Twofold fast arithmetic
نویسنده
چکیده
Can we assure math computations by automatic verifying floating-point accuracy? We define fast arithmetic (based on Dekker [1]) over twofold approximations z ≈ z0 + z1, such that z0 is standard result and z1 assesses inaccuracy ∆z0 = z − z0. We propose on-fly tracking z1, detecting if ∆z0 appears too high. We believe permanent tracking is worth its cost. C++ test code for Intel AVX available via web.
منابع مشابه
Twofold exp and log
This article is about twofold arithmetic [1, 2]. Here I introduce algorithms and experimental code for twofold variant of C/C++ standard functions exp() and log(), and expm1() and log1p(). Twofold function y0 + y1 ≈ f(x0 + x1) is nearly 2x-precise so can assess accuracy of standard one. Performance allows assessing on-fly: twofold texp() over double is ~10x times faster than expq() by GNU quadm...
متن کاملA Family of Recursive Least-squares Adaptive Algorithms Suitable for Fixed-point Implementation
The main feature of the least-squares adaptive algorithms is their high convergence rate. Unfortunately, they encounter numerical problems in finite precision implementation and especially in fixed-point arithmetic. The objective of this paper is twofold. First, an analysis of the finite precision effects of the recursive least-squares (RLS) algorithm is performed, outlining some specific probl...
متن کاملBroth microdilution testing of Pseudomonas aeruginosa and aminoglycosides: need for employing dilutions differing by small arithmetic increments.
The use of dilutions differing by small arithmetic increments was studied as a means for improving the definition and measurement of minimum inhibitory concentrations and precision parameters for testing Pseudomonas aeruginosa versus the aminoglycosides by the broth microdilution test. For five strains of P. aeruginosa versus gentamicin, tobramycin, and amikacin, comparisons were made of minimu...
متن کاملTwofold fast summation
Debugging accumulation of floating-point errors is hard; ideally, computer should track it automatically. Here we consider twofold approximation of exact real with value + error pair of floating-point numbers. Normally, value + error sum is more accurate than value alone, so error can estimate deviation between value and its exact target. Fast summation algorithm, that provides twofold sum of ∑...
متن کاملDesign and Analysis of Fast Text Compression Based on Quasi-Arithmetic Coding
We give a detailed algorithm for fast text compression. Our algorithm, related to the PPM method, simpli es the modeling phase by eliminating the escape mechanism and speeds up coding by using a combination of quasi-arithmetic coding and Rice coding. We provide details of the use of quasi-arithmetic code tables, and analyze their compression performance. Our Fast PPM method is shown experimenta...
متن کامل