| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
|
This is our mail of April 17 with an ttachment. -------- Original-Nachricht --------
With the attachment this time! -------- Original-Nachricht --------
Dear colleagues: The text of paragraph 11.11.11 Complete arithmetic, dot product function on page 48 of Draft 7.1 is not yet in full accordance with the intention of the motion. The following changes are requested: (see our mail of October 11, 2012) 1. Change the third line after 11.11.11: "[4], for the parent format F of at least one such type.", and add another paragraph after it: [4], for the parent format F of these types. An implementation shall at least provide complete formats corresponding to binary64 if it provides binary floating-point arithmetic and decimal64 if it provides decimal floating-point. Complete formats corresponding to other floating-point formats may be provided. 2. In the penultimate paragraph, replace "exactly and rounds it once to give a result of format F" to exactly, giving a complete format result. The result may be converted if necessary to a specified result format by application of the convert operation. We attach a simplified version of the position paper. It is entitled: The Exact Dot Product. For more details on the topic see [5], [6], and [7] there. Best wishes Ulrich Kulisch and Van Snyder -- Karlsruher Institut für Technologie (KIT) Institut für Angewandte und Numerische Mathematik D-76128 Karlsruhe, Germany Prof. Ulrich Kulisch Telefon: +49 721 608-42680 Fax: +49 721 608-46679 E-Mail: ulrich.kulisch@xxxxxxx www.kit.edu www.math.kit.edu/ianm2/~kulisch/ KIT - Universität des Landes Baden-Württemberg und nationales Großforschungszentrum in der Helmholtz-Gesellschaft |
Attachment:
DotProdP1788.pdf
Description: Adobe PDF document