News Articles

Articles

Cozy up with CUDA, Wilmott Magazine

‘CUDA’ is the magic word to direct the newly enhanced SciFinance to CUDA-enable your Monte Carlo pricing models

"...Obviously, speed of model development is important, but so is the execution time of the pricing model. SciComp has partnered with NVIDIA, and we are now taking advantage of their graphics processing unit (GPU) technology. NVIDIA developed a programming language called CUDA that allows you to develop programs that efficiently run on the GPU card. Code that takes advantage of the GPU’s parallelism exhibits tremendous increases in execution speed compared to serial code.

SciFinance can now automatically create CUDA-enabled pricing and risk models, so a customer need only add the keyword ‘CUDA’ to an existing serial model specification. Depending on the type of pricing model, it can run from 30 to 120 times faster than the comparable serial code. One of the best parts from a customer’s perspective is that they do not need to have any CUDA or parallel computing experience, SciFinance takes care of all of that."

Request a password to the SciComp Resource Center to read the entire article.

Wall Street Accelerates Options Analysis with GPU Technology, Wall Street and Technology

"...One of the tools NVIDIA has introduced, and a key driver behind the adoption of GPUs in financial markets, is a programming environment called CUDA, which was released two years ago. It is this CUDA architecture inside the chip (as well as a CUDA toolkit and a C-compiler) that enables software developers to tap into the parallel architecture of the GPU using the C programming language.

Based in Austin, Texas, SciComp, a developer of derivatives software, began using NVIDIA's CUDA technology in 2007 so that it could parallelize its code. The vendor's SciFinance product "actually generates the source code for the pricing and risk models," explains Curt Randall, EVP and head of financial engineering as well as a founder of SciComp.

The thousands of lines of C or C++ code generated by SciFinance can help a quant produce a new pricing model in a matter of hours, says Randall. "The compiler and the language make the hardware accessible to general programming as opposed to graphics programming," says Randall, who contends that the NVIDIA hardware can enable the code to run 30 to 50 times faster than standard PCs. "Programmers don't have to understand the nuances of how to parallelize code. All of that is done for them automatically."

Tesla 10 & CUDA 2.0: Real-World Applications & Financials: SciComp's SciFinance Beyond3D

"..Traditionally, traders who imagined a deal talked to ‘quants’ (PhDs in mathematics or physics who are specialized in finance) and the mathematical equations they came up with for the pricing model then had to be transformed into a program through weeks of hand-coding. After this error-prone process, the program had to be run to estimate the price.

SciFinance’s approach is to automate program generation; instead of manual coding, the quants can ‘simply’ describe their equations in a specialized language. The process is much quicker and much more reliable. Therefore, the relative length of running the program increases and it becomes more of a bottleneck. The solution they came up with is, of course, FPGAs. Uhhh, wait, no, I meant CUDA obviously:

What may not be as obvious in that image is that the only thing you need to add to get GPGPU acceleration is literally ‘CUDA’; it’s a single keyword, not a fundamentally different way to formulate the math equations. This allows SciComp’s customers to save even more time while also improving accuracy (through a higher number of iterations), both of which result in obvious financial benefits.

The difference between merely being able to demonstrate the acceleration of a basic stock option pricing model and this is huge. The amount of work involved is likely ridiculously higher, and it’s a pretty good example of how much far along CUDA is in terms of software..."

SciComp Accelerates Derivatives Pricing with CUDA, InsideHPC

"...SciComp is one of the first companies to fully embrace the potential that GPUs have in the field of computational finance,” said Andy Keane, general manager of the GPU Computing business at NVIDIA. “The ability to not just deliver small and incremental increases in performance, but instead to deliver 100X and reduce weeks of hand coding to immediate, real-time results is incredibly powerful. We look forward to working closely with SciComp going forward to bring more defining improvements to the SciFinance generated pricing models and in turn their customers’ businesses."

General Monte Carlo Greeks in Practice, Wilmott Magazine

"...The simplest general approach for estimating the Greeks is based on finite differences, in which the Monte Carlo pricing function is called to revalue the derivative at perturbed parameters and a finite difference is applied to approximate the partial derivatives (price sensitivities). The advantage of such estimation lies in its independence of the underlying model and payoff structure, enabling a generic implementation with little additional programming..."

Computer Software That Writes Itself Newsweek International

"One automatic programming tool has already made it into the financial marketplace. SciComp, based in Austin, Texas, has developed a product that helps investment banks design programs to price financial derivatives. It takes complex mathematical models and translates them into something a computer can solve, allowing banks to flexibly change pricing models as they introduce new products and guidelines."

The Electronic Quant, Risk Magazine

"SciComp, based in Austin, Texas, has produced an artificial intelligence package that will turn an analyst's options pricing model into a computer program."

Code-Crunching for Option Pricing, Derivatives Strategy

"Anything you can write in terms of a partial differential equation, the software can price."

If the Skew Fits, Risk Magazine

"Smiles, smirks, skews, sneers...these rather whimsical terms inthe volatility vernacular describe an important reality: implied Black-Scholes volatilities are not constant...In this article, we describe a very simple model of the local volatility type that nevertheless provides a rich framework for capturing volatility smiles and skews."

SciComp Automates Algorithm Production at Merrill, Derivatives & Risk Technology

Merrill to Use SciFinance to Price Equity Derivatives Derivatives & Risk Technology

Bear, Stearns Chooses SciFinance for Risk Control, Wall Street & Technology

Bear Stearns Securities Takes SciFinance, Derivatives & Risk Technology

 

News clippings

 

Computer Software That Writes Itself, Newsweek International

"The benefits of automatic software are compelling. Companies would need fewer programmers and could ratchet up productivity. Humans writing computer code are also prone to errors. "If a programmer can sit down, specify what you want and push a button, you end up much more productive," says Doug Smith, a researcher at the Kestrel Institute, a nonprofit RD center in Palo Alto, California. "It's the next stage in the evolution of computer programming."

SciComp Inc., Risk.net

"...SciComp extended the instrument coverage of its SciFinance automated pricing model generator to include, among others, cash collateralised debt obligations (CCDOs), CDO-squareds and credit-linked range accrual notes. A new generic short rate and hazard rate calibration module calibrates both the short and hazard rates (correlated), using models such as one-factor, Gaussian and extended exponential Vasichek."

BOCI Choose Scicomp, Wilmott Magazine

"Deciding factors for BOCI purchasing SciFinace included the ability to solve problems through a high-level declarative language rather than hand coding," said George Advani, a quantitative analyst for BOCI. We will use SciFinance to value and hedge equity derivative instruments including baskets and exotics."

SciComp strategy, Risk Magazine

"...the company plans to release a new examples catalogue for its automated pricing model generator SciFinance, which will enable users to code and price very exotic derivatives. An upgraded version of SciIntegrator will provide extra functionality for developers using Microsoft's .Net and Java..."

SciComp: Weapon of math instruction, Wilmott Magazine

"Numerical solution of partial differential equations (PDEs) is the optimal method for pricing convertible bonds. But writing PDE codes by hand can be time consuming and tricky. Not anymore."

SciComp releases SciCal calibrator for SV and SVJ, Numa.web

"SciComp Inc. has released SciCal software for calibrating stochastic volatility (SV) and stochastic volatility + jumps (SVJ) pricing models to market data. The modern search variant used in SciCal is very fast --SciCal can calibrate from a handful to hundreds of options to market data in under a minute. "

Turbo-charged models, Risk Magazine

"...Fortis originally used a Monte Carlo application developed in-house, but has now adopted SciMC from Texas-based SciComp, which will automatically generate Monte Carlo code from simple specifications the user makes in a special language called Aspen."

Fast Monte Carlo programming, Risk Magazine

"Users own the code the product produces, which does not need run-time licenses. Using the company's ASPEN language, users can specify any number of stochastic differential equations that describe underlying processes, a payout discount and sensitivity functions."

Texas-based SciComp has developed SciMC, Finextra.com

"SciMC handles sophisticated mathematical and financial features including exotic path dependencies, jumps, high dimensionality, deterministic sequences, choice of variance reduction techniques, and American or Bermudan exercise."

ABN Amro equities unit in SciComp deal, Risk News

"SciFinance automatically generates custom codes - called C codes - based on brief specifications of derivatives structures. ABN Amro believes the system will reduce both the time and costs involved in developing a new model, as it can automate the process as well as formalising the concepts that analysts and traders use."

ABN AMRO has licensed SciFinance and SciXL, Finextra.com

"Andrew Greaves, managing director at ABN Amro, says: "We will use SciXL to run models in Excel without extra programming steps. Together, SciFinance and SciXL will reduce the time and cost involved with new model development."

Dutch bank buys SciComp software, Austin Business Journal

SciComp awarded patent, Austin American-Statesman

SciComp has been issued a patent, Wall Street & Technology Week

SciXL from SciComp takes the pain out of producing Microsoft Excel add-ins, Financial Engineering News

"SciXL is a breakthrough, since it builds the add-in for you, and creates and maintains all the code that links the add-in programs to Excel. In addition, add-ins written in C++ perform much faster and are more stable than those written in Visual Basic."