Speeding Up Monte Carlo Computations by Parallel Processing Using a GPU for Uncertainty Evaluation in accordance with GUM Supplement 2

Abstract

The GUM Supplement 2 describes a Monte Carlo Method (MCM) for evaluating measurement models with more than one output quantity. Such models are common in electrical metrology where the measurands may be complex-valued quantities, such as S-parameters. The Standards and Calibration Laboratory (SCL) developed a software tool six years ago in accordance with GUM Supplement 2. The SCL software tool was written in Visual C++ and Visual Basic for Application (VBA), with Microsoft Excel as front-end user interface. As MCM involves large numbers of repetitive computations, this old software tool has a long processing time, especially for complicated measurement models. Nowadays many personal computers are equipped with a graphics processing unit (GPU) containing up to thousands of floating point cores. As MCM is well suited to parallel processing, SCL has ported the algorithm to a GPU using the Open Computing Language (OpenCL). The new tool is an add-on module to Microsoft Excel. GPUs from the major suppliers Nvidia, AMD and Intel are supported. The uncertainty computation time may be reduced significantly. This paper describes the design and implementation of this new software tool.