Subject:
|
Re: Intermittend IC floating point divide problem
|
Newsgroups:
|
lugnet.robotics.handyboard
|
Date:
|
Mon, 23 Nov 1998 20:24:47 GMT
|
Original-From:
|
Eric Noyau <NOYAU@APPLE.COMstopspammers>
|
Viewed:
|
6959 times
|
| |
|
|
Somebody already did part of the job. If you go to http://www.cs.dartmouth.edu/~jonh/robots/skguide.html you can download a pcode source for a road warrior kind of robot with some "corrections to the floating point math library to produce correct results for several previously buggy operations".
I haven't looked at it yet but it might be interesting to you or your programmer assigned to the task. Poor guy! I'll recommend a good read of the The Art Of Computer Programming; Fundamental Algorithms by Donald E. Knuth (Do not forget to get the errata from his Web site: there is a bug in his division algorithm).
-- Eric
At 09:31 11/23/98 , Randy Sargent wrote:
>
> Dear Richard and Javaid,
>
> I have assigned a programmer to track down the floating point divide
> problem, and will keep you posted on its resolution. We believe it is a
> problem with the floating point library that is assembled into the pcode,
> so it is likely we should be able to release a new version of the pcode
> with a fix to the problem.
>
> -- Randy
>
> > Thankyou for your kind response to my student Javaid Iqbal on errors in
> > using floating point divide within IC software. The errors have
> > effectively brought our use to a full stop at present because we cannot
> > rely on results nor bypass what has become a random occurance of the
> > error. Would you know of anybody who can fix the bug or a different
> > software environment to use?
> >
> > Kind regards
> >
> > Richard Willgoss
> > UNSW, Sydney, Australia
>
> ----------------------------------------------------------------
> Randy Sargent Newton Research Labs
> President Robotic Systems and Software
> rsargent@newtonlabs.com http://www.newtonlabs.com/
|
|
Message is in Reply To:
 | | Intermittend IC floating point divide problem
|
| Dear Richard and Javaid, I have assigned a programmer to track down the floating point divide problem, and will keep you posted on its resolution. We believe it is a problem with the floating point library that is assembled into the pcode, so it is (...) (28 years ago, 23-Nov-98, to lugnet.robotics.handyboard)
|
2 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
