The IEEE 754-2008 standard recommends the correct rounding of elementary functions. This requires to solve the Table Maker’s Dilemma which implies a huge amount of CPU computation time. We consider in this paper accelerating such computations, namely Lefevre algorithm, on Graphics Processing Units (GPU) which are massively parallel architectures with a partial SIMD execution (Single […]

June 2, 2013 by hgpu

The IEEE 754-2008 standard recommends the correct rounding of elementary functions. This requires to solve the Table Maker’s Dilemma which implies a huge amount of CPU computation time. We consider in this paper accelerating such computations, namely Lef’evre algorithm, on Graphics Processing Units (GPU) which are massively parallel architectures with a partial SIMD execution (Single […]

November 14, 2012 by hgpu

The goal of this project is to implement the GMP library in CUDA and evaluate its performance. GMP (GNU Multiple Precision) is a free library for arbitrary precision arithmetic, operating on signed integers, rational numbers, and floating point numbers. There is no practical limit to the precision except the ones implied by the available memory […]

February 7, 2012 by hgpu

Since 1985, the IEEE 754 standard defines formats, rounding modes and basic operations for floating-point arithmetic. In 2008 the standard has been extended, and recommendations have been added about the rounding of some elementary functions such as trigonometric functions (cosine, sine, tangent and their inverses), exponentials, and logarithms. However to guarantee the exact rounding of […]

November 23, 2011 by hgpu

Accurate table methods allow for very accurate and efficient evaluation of elementary functions. We present new single-table approaches to logarithm and exponential evaluation, by which we mean that a single table of values works for both log(x) and log(1+x), and a single table for ex and ex-1. This approach eliminates special cases normally required to […]

September 11, 2011 by hgpu

Data-parallel architectures like SIMD (Single Instruction Multiple Data) or SIMT (Single Instruction Multiple Thread) have been adopted in many recent CPU and GPU architectures. Although some SIMD and SIMT instruction sets include double-precision arithmetic and bitwise operations, there are no instructions dedicated to evaluating elementary functions like trigonometric functions in double precision. Thus, these functions […]

November 20, 2010 by hgpu