A precision- and range-independent tool for testing floating-point arithmetic I: Basic operations, square root, and remainder

Citation
B. Verdonk et al., A precision- and range-independent tool for testing floating-point arithmetic I: Basic operations, square root, and remainder, ACM T MATH, 27(1), 2001, pp. 92-118
Citations number
28
Language
INGLESE
art.tipo
Article
Categorie Soggetti
Computer Science & Engineering
Journal title
ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE
ISSN journal
0098-3500 → ACNP
Volume
27
Issue
1
Year of publication
2001
Pages
92 - 118
Database
ISI
SICI code
0098-3500(200103)27:1<92:APARTF>2.0.ZU;2-S
Abstract
This paper introduces a precision- and range-independent tool for testing t he compliance of hardware or software implementations of (multiprecision) f loating-point arithmetic with the principles of the IEEE standards 754, and 854. The tool consists of a driver program, offering many options to test only specific aspects of the IEEE standards, and a large set of test vector s, encoded in a precision-independent syntax to allow the testing of basic and extended hardware formats as well as multiprecision floating-point impl ementations. The suite of test vectors stems on one hand from the integrati on and fully precision- and range-independent generalization of existing ha rdware test sets, and on the other hand from the systematic testing of exac t rounding for all combinations of round and sticky bits that can occur. Th e former constitutes only 50% of the resulting test set. In the latter we e specially focus on hard-to-round cases. In addition, the test suite implici tly tests properties of floating-point operations, following the idea of Pa ranoia, and it reports which of the three IEEE-compliant underflow mechanis ms is used by the floating-point implementation under consideration. We als o check whether that underflow mechanism is used consistently. The tool is backward compatible with the UCBTEST package and with Coonen's test syntax.