These matrices are required to compute the cointegrating vectors a, the parameters a, the long run matrix a=aa',
and the variance matrix of the residuals, once r, the # of cointegrating vectors, is defined.
See the procedure COINTEGR
*/
proc (3) - johansen(levels,p);
local N,T,diff,x,i,yOt,ypt,q,rOt,rpt,sOO,sOp,spO,spp,c,cc,
lambda, vlambda,lr;
N=cols(levels); T=rows(levels);
diff=(levels-lag 1 (levels)) ;
x=diff;
i=l;do while i<=p-l;
x=x~lagn(diff,i);
i=i+l;
endo;
x=packr(x);
yOt=x[.,l:N];
ypt=packr (lagn(levels,p)) ;
x=x[.,N+l:p*N];
q=invpd(x'x); R0t=y0t-x*q*x'y0t;
Rpt=ypt-x*q*x'ypt;
sOO=rOt'rOt/T;
sOp=rOt'rpt/T;
spθ=sθp';
spp=rpt'rpt/T;
c=chol(spp)';
c=inv(c);
{lambda, vlambda}=eigrs2(cc*spθ*invpd(sθθ)*sθp*cc');
retp(rev(-T* cumsumc(ln( 1 -lambda))),rev(-T* 1 n( 1 -lambda)),
(rev((cc'vlambda)')')~sθθ~sθp);
endp;
/*PROC CI_VEC
Pedro J. F. de Lima, May 1993
Department of Economics, The Johns Hopkins University
FORMAT
{beta, alpha, lrun, omega} = ci_vec(mat,r);
Input
mat - n x 3n matrix = to various as defined in proc johansen r - number of cointegrating vectors
OUTPUT
beta - estimate of the cointegrating vector a.
alpha - estimate of a.
lrun - estimate of the long run matrix a=aa'
omega - estimate of β, the variance matrix of the VAR(p) innovations */
proc (4) = ci_vec(mat,r); local e,n,beta,alpha;
beta=mat[.,l:r];
n=rows(mat);
alpha= - mat[.,2*n+l:3*n]*beta;
retp(beta,alpha,-mat[.,2*n+l :3 *n] *beta*beta',mat[.,n+l :2*n]-alpha*alpha'); endp;

Claims

CLAIMS What is claimed is:
1. Analytical computing apparatus comprising:
one or more analytical processing units, each unit being formed of a respective field programmable gate array and performing a respective analytical operation; and
an application interface enabling use of respective desired ones of the analytical processing units by one or more computer application programs, different computer application programs utilizing different ones of the analytical processing units, each computer application program performing a respective analytical computation, such that the analytical computing apparatus provides a hardware accelerated numerical system and analysis.
2. Apparatus as claimed in Claim 1 wherein the respective analytical operations of the analytical processing units include any of algebraic operations, statistical operations and real time matrix operations.
3. Apparatus as claimed in Claim 2 wherein the algebraic operations include matrix mathematics operations and linear equation solving.
4. Apparatus as claimed in Claim 2 wherein the statistical operations include random number generation, variance-covariance matrix construction, correlation matrix construction, maximum likelihood estimation, and non- stationary correlation among a time series.
5. Apparatus as claimed in Claim 2 wherein the real time matrix operations include option pricing, pricing mortgage backed securities, interest rate prediction and foreign exchange rate prediction.
6. Apparatus as claimed in Claim 1 wherein the computer application programs include analytical financial applications implementing any of portfolio value at risk calculations,, portfolio optimization, alpha back testing, algorithm trading and multivariant regulation analysis.
7. A method for analytical computing comprising:
providing one or more analytical processing units, each unit being formed of a respective field programmable gate array and performing a respective analytical operation; and
enabling use of respective desired ones of the analytical processing units by one or more computer application programs, different computer application programs utilizing different ones of the analytical processing units, each computer application program performing a respective analytical computation, such that a hardware accelerated numerical system and analysis results.
8. The method as claimed in Claim 7 wherein the respective analytical
operations of the analytical processing units include any of algebraic operations, statistical operations and real time matrix operations.
9. The method as claimed in Claim 8 wherein the algebraic operations include matrix mathematics operations and linear equation solving.
10. The method as claimed in Claim 8 wherein statistical operations include random number generation, variance-covariance matrix construction, correlation matrix construction, maximum likelihood estimation, and non- stationary correlation among a time series.
11. The method as claimed in Claim 8 wherein the real time matrix operations include option pricing, pricing mortgage backed securities, interest rate prediction and foreign exchange rate prediction.
12. The method as claimed in Claim 7 wherein the computer application
programs include analytical financial applications implementing any of portfolio value at risk calculations, portfolio optimization, alpha back testing, algorithm trading and multivariant regulation analysis.
13. Computer apparatus for analytical computing, comprising
a plurality of analytical processing means, each processing means being formed of a respective field programmable gate array (FPGA) and performing a respective analytical operation; and
application interface means for enabling use of respective desired ones of the analytical processing means by one or more computer application programs, different computer application programs utilizing different ones of the analytical processing means, each computer application program performing a respective analytical computation,
wherein the plurality of analytical processing means together with the application interface means provides a hardware accelerated numerical system and analysis.
14. Computer apparatus as claimed in Claim 13 wherein the respective analytical operations of the analytical processing means include any of algebraic operations, statistical operations and real time matrix operations.
15. Computer apparatus as claimed in Claim 14 wherein the algebraic operations include matrix mathematics operations and linear equation solving.
16. Computer apparatus as claimed in Claim 14 wherein the statistical operations include random number generation, variance-covariance matrix construction, correlation matrix construction, maximum likelihood estimation, and non- stationary correlation among a time series.
17. Computer apparatus as claimed in Claim 14 wherein the real time matrix operations include option pricing, pricing mortgage backed securities, interest rate prediction and foreign exchange rate prediction.
18. Computer apparatus as claimed in Claim 13 wherein the computer
application programs include analytical financial applications implementing any of portfolio value at risk calculations, portfolio optimization, alpha back testing, algorithm trading and multi- variant regulation analysis.
19. Computer apparatus as claimed in Claim 13 further comprising an
application program coupled to one of the analytical processing means in a manner such that the application program is carried on the FPGA of the one analytical processing means.
20. Computer apparatus as claimed in Claim 19 wherein the application program coupled to the one analytical processing means is carried on a respective PCI board.
21. A computer system for analytical computing comprising:
a CPU coupled to a mother board;
a plurality of processing boards each coupled to the mother board and having one or more respective analytical processing units, each processing unit being formed of a respective field programmable gate array and performing a respective analytical operation;
working memory coupled to the motherboard; and
system communication interface for elements on the motherboard and for enabling use of respective desired ones of the analytical processing units by one or more computer application programs, different computer application programs utilizing different ones of the analytical processing units, each computer application program performing a respective analytical computation, in a manner such that a hardware accelerated numerical system and analysis results.
22. A computer system as claimed in Claim 21 wherein the respective analytical operations of the analytical processing units include any of algebraic operations, statistical operations and real time matrix operations.
23. A computer system as claimed in Claim 22 wherein: the algebraic operations include matrix mathematics operations and linear equation solving;
the statistical operations include random number generation, variance-covariance matrix construction, correlation matrix construction, maximum likelihood estimation, and non-stationary correlation among a time series; and
the real time matrix operations include option pricing, pricing mortgage backed securities, interest rate prediction and foreign exchange rate prediction.
PCT/US2006/044902 2005-12-02 2006-11-20 Methods and systems for computing platform WO2007067338A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CA002631517A CA2631517A1 (en) 2005-12-02 2006-11-20 Methods and systems for computing platform
EP06838067A EP1955191A1 (en) 2005-12-02 2006-11-20 Methods and systems for computing platform

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
US74168005P 2005-12-02 2005-12-02
US60/741,680 2005-12-02
US11/599,911 2006-11-15
US11/599,911 US7716100B2 (en) 2005-12-02 2006-11-15 Methods and systems for computing platform

Publications (1)

Publication Number Publication Date
WO2007067338A1 true WO2007067338A1 (en) 2007-06-14

Family

ID=37983589

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2006/044902 WO2007067338A1 (en) 2005-12-02 2006-11-20 Methods and systems for computing platform

Country Status (5)

Country Link
US (1) US7716100B2 (en)
EP (1) EP1955191A1 (en)
CA (1) CA2631517A1 (en)
TW (1) TW200741479A (en)
WO (1) WO2007067338A1 (en)

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006072891A (en) * 2004-09-06 2006-03-16 Sony Corp Method and device for generating pseudo random number sequence with controllable cycle based on cellular automata
US9572494B2 (en) * 2008-08-12 2017-02-21 New Jersy Institute of Technology Method and apparatus for multi-spectral imaging and analysis of skin lesions and biological tissues
US8755515B1 (en) 2008-09-29 2014-06-17 Wai Wu Parallel signal processing system and method
US20100169403A1 (en) * 2008-12-31 2010-07-01 L3 Communications Integrated Systems, L.P. System for matrix partitioning in large-scale sparse matrix linear solvers
US8380570B2 (en) * 2009-10-27 2013-02-19 Yahoo! Inc. Index-based technique friendly CTR prediction and advertisement selection
GB201000089D0 (en) * 2010-01-05 2010-02-17 Mura Michael E Numerical modelling apparatus
JP5606114B2 (en) * 2010-03-19 2014-10-15 株式会社東芝 Power generation amount prediction device, prediction method, and prediction program
US20120209659A1 (en) * 2011-02-11 2012-08-16 International Business Machines Corporation Coupling demand forecasting and production planning with cholesky decomposition and jacobian linearization
US20120259792A1 (en) * 2011-04-06 2012-10-11 International Business Machines Corporation Automatic detection of different types of changes in a business process
US9311433B2 (en) * 2011-05-27 2016-04-12 Airbus Operations S.L. Systems and methods for improving the execution of computational algorithms
FR2987483B1 (en) * 2012-02-29 2014-03-07 Sagem Defense Securite METHOD OF ANALYSIS OF FLIGHT DATA
WO2015057837A1 (en) * 2013-10-15 2015-04-23 Analog Devices, Inc. Sampling variables from probabilistic models
WO2016136237A1 (en) 2015-02-23 2016-09-01 Okinawa Institute of Science and Technology Graduate University System and method of determining forecast error for renewable energy fluctuations
TWI588771B (en) * 2016-06-08 2017-06-21 南臺科技大學 Thesis analysis system
CN110851505B (en) * 2019-11-20 2023-12-22 鹏城实验室 Data processing framework, method and system

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6539438B1 (en) 1999-01-15 2003-03-25 Quickflex Inc. Reconfigurable computing system and method and apparatus employing same

Family Cites Families (84)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE3482532D1 (en) * 1983-07-06 1990-07-19 Secr Defence Brit PROCESSOR WITH FORCED ADJUSTMENT.
US4763294A (en) * 1985-12-19 1988-08-09 Wang Laboratories, Inc. Method and apparatus for floating point operations
US4872133A (en) * 1988-02-18 1989-10-03 Motorola, Inc. Floating-point systolic array including serial processors
DE68920388T2 (en) * 1988-09-19 1995-05-11 Fujitsu Ltd Parallel computer system using a SIMD process.
US5072371A (en) * 1989-03-01 1991-12-10 The United States Of America As Represented By The United States Department Of Energy Method for simultaneous overlapped communications between neighboring processors in a multiple
US5020059A (en) * 1989-03-31 1991-05-28 At&T Bell Laboratories Reconfigurable signal processor
US4962381A (en) * 1989-04-11 1990-10-09 General Electric Company Systolic array processing apparatus
US5138695A (en) * 1989-10-10 1992-08-11 Hnc, Inc. Systolic array image processing system
US5175856A (en) * 1990-06-11 1992-12-29 Supercomputer Systems Limited Partnership Computer with integrated hierarchical representation (ihr) of program wherein ihr file is available for debugging and optimizing during target execution
US5477221A (en) * 1990-07-10 1995-12-19 The United States Of America As Represented By The Administrator Of The National Aeronautics And Space Administration Pipeline synthetic aperture radar data compression utilizing systolic binary tree-searched architecture for vector quantization
US5274832A (en) * 1990-10-04 1993-12-28 National Semiconductor Corporation Systolic array for multidimensional matrix computations
JP2647330B2 (en) * 1992-05-12 1997-08-27 インターナショナル・ビジネス・マシーンズ・コーポレイション Massively parallel computing system
US5802290A (en) * 1992-07-29 1998-09-01 Virtual Computer Corporation Computer network of distributed virtual computers which are EAC reconfigurable in response to instruction to be executed
US5509134A (en) * 1993-06-30 1996-04-16 Intel Corporation Method and apparatus for execution of operations in a flash memory array
US6456982B1 (en) * 1993-07-01 2002-09-24 Dragana N. Pilipovic Computer system for generating projected data and an application supporting a financial transaction
US5675803A (en) * 1994-01-28 1997-10-07 Sun Microsystems, Inc. Method and apparatus for a fast debugger fix and continue operation
US5671377A (en) * 1994-07-19 1997-09-23 David Sarnoff Research Center, Inc. System for supplying streams of data to multiple users by distributing a data stream to multiple processors and enabling each user to manipulate supplied data stream
US6052773A (en) * 1995-02-10 2000-04-18 Massachusetts Institute Of Technology DPGA-coupled microprocessors
US5570040A (en) * 1995-03-22 1996-10-29 Altera Corporation Programmable logic array integrated circuit incorporating a first-in first-out memory
US5655122A (en) * 1995-04-05 1997-08-05 Sequent Computer Systems, Inc. Optimizing compiler with static prediction of branch probability, branch frequency and function frequency
US6077315A (en) * 1995-04-17 2000-06-20 Ricoh Company Ltd. Compiling system and method for partially reconfigurable computing
US5903771A (en) * 1996-01-16 1999-05-11 Alacron, Inc. Scalable multi-processor architecture for SIMD and MIMD operations
GB2309558A (en) * 1996-01-26 1997-07-30 Ibm Load balancing across the processors of a server computer
US5737766A (en) * 1996-02-14 1998-04-07 Hewlett Packard Company Programmable gate array configuration memory which allows sharing with user memory
US5926636A (en) * 1996-02-21 1999-07-20 Adaptec, Inc. Remote procedural call component management method for a heterogeneous computer network
US5956518A (en) * 1996-04-11 1999-09-21 Massachusetts Institute Of Technology Intermediate-grain reconfigurable processing device
US5715453A (en) * 1996-05-31 1998-02-03 International Business Machines Corporation Web server mechanism for processing function calls for dynamic data queries in a web page
US5987480A (en) * 1996-07-25 1999-11-16 Donohue; Michael Method and system for delivering documents customized for a particular user over the internet using imbedded dynamic content
US5892962A (en) * 1996-11-12 1999-04-06 Lucent Technologies Inc. FPGA-based processor
US5784108A (en) * 1996-12-03 1998-07-21 Zapex Technologies (Israel) Ltd. Apparatus for and method of reducing the memory bandwidth requirements of a systolic array
US6128663A (en) * 1997-02-11 2000-10-03 Invention Depot, Inc. Method and apparatus for customization of information content provided to a requestor over a network using demographic information yet the user remains anonymous to the server
US5953502A (en) * 1997-02-13 1999-09-14 Helbig, Sr.; Walter A Method and apparatus for enhancing computer system security
US6631518B1 (en) * 1997-03-19 2003-10-07 International Business Machines Corporation Generating and utilizing organized profile information
US6286135B1 (en) * 1997-03-26 2001-09-04 Hewlett-Packard Company Cost-sensitive SSA-based strength reduction algorithm for a machine with predication support and segmented addresses
US6215898B1 (en) * 1997-04-15 2001-04-10 Interval Research Corporation Data processing system and method
US6507898B1 (en) * 1997-04-30 2003-01-14 Canon Kabushiki Kaisha Reconfigurable data cache controller
US5966534A (en) * 1997-06-27 1999-10-12 Cooke; Laurence H. Method for compiling high level programming languages into an integrated processor with reconfigurable logic
US6226776B1 (en) * 1997-09-16 2001-05-01 Synetry Corporation System for converting hardware designs in high-level programming language to hardware implementations
US6061706A (en) * 1997-10-10 2000-05-09 United Microelectronics Corp. Systolic linear-array modular multiplier with pipeline processing elements
US6009410A (en) * 1997-10-16 1999-12-28 At&T Corporation Method and system for presenting customized advertising to a user on the world wide web
US5915123A (en) * 1997-10-31 1999-06-22 Silicon Spice Method and apparatus for controlling configuration memory contexts of processing elements in a network of multiple context processing elements
US6076152A (en) * 1997-12-17 2000-06-13 Src Computers, Inc. Multiprocessor computer architecture incorporating a plurality of memory algorithm processors in the memory subsystem
US6434687B1 (en) * 1997-12-17 2002-08-13 Src Computers, Inc. System and method for accelerating web site access and processing utilizing a computer system incorporating reconfigurable processors operating under a single operating system image
US6079032A (en) * 1998-05-19 2000-06-20 Lucent Technologies, Inc. Performance analysis of computer systems
US6192439B1 (en) * 1998-08-11 2001-02-20 Hewlett-Packard Company PCI-compliant interrupt steering architecture
US6243791B1 (en) * 1998-08-13 2001-06-05 Hewlett-Packard Company Method and architecture for data coherency in set-associative caches including heterogeneous cache sets having different characteristics
WO2000049496A1 (en) * 1999-02-15 2000-08-24 Koninklijke Philips Electronics N.V. Data processor with a configurable functional unit and method using such a data processor
US6594822B1 (en) * 1999-02-19 2003-07-15 Nortel Networks Limited Method and apparatus for creating a software patch by comparing object files
US6460122B1 (en) * 1999-03-31 2002-10-01 International Business Machine Corporation System, apparatus and method for multi-level cache in a multi-processor/multi-controller environment
GB2352548B (en) * 1999-07-26 2001-06-06 Sun Microsystems Inc Method and apparatus for executing standard functions in a computer system
US6385757B1 (en) * 1999-08-20 2002-05-07 Hewlett-Packard Company Auto design of VLIW processors
US6684193B1 (en) * 1999-10-05 2004-01-27 Rapt Technologies Corporation Method and apparatus for multivariate allocation of resources
FR2802661B1 (en) * 1999-12-21 2003-10-31 Bull Sa HIGH SPEED RANDOM NUMBER GENERATOR
US6526572B1 (en) * 2000-02-09 2003-02-25 Hewlett-Packard Company Mechanism for software register renaming and load speculation in an optimizer
US8095508B2 (en) * 2000-04-07 2012-01-10 Washington University Intelligent data storage and processing using FPGA devices
US7139743B2 (en) * 2000-04-07 2006-11-21 Washington University Associative database scanning and information retrieval using FPGA devices
US6362650B1 (en) * 2000-05-18 2002-03-26 Xilinx, Inc. Method and apparatus for incorporating a multiplier into an FPGA
US6594736B1 (en) * 2000-08-15 2003-07-15 Src Computers, Inc. System and method for semaphore and atomic operation management in a multiprocessor
US6751792B1 (en) * 2000-10-04 2004-06-15 Sun Microsystems, Inc. Using value-expression graphs for data-flow optimizations
US7155602B2 (en) * 2001-04-30 2006-12-26 Src Computers, Inc. Interface for integrating reconfigurable processors into a general purpose computing system
US20030046492A1 (en) * 2001-08-28 2003-03-06 International Business Machines Corporation, Armonk, New York Configurable memory array
US6985918B2 (en) * 2001-10-17 2006-01-10 Hewlett-Packard Development Company, L.P. Random number generators implemented with cellular array
US6839893B2 (en) * 2001-10-18 2005-01-04 International Business Machines Corporation Debugger capable of providing warnings for unreachable breakpoints
US6931488B2 (en) * 2001-10-30 2005-08-16 Sun Microsystems, Inc. Reconfigurable cache for application-based memory configuration
US6836823B2 (en) * 2001-11-05 2004-12-28 Src Computers, Inc. Bandwidth enhancement for uncached devices
US7024519B2 (en) * 2002-05-06 2006-04-04 Sony Computer Entertainment Inc. Methods and apparatus for controlling hierarchical cache memory
US7093023B2 (en) * 2002-05-21 2006-08-15 Washington University Methods, systems, and devices using reprogrammable hardware for high-speed processing of streaming data to find a redefinable pattern and respond thereto
US6714041B1 (en) * 2002-08-30 2004-03-30 Xilinx, Inc. Programming on-the-fly (OTF)
US7225324B2 (en) * 2002-10-31 2007-05-29 Src Computers, Inc. Multi-adaptive processing systems and techniques for enhancing parallelism and performance of computational functions
US6983456B2 (en) * 2002-10-31 2006-01-03 Src Computers, Inc. Process for converting programs in high-level programming languages to a unified executable for hybrid computing platforms
US20040225483A1 (en) * 2003-02-24 2004-11-11 Michal Okoniewski Fdtd hardware acceleration system
US8271369B2 (en) * 2003-03-12 2012-09-18 Norman Gilmore Financial modeling and forecasting system
US20040186804A1 (en) * 2003-03-19 2004-09-23 Anindya Chakraborty Methods and systems for analytical-based multifactor multiobjective portfolio risk optimization
US7149867B2 (en) * 2003-06-18 2006-12-12 Src Computers, Inc. System and method of enhancing efficiency and utilization of memory bandwidth in reconfigurable hardware
US7366326B2 (en) * 2003-06-24 2008-04-29 University Of Maryland, Baltimore County Real-time implementation of field programmable gate arrays (FPGA) design in hyperspectral imaging
US20050033672A1 (en) * 2003-07-22 2005-02-10 Credit-Agricole Indosuez System, method, and computer program product for managing financial risk when issuing tender options
US7174432B2 (en) * 2003-08-19 2007-02-06 Nvidia Corporation Asynchronous, independent and multiple process shared memory system in an adaptive computing architecture
US20050177485A1 (en) * 2004-02-09 2005-08-11 William Peter Method for rapid and accurate pricing of options and other derivatives
US20050209959A1 (en) * 2004-03-22 2005-09-22 Tenney Mark S Financial regime-switching vector auto-regression
US20050251375A1 (en) * 2004-05-07 2005-11-10 Lukasz Salwinski Method for simulating the dynamics of biological networks in silico
US7444454B2 (en) * 2004-05-11 2008-10-28 L-3 Communications Integrated Systems L.P. Systems and methods for interconnection of multiple FPGA devices
US7506297B2 (en) * 2004-06-15 2009-03-17 University Of North Carolina At Charlotte Methodology for scheduling, partitioning and mapping computational tasks onto scalable, high performance, hybrid FPGA networks
US20050288800A1 (en) * 2004-06-28 2005-12-29 Smith William D Accelerating computational algorithms using reconfigurable computing technologies
EP1784719A4 (en) * 2004-08-24 2011-04-13 Univ Washington Methods and systems for content detection in a reconfigurable hardware

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6539438B1 (en) 1999-01-15 2003-03-25 Quickflex Inc. Reconfigurable computing system and method and apparatus employing same

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
DAVIS D ET AL: "A JAVA DEVELOPMENT AND RUNTIME ENVIRONMENT FOR RECONFIGURABLE COMPUTING", PROCEEDINGS. IEEE SYMPOSIUM ON PARALLEL AND DISTRIBUTED PROCESSING. PROCEEDINGS OF SPDP, March 1998 (1998-03-01), pages 43 - 48, XP000827565 *
KIM M ET AL: "IMPLEMENTATION OF FPGA BASED FAST UNITARY MUSIC DOA ESTIMATOR", IEICE TRANSACTIONS ON ELECTRONICS, ELECTRONICS SOCIETY, TOKYO, JP, vol. E87-C, no. 9, September 2004 (2004-09-01), pages 1485 - 1494, XP001209718, ISSN: 0916-8524 *

Also Published As

Publication number Publication date
EP1955191A1 (en) 2008-08-13
TW200741479A (en) 2007-11-01
CA2631517A1 (en) 2007-06-14
US20070192241A1 (en) 2007-08-16
US7716100B2 (en) 2010-05-11

Similar Documents

Publication Publication Date Title
US7716100B2 (en) Methods and systems for computing platform
Salle et al. Efficient sampling and meta-modeling for computational economic models
Luo et al. The t copula with multiple parameters of degrees of freedom: bivariate characteristics and application to risk management
Arvanitis et al. Stochastic spanning
Bargigli et al. Random digraphs with given expected degree sequences: a model for economic networks
Ambikasaran et al. Fast direct methods for Gaussian processes and the analysis of NASA Kepler mission data
US20030101122A1 (en) Generator libraries
Ji et al. A stochastic linear goal programming approach to multistage portfolio management based on scenario generation via linear programming
Dimov et al. Computational challenges in the numerical treatment of large air pollution models
Hearin et al. DSPS: Differentiable stellar population synthesis
Angün A risk-averse approach to simulation optimization with multiple responses
Ökten et al. Parameterization based on randomized quasi-Monte Carlo methods
Osthus et al. Dynamic linear models for forecasting of radiation belt electrons and limitations on physical interpretation of predictive models
Chen et al. A regression-based calibration method for agent-based models
Aunsri et al. A time-varying Bayesian compressed vector autoregression for macroeconomic forecasting
Gourieroux et al. Generalized covariance estimator
Cai et al. Forecasting large scale conditional volatility and covariance using neural network on GPU
Fokianos et al. Binary time series models driven by a latent process
Dixon et al. Monte Carlo–based financial market value-at-risk estimation on GPUs
Chan et al. Handbook of financial risk management: Simulations and case studies
Artemiev et al. Numerical solution of stochastic differential equations on Supercomputers
Hintz et al. Computational Challenges of t and Related Copulas.
Yuan et al. pyvine: the Python package for regular vine copula modeling, sampling and testing
Poncet et al. Monte Carlo Simulations
Halada et al. Optimal multistage portfolio management using a parallel interior point method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application
WWE Wipo information: entry into national phase

Ref document number: 2631517

Country of ref document: CA

WWE Wipo information: entry into national phase

Ref document number: 2006838067

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: DE