Credit Modeling with Supercomputing

By Bill Blake

September 14, 2007

The creation and deployment of new numerical methods for economic and financial modeling is becoming a critical competitive weapon for banks, hedge funds and other investment firms. From a high performance computing (HPC) perspective, when quantitative analysts are asked how fast these new computations need to be processed, their answer is usually “at least fifteen minutes faster than our competitors with a lot of extra credit for finishing before the close of daily trading.” Consequently, computing requirements on Wall Street are growing exponentially as algorithms and models become more complex to support new investment opportunities, while incorporating ever larger data sets.

But the desktop computers that are used to develop these ever-growing financial codes are inadequate to support full scale production deployment, prompting investments firms to turn to HPC systems such as parallel servers, clusters or grids. Fortunately these systems now employ cost-effective multi-core processors from Intel and AMD, and as a result parallel supercomputing is finally accessible to Wall Street firms.

The problem is the new parallel hardware is like a fast highway that leads to a software wall. Parallel HPCs are unfamiliar computing platforms to most financial analysts, who are accustomed to working with popular mathematical tools such as MATLAB, Python and R to produce their financial models. Computing these models on parallel systems typically requires a team of highly trained programmers to rewrite hundreds of lines of analyst-written VHLL code into thousands of lines of complex C, C++, or FORTRAN involving the Message Passing Interface (MPI) or equivalent manual parallelization techniques. This redesign can take months of time while preventing the interactive experimentation and refinement of the models financial analysts require. These complex coding techniques require programmers versed in parallel programming. And programmers with those skills are not only expensive, they are in short supply in the financial services sector.

Consider the case of Julius Finance, a Wall Street research company that specializes in credit modeling analysis. The company focuses on credit derivative products, analyzing the relative valuation of synthetic collateralized debt obligations (CDOs). The computationally challenging analysis of credit factors such as spread, credit rating, foreign exchange and interest rates for a wide variety of corporate investments has made the credit derivative market a black art at best.

Until now, these financial products have been priced by investment firms using Copula models, a popular approach for modeling dependencies between random variables thanks to their relative mathematical simplicity. But the trade-off for this simplicity has been inconsistent, unconvincing results. “Existing mathematical frameworks for CDO valuation are far from compelling…to put it mildly,” says Peter Cotton, CEO of Julius Finance. “This is not surprising, as rigorous evaluation of credit models is prohibitively time consuming in any conventional research setup.”

Cotton knew that employing new, more sophisticated algorithmic models on massive amounts of variable financial data would give the company a tremendous competitive advantage when it came to making more accurate predictions about a portfolio’s potential.

The company installed a Linux-based cluster to provide the necessary processing power and memory capacity. But rather than employ computer scientists to parallelize the models, Julius Finance took a different approach, using Star-P software to transparently bridge analysts’ desktops with the Linux cluster. This way, the company’s analysts could continue to work in their familiar MATLAB environment, and enable their applications to run on the parallel clusters without reprogramming.

This interactive supercomputing approach allows for continual feedback and refinement from prototype to production, resulting in higher quality models, algorithms and, ultimately, much more accurate portfolio predictions. The company gained a quantum leap in computational performance to handle the massive data sets and model complexities, without having to lose the interactivity and ease of use of their desktop environment. “We took this approach to reduce prototyping time and facilitate memory intensive experiments…looking under more rocks, as it were, and finding very interesting things,” says Cotton.

Julius Finance is part of a growing trend on Wall Street establishing HPCs as a critical resource in the IT data center. The reason: as financial applications become more complex and more compute-intensive, the ability to offer real-time results diminishes with desktop-bound computing. And the big challenge on Wall Street is in providing actionable financial analysis before the window of opportunity closes. Shrinking the “time to solution at full scale” can offer tremendous competitive differentiation to investment firms.

Beyond the specific area of credit modeling, speeding up computations and scenario analyses is critical to all aspects of  financial services — including trading desks, risk management desks, etc. — because each component, while a relatively small part of the overall environment, is potentially computationally expensive. Whether the reactions are in nearly real time, on an hourly basis, or at the end of the day, the decisions could often be improved if they included more trajectories, more scenarios.  The models are dynamic — with frequent updates with new parameters — so flexibility in algorithm development and production deployment is key.

This new interactive supercomputing model can be generalized to a variety of financial analytical applications ranging from numerically-intensive workloads in simulation, optimization and valuation to data-intensive workloads performing pattern detection for fraud detection and trading opportunities, such as:

Monte Carlo Simulation – These simulations have many advantages, including the ease of implementation and applicability to multi-dimensional problems commonly encountered in finance.  However, calculation using Monte Carlo techniques is very time consuming due to the need for simulating many trajectories with multiple parameters. 

Portfolio Optimization — Taking an interactive supercomputing approach, analysts can run their models on parallel systems to optimize thousands of individual portfolios overnight based on the previous day’s trading results. Commercial and open source optimization libraries such as Axioma or CPLEX can typically be plugged in and executed in parallel – all from within the analyst’s desktop application.

Valuation of Financial Derivatives — Valuing financial derivatives is computationally intensive, and requires large amounts of computer time. A re-insurance firm, for example, may need to value and compute hedge strategies for hundreds of thousands of policy holders in its portfolio on a regular and timely basis. Analysts need to be able to explore new valuation methodologies from their desktop, using high performance computers to run billions of complex scenarios.

Detection of Credit Card Fraud — The rise of identity theft together with the popularity of online shopping has resulted in a huge increase in credit card fraud. As thieves become increasingly shrewd in exploiting security weaknesses, banks and credit card companies need to be extremely agile to stay ahead of them. Parallel HPCs enable a bank to easily run more sophisticated fraud detection algorithms against tens of millions of credit card accounts.

Hedge Fund Trading — In balancing a large portfolio of stocks, analysts need to search for short- and long-term patterns, identify correlations between securities, and develop forecasts. Intense computations are required against terabyte-sized “tick store” databases — potentially a decade or more of trading data for thousands of securities. HPCs allow for faster reaction time to market conditions, enabling analysts to evaluate more sophisticated algorithms that take into account larger data sets.

Until now, investment firms faced an “either-or” dilemma by choosing to live with the performance limitations of their desktop systems, or engaging a team of programmers to re-code their algorithms to take advantage of powerful parallel servers or clusters. But that situation changes with new interactive supercomputing models that provide a “both-and” opportunity combining the productivity breakthroughs of easy to use desktop development with a seamless transition to deployment of large, complex financial simulations on parallel servers. Analysts can focus on rapidly delivering the most accurate, comprehensive and actionable intelligence by leveraging abundant parallel system resources without the need for scarce human resources.

—–

About the Author

Bill Blake is the Chief Executive Officer of Interactive Supercomputing Inc. He brings more than two decades of senior executive experience in developing high performance computing systems. He joins ISC from Netezza, where he was senior vice president of product development for the high-performance data warehouse appliance company. Bill previously was vice president of high performance technical computing at Compaq, where he led development and marketing efforts. He received undergraduate and graduate degrees in Electrical Engineering at the Lowell Technological Institute, and is a member of the Institute of Electrical and Electronics Engineers (IEEE), the Association for Computing Machinery (ACM), and the American Association of Artificial Intelligence. Bill is a member of the board of directors of supercomputing pioneer, Cray Inc. Supercomputers, as well as Etnus, Inc., a provider of analytical software for developing complex computer code.

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industy updates delivered to you every week!

Data Vortex Users Contemplate the Future of Supercomputing

October 19, 2017

Last month (Sept. 11-12), HPC networking company Data Vortex held its inaugural users group at Pacific Northwest National Laboratory (PNNL) bringing together about 30 participants from industry, government and academia t Read more…

By Tiffany Trader

AI Self-Training Goes Forward at Google DeepMind

October 19, 2017

DeepMind, Google’s AI research organization, announced today in a blog that AlphaGo Zero, the latest evolution of AlphaGo (the first computer program to defeat a Go world champion) trained itself within three days to play Go at a superhuman level (i.e., better than any human) – and to beat the old version of AlphaGo – without leveraging human expertise, data or training. Read more…

By Doug Black

Researchers Scale COSMO Climate Code to 4888 GPUs on Piz Daint

October 17, 2017

Effective global climate simulation, sorely needed to anticipate and cope with global warming, has long been computationally challenging. Two of the major obstacles are the needed resolution and prolonged time to compute Read more…

By John Russell

HPE Extreme Performance Solutions

Transforming Genomic Analytics with HPC-Accelerated Insights

Advancements in the field of genomics are revolutionizing our understanding of human biology, rapidly accelerating the discovery and treatment of genetic diseases, and dramatically improving human health. Read more…

Student Cluster Competition Coverage New Home

October 16, 2017

Hello computer sports fans! This is the first of many (many!) articles covering the world-wide phenomenon of Student Cluster Competitions. Finally, the Student Cluster Competition coverage has come to its natural home: H Read more…

By Dan Olds

Data Vortex Users Contemplate the Future of Supercomputing

October 19, 2017

Last month (Sept. 11-12), HPC networking company Data Vortex held its inaugural users group at Pacific Northwest National Laboratory (PNNL) bringing together ab Read more…

By Tiffany Trader

AI Self-Training Goes Forward at Google DeepMind

October 19, 2017

DeepMind, Google’s AI research organization, announced today in a blog that AlphaGo Zero, the latest evolution of AlphaGo (the first computer program to defeat a Go world champion) trained itself within three days to play Go at a superhuman level (i.e., better than any human) – and to beat the old version of AlphaGo – without leveraging human expertise, data or training. Read more…

By Doug Black

Student Cluster Competition Coverage New Home

October 16, 2017

Hello computer sports fans! This is the first of many (many!) articles covering the world-wide phenomenon of Student Cluster Competitions. Finally, the Student Read more…

By Dan Olds

Intel Delivers 17-Qubit Quantum Chip to European Research Partner

October 10, 2017

On Tuesday, Intel delivered a 17-qubit superconducting test chip to research partner QuTech, the quantum research institute of Delft University of Technology (TU Delft) in the Netherlands. The announcement marks a major milestone in the 10-year, $50-million collaborative relationship with TU Delft and TNO, the Dutch Organization for Applied Research, to accelerate advancements in quantum computing. Read more…

By Tiffany Trader

Fujitsu Tapped to Build 37-Petaflops ABCI System for AIST

October 10, 2017

Fujitsu announced today it will build the long-planned AI Bridging Cloud Infrastructure (ABCI) which is set to become the fastest supercomputer system in Japan Read more…

By John Russell

HPC Chips – A Veritable Smorgasbord?

October 10, 2017

For the first time since AMD's ill-fated launch of Bulldozer the answer to the question, 'Which CPU will be in my next HPC system?' doesn't have to be 'Whichever variety of Intel Xeon E5 they are selling when we procure'. Read more…

By Dairsie Latimer

Delays, Smoke, Records & Markets – A Candid Conversation with Cray CEO Peter Ungaro

October 5, 2017

Earlier this month, Tom Tabor, publisher of HPCwire and I had a very personal conversation with Cray CEO Peter Ungaro. Cray has been on something of a Cinderell Read more…

By Tiffany Trader & Tom Tabor

Intel Debuts Programmable Acceleration Card

October 5, 2017

With a view toward supporting complex, data-intensive applications, such as AI inference, video streaming analytics, database acceleration and genomics, Intel i Read more…

By Doug Black

Reinders: “AVX-512 May Be a Hidden Gem” in Intel Xeon Scalable Processors

June 29, 2017

Imagine if we could use vector processing on something other than just floating point problems.  Today, GPUs and CPUs work tirelessly to accelerate algorithms Read more…

By James Reinders

NERSC Scales Scientific Deep Learning to 15 Petaflops

August 28, 2017

A collaborative effort between Intel, NERSC and Stanford has delivered the first 15-petaflops deep learning software running on HPC platforms and is, according Read more…

By Rob Farber

Oracle Layoffs Reportedly Hit SPARC and Solaris Hard

September 7, 2017

Oracle’s latest layoffs have many wondering if this is the end of the line for the SPARC processor and Solaris OS development. As reported by multiple sources Read more…

By John Russell

US Coalesces Plans for First Exascale Supercomputer: Aurora in 2021

September 27, 2017

At the Advanced Scientific Computing Advisory Committee (ASCAC) meeting, in Arlington, Va., yesterday (Sept. 26), it was revealed that the "Aurora" supercompute Read more…

By Tiffany Trader

How ‘Knights Mill’ Gets Its Deep Learning Flops

June 22, 2017

Intel, the subject of much speculation regarding the delayed, rewritten or potentially canceled “Aurora” contract (the Argonne Lab part of the CORAL “ Read more…

By Tiffany Trader

Google Releases Deeplearn.js to Further Democratize Machine Learning

August 17, 2017

Spreading the use of machine learning tools is one of the goals of Google’s PAIR (People + AI Research) initiative, which was introduced in early July. Last w Read more…

By John Russell

GlobalFoundries Puts Wind in AMD’s Sails with 12nm FinFET

September 24, 2017

From its annual tech conference last week (Sept. 20), where GlobalFoundries welcomed more than 600 semiconductor professionals (reaching the Santa Clara venue Read more…

By Tiffany Trader

Nvidia Responds to Google TPU Benchmarking

April 10, 2017

Nvidia highlights strengths of its newest GPU silicon in response to Google's report on the performance and energy advantages of its custom tensor processor. Read more…

By Tiffany Trader

Leading Solution Providers

Graphcore Readies Launch of 16nm Colossus-IPU Chip

July 20, 2017

A second $30 million funding round for U.K. AI chip developer Graphcore sets up the company to go to market with its “intelligent processing unit” (IPU) in Read more…

By Tiffany Trader

Amazon Debuts New AMD-based GPU Instances for Graphics Acceleration

September 12, 2017

Last week Amazon Web Services (AWS) streaming service, AppStream 2.0, introduced a new GPU instance called Graphics Design intended to accelerate graphics. The Read more…

By John Russell

EU Funds 20 Million Euro ARM+FPGA Exascale Project

September 7, 2017

At the Barcelona Supercomputer Centre on Wednesday (Sept. 6), 16 partners gathered to launch the EuroEXA project, which invests €20 million over three-and-a-half years into exascale-focused research and development. Led by the Horizon 2020 program, EuroEXA picks up the banner of a triad of partner projects — ExaNeSt, EcoScale and ExaNoDe — building on their work... Read more…

By Tiffany Trader

Delays, Smoke, Records & Markets – A Candid Conversation with Cray CEO Peter Ungaro

October 5, 2017

Earlier this month, Tom Tabor, publisher of HPCwire and I had a very personal conversation with Cray CEO Peter Ungaro. Cray has been on something of a Cinderell Read more…

By Tiffany Trader & Tom Tabor

Cray Moves to Acquire the Seagate ClusterStor Line

July 28, 2017

This week Cray announced that it is picking up Seagate's ClusterStor HPC storage array business for an undisclosed sum. "In short we're effectively transitioning the bulk of the ClusterStor product line to Cray," said CEO Peter Ungaro. Read more…

By Tiffany Trader

Intel Launches Software Tools to Ease FPGA Programming

September 5, 2017

Field Programmable Gate Arrays (FPGAs) have a reputation for being difficult to program, requiring expertise in specialty languages, like Verilog or VHDL. Easin Read more…

By Tiffany Trader

IBM Advances Web-based Quantum Programming

September 5, 2017

IBM Research is pairing its Jupyter-based Data Science Experience notebook environment with its cloud-based quantum computer, IBM Q, in hopes of encouraging a new class of entrepreneurial user to solve intractable problems that even exceed the capabilities of the best AI systems. Read more…

By Alex Woodie

HPC Chips – A Veritable Smorgasbord?

October 10, 2017

For the first time since AMD's ill-fated launch of Bulldozer the answer to the question, 'Which CPU will be in my next HPC system?' doesn't have to be 'Whichever variety of Intel Xeon E5 they are selling when we procure'. Read more…

By Dairsie Latimer

  • arrow
  • Click Here for More Headlines
  • arrow
Share This