| |
 |
|
|
Science Forum Index » Math - Numerical Analysis Forum » Cross-platform way to set FP control bits?...
Page 1 of 1
|
| Author |
Message |
| Phil Hobbs... |
 Posted: Tue May 13, 2008 2:11 pm |
|
|
|
Guest
|
I have a couple of largish simulation programs that run on Windows and
Linux (and OS/2 because it has a debugger that I love). I need to be
able to ignore underflow errors, generate zero instead of NaN on an
underflow, and turn off denormalized numbers. I can do that on some
compilers and not others... is there a cross-platform solution to
setting reasonably consistent FP processor behaviour?
Thanks,
Phil Hobbs |
|
|
| Back to top |
|
| Phil Hobbs... |
| Posted: Mon May 19, 2008 1:23 pm |
|
|
|
Guest
|
Phil Hobbs wrote:
Quote: I have a couple of largish simulation programs that run on Windows and
Linux (and OS/2 because it has a debugger that I love). I need to be
able to ignore underflow errors, generate zero instead of NaN on an
underflow, and turn off denormalized numbers. I can do that on some
compilers and not others... is there a cross-platform solution to
setting reasonably consistent FP processor behaviour?
Thanks,
Phil Hobbs
Nobody? I'd settle for a good way to do it in gcc, which seems to be
the toughest case, if there really isn't a cross-platform library for
this...
Thanks,
Phil Hobbs |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Sun Jul 27, 2008 1:38 am
|
|
