Numerical Technologies invests a large part of its R&D towards the application of HPC to financial risk management. It focuses its efforts on cluster computing and GPU computing and develops financial risk management software applications that take full advantage of HPC systems.

GPU computing uses a computer’s GPU (or specialized accelerator) to accelerate the processing of calculation-intensive tasks, enabling users to attain massively parallel processing power on servers and even on personal computers. There has been increasing commercial application of GPU-accelerated systems in recent years and NVIDIA reported that the November 2011 Top500 Supercomputer Sites lists 35 Tesla-GPU-accelerated supercomputers, 3 of which are in the top 5. However, despite its advantages, the growth of GPU computing in the finance industry has been hindered by the difficulty associated with CUDA, the de facto GPU computing standard from NVIDIA.

Numerical Technologies recognizes the economic potential of GPU computing and is at the forefront of GPU computing R&D in financial risk management. As part of its research, Numerical Technologies has developed NtParallel an experimental API library that encapsulates CUDA programs and helps businesses tap the power of GPU computing without the difficult CUDA development. You can call NtParallel from C, C++, Java, and VBA for Microsoft Excel.

NtParallel benchmark tests comparing Mandelbrot and Black Scholes computation speed on NtParallel and normal CPU.

Learn more about NtParallel at www.ntparallel.com. For information about GPU computing, see What is GPU Computing? from the NVIDIA Web site.