Timeline for Is the behaviour of floating-point arithmetic defined by the c++ standard?
Current License: CC BY-SA 3.0
15 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 4, 2015 at 22:31 | history | edited | user53141 | CC BY-SA 3.0 | Grammar and a typo |
| Dec 4, 2015 at 18:48 | answer | added | Bart van Ingen Schenau | timeline score: 2 | |
| Apr 27, 2015 at 11:29 | audit | Suggested edits | |||
| Apr 27, 2015 at 13:03 | |||||
| Apr 24, 2015 at 6:15 | comment | added | rwong | I think the drafts on this page (en.cppreference.com) might provide some insights for this question. | |
| Apr 24, 2015 at 5:03 | comment | added | Thomas Eding | No, but you can query and static_assert on some models stackoverflow.com/questions/5777484/… | |
| Apr 24, 2015 at 4:30 | comment | added | Basile Starynkevitch | It practically means that to understand the semantics of a C++ program dealing with floating point, you need to state that the program is conformant to C++11 and to IEEE754... (practically the case on most C++ implementations). Then I believe that its behavior can be defined (but I am not sure it would be unique). | |
| Apr 24, 2015 at 0:51 | comment | added | quant | Does this mean that a c++ program with floating point arithmetic cannot be defined by the standard? | |
| Apr 24, 2015 at 0:16 | comment | added | Basile Starynkevitch | But the practical thing to know is floating-point-gui.de | |
| Apr 24, 2015 at 0:12 | comment | added | Basile Starynkevitch | Indeed, C++ standards are costly, but their latest draft is nearly equivalent to them, and you can download them freely, e.g. for C++11 download n3337 | |
| Apr 24, 2015 at 0:09 | comment | added | Basile Starynkevitch | I would believe that a weird implementation where the floating point types are all singleton types (i.e. you have only one floating point value, NAN) is conformant to the letter (but not the spirit) of the C++11 standard. | |
| Apr 23, 2015 at 23:07 | comment | added | Ixrec | @RobertHarvey I mean there are concrete limits to how irregular any particular operation can be. | |
| Apr 23, 2015 at 23:05 | comment | added | Robert Harvey | @ixrec: By deterministic, do you mean consistently irregular? | |
| Apr 23, 2015 at 22:26 | comment | added | Ixrec | I believe SO has an excellent answer to this already: stackoverflow.com/questions/24157094/…. The short version is that the C++ standard says almost nothing, but most real-world implementations follow the IEEE 754 standard which is close to being fully deterministic. | |
| Apr 23, 2015 at 22:20 | history | edited | quant | CC BY-SA 3.0 | deleted 2 characters in body |
| Apr 23, 2015 at 22:12 | history | asked | quant | CC BY-SA 3.0 |