|
From: | Michael Godfrey |
Subject: | Re: floating point precision control |
Date: | Wed, 13 Jul 2016 19:40:15 -0400 |
User-agent: | Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.1.1 |
John, I think that this is right. Trying to use some capabilities beyond 64-bit IEEE may work in some cases, but would necessarily introduce "unexpected" and variable results which would depend in conditions which would be hard to control or maintain. Reliability and stability are very important. They dominate occasional improvements in precision. Michael On 07/12/2016 03:28 PM, John W. Eaton
wrote:
Maybe we should just try to do the best we can with 64-bit IEEE arithmetic and not try to use extended precision since it is impossible to always use 80-bit precision everywhere (intermediate values may be stored in 64-bit memory locations, for example) and a mixture of 64- and 80-bit arithmetic seems like it may be asking for unpredictable results? |
[Prev in Thread] | Current Thread | [Next in Thread] |