At a customer of mine we had a graphic package which did the
presentation layer of the scientific and technical applications (called
GKS,
graphic kernel system). The output of GKS went to plotters (Calcomp)
and graphic displays (via GDDM). The interesting thing about this was
that you had not to code for GDDM or Calcomp; same interface for both.
The applications normally used double float for their computations,
but inside the graphic package, all floats were single precision, because
double would make no sense, given the precision of the output decives;
in fact you will see no difference when computing coordinates for the
plotter or GDDM. So on entry to the graphics package, the values were
truncated; all GKS functions had single precision arguments.
This is kind of CAD/CAM, too.
Kind regards
Bernd
Am 14.03.2017 um 02:15 schrieb Steve Beaver:
It’s a silly question, but why are you using single precession floating point.
The ONLY times I have used floating point is having to do with extremely large
numbers or some extremely small number of very fine precision
Used in CAD/CAM
Steve
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
