[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [rtl] More questions about floating point
Janet.
All of my experience with hands-on control of the FPU comes from a
pseudo-DOS environment but the FPU's the same so:
The Microsoft Assembler x86 Instruction Set book has a wonderful
section on the FPU. There are separately maskable interrupts (actually
interrupt causes -- there's only one interrupt) for:
* Loss of Precision
* Underflow
* Overflow
* Divide By Zero
* Denormalized Operand
* Invalid Operation.
Therefore the answer to your question is that what the system does depends
on how it's setup. The Intel mnemonic for storing the control word is FSTCW
but I don't know what "as" uses. If you need more specific information,
contact me off-list.
Fellow Avionics Support Systems Engineer & (SGI) Advanced Signal Processing
Laboratory Administrator
Radar Systems Engineering Department
Electronic Systems and Sensors
MS 520
Northrop Grumman Corporation
Baltimore-Washington International Airport
7323 Aviation Boulevard
Baltimore Maryland 21240
Voice: (410) 993 - 2096 Mornings; all-day voice-mail
(410) 969 - 8068 Afternoons with answering machine
FAX: (410) 993 - 8084 On-site
(410) 969 - 8068 Afternoons; call first to arrange
E-Mail: Mornings: mailto:Norman_A_Dresner@mail.Northgrum.com
Afternoons: mailto:ndrez@att.net
> -----Original Message-----
> From: Estabridis, Janet P [SMTP:EstabridisJP@navair.navy.mil]
> Sent: Tuesday, October 02, 2001 11:05 PM
> To: 'Norm Dresner '; 'rtl@fsmlabs.com '; Estabridis, Janet P
> Subject: RE: [rtl] More questions about floating point
>
> I'll be waiting. Maybe you can answer this for me. Does the system hang
> or crash if you do get an overflow on a divide ? I am experiencing some
> problems and I'm trying to track it down. I am not always sure my numbers
> will be well bounded when I do the operations (coordinate translation and
> rotation) due to the fact that I am receiving the data via a remote site.
>
>
> Janet
>
> -----Original Message-----
> From: Norm Dresner
> To: rtl@fsmlabs.com; EstabridisJP@navair.navy.mil
> Sent: 10/2/01 4:55 PM
> Subject: Re: [rtl] More questions about floating point
>
> I assume we're talking about x86 architecture. What you
> want to do is to read the status register of the FPU. IIRC
> there are FPU-instructions for doing just that.
> Unfortunately I don't know the linux assembler well enough
> to come up with a Q&D means of doing that but hopefully one
> of the gurus will jump in and save the day.
>
> Norm
>
> ----- Original Message -----
> From: Estabridis, Janet P <EstabridisJP@navair.navy.mil>
> To: <rtl@fsmlabs.com>
> Sent: Tuesday, October 02, 2001 4:49 PM
> Subject: [rtl] More questions about floating point
>
>
> > Hi,
> >
> > Can you check for a floating point overflow if you are
> using floating point
> > in a kernel thread ?
> >
> > Janet Estabridis
> > -- [rtl] ---
> > To unsubscribe:
> > echo "unsubscribe rtl" | mail majordomo@rtlinux.org OR
> > echo "unsubscribe rtl <Your_email>" | mail
> majordomo@rtlinux.org
> > --
> > For more information on Real-Time Linux see:
> > http://www.rtlinux.org/
> >
> -- [rtl] ---
> To unsubscribe:
> echo "unsubscribe rtl" | mail majordomo@rtlinux.org OR
> echo "unsubscribe rtl <Your_email>" | mail majordomo@rtlinux.org
> --
> For more information on Real-Time Linux see:
> http://www.rtlinux.org/