Accelerating financial applications on the GPU

Grauer-Gray, Scott; Killian, William; Schleicher, Robert; Cavazos, John

doi:10.1145/2458523.2458536

Cited by 34 publications

(13 citation statements)

References 14 publications

(12 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In the parallel implementation, a different random number generator is required for each simulation [4]. To ensure the results are statistically uncorrelated, the generators should have the same seed but different sequence for each simulation.…”

Section: Monte Carlo Methodsmentioning

confidence: 99%

GPU option pricing

Suo

Zhu

Attridge

et al. 2015

Proceedings of the 8th Workshop on High Performance Computational Finance

View full text Add to dashboard Cite

In this paper, we explore the possible approaches to harness extra computing power from commodity hardware to speedup pricing calculation of individual options. Specifically, we leverage two parallel computing platforms: Open Computing Language (OpenCL) and Compute United Device Architecture (CUDA). We propose several parallel implementations of the two most popular numerical methods of option pricing: Lattice model and Monte Carlo method. In the end, we show that the parallel implementations achieve significant performance improvement over serial implementations.

show abstract

Section: Monte Carlo Methodsmentioning

confidence: 99%

GPU option pricing

Suo

Zhu

Attridge

et al. 2015

Proceedings of the 8th Workshop on High Performance Computational Finance

View full text Add to dashboard Cite

show abstract

“…In [18], the authors reported an improvement in optimizing the Monte Carlo approach for the averages of 10 000 iterations, which overcame the limitations of the CPUs even for a moderate set of options. In [9], a significant speedup was achieved for the parallelization of the Black-Scholes, Monte Carlo, Bonds, and Repo code paths from the QuantLib library using hand-written CUDA and OpenCL codes that were specifically targeted for the GPU. In [7], the authors employed the Least Squared Monte Carlo method on a GPU, which led to a reduction of computing time for the numerical pricing of European Multi-dimensional options based on the COS Method.…”

Section: Literature and Related Work 21 The Performance Of Option Pmentioning

confidence: 99%

Lower Precision calculation for option pricing

Ścibisz-Mordelska

Nielek

2017

csci

View full text Add to dashboard Cite

The problem of option pricing is one of the most critical issues and fundamental building blocks in mathematical finance. The research includes the deployment of a lower-precision type in two option-pricing algorithms: the Black-Scholes and Monte Carlo simulations. We make an assumption that the shorter the number used for calculations is (in bits), the more operations we are able to perform at the same time. The results are examined by a comparison to the outputs of single-and double-precision types. The major goal of the study is to indicate whether the lower-precision types can be used in financial mathematics. The findings indicate that Black-Scholes provided more precise outputs than the basic implementation of the Monte Carlo simulation. Modification of the Monte Carlo algorithm is also proposed. The research shows the limitations and opportunities of the lower-precision type usage. In order to benefit from the application in terms of the time of calculation, the improved algorithms can be implemented on a GPU or FPGA. We conclude that, under particular restrictions, the lower-precision calculation can be used in mathematical finance.

show abstract

“…In the end, the most optimal configuration that allows the fastest search gets chosen. Auto-tuning is widely applied in computer science (Williams, 2008), but has also seen applications in other domains such as computational finance (Grauer-Gray et al, 2013) or astronomy (Sclocco et al, 2012). Auto-tuning for radio transient surveys also shows promising results in terms of performance portability (Sclocco et al, 2015).…”

Section: Introductionmentioning

confidence: 99%

The Apertif Monitor for Bursts Encountered in Real-time (AMBER) auto-tuning optimization with genetic algorithms

Mikhailov

Sclocco

2018

Astronomy and Computing

View full text Add to dashboard Cite

Real-time searches for faint radio pulses from unknown radio transients are computationally challenging. Detections become further complicated due to continuously increasing technical capabilities of transient surveys: telescope sensitivity, searched area of the sky, number of antennas or dishes, temporal and frequency resolution. The new Apertif transient survey on the Westerbork telescope happens in real-time on GPUs by means of the single-pulse search pipeline AMBER (Sclocco, 2017). AMBER initially carries out auto-tuning: it finds the most optimal configuration of user-controlled parameters per each of four pipeline kernels so that each kernel performs its task as fast as possible. The pipeline uses a brute-force (BF) exhaustive search which in total takes 5 -24 hours to run depending on the processing cluster architecture. We apply more heuristic, biologically driven genetic algorithms (GAs) to limit the exploration of the total parameter space, tune all four kernels together and reduce the tuning time to few hours. Our results show that after only few hours of tuning, GAs always find similar or even better configurations for all kernels together than the combination of single kernel configurations tuned by the BF approach. At the same time, by means of their genetic operators, GAs converge into better solutions than those obtained by pure random searches. The explored multi-dimensional parameter space is very complex and has multiple local optima as the evolution of randomly generated configurations does not always guarantee global solution. 1 Other options, such as FPGAs and ASICs, are also available. However, FPGAs are very hard to program, and floating point performance is not comparable with GPUs, whereas ASIC are expensive to design and produce.

show abstract

Accelerating financial applications on the GPU

Cited by 34 publications

References 14 publications

GPU option pricing

GPU option pricing

Lower Precision calculation for option pricing

The Apertif Monitor for Bursts Encountered in Real-time (AMBER) auto-tuning optimization with genetic algorithms

Contact Info

Product

Resources

About