MPC8xx fpu support and compiling glib.

MPC8xx fpu support and compiling glib.

[Conn Clark]

Hello All,


	I was wondering a few things about MPC8xx FPU support. If
I build a kernel with fpu-emulation do I have to pass the CFLAGS
"-msoft-float -D_SOFT_FLOAT" when compiling glibc? Does it save me
any ram if I don't? I know if I don't my performance will drop but
our aplication doesn't use floating point except for ping mabey.


Thanks in advance,

	Conn
--

*****************************************************************
  If you live at home long enough, your parents will move out.
*****************************************************************

Conn Clark
Engineering Stooge				clark@esteem.com
Electronic Systems Technology Inc.		www.esteem.com

Stock Ticker Symbol				ELST

"clark@esteem.com" Copyright 2000 By Electronic Systems Technology
 This email address may be used to communicate to Conn Clark
 provided it is not being used for advertisement purposes, unless
 prior written consent is given. This email address may not be
 sold under any circumstances. All other rights reserved.

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: MPC8xx fpu support and compiling glib.

[Wolfgang Denk]

In message <3CB60E13.BA0A4A6@esteem.com> you wrote:
>
> 	I was wondering a few things about MPC8xx FPU support. If
> I build a kernel with fpu-emulation do I have to pass the CFLAGS

That makes only sense when your application AND libraies DO  NOT  use
soft-float.

> "-msoft-float -D_SOFT_FLOAT" when compiling glibc? Does it save me

You cannot mix FPU-code and  soft-float  in  one  application/library
setup.

> any ram if I don't? I know if I don't my performance will drop but
> our aplication doesn't use floating point except for ping mabey.

Ummm... why should soft-float result in _smaller_ code? On contrary -
the code will be bigger, but run faster since you don;t have to  trap
into the kernel for each FP instruction.

Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
To know how another being, another creature feels -  that  is  impos-
sible.                  - Terry Pratchett, _The Dark Side of the Sun_

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: MPC8xx fpu support and compiling glib.

[Wolfgang Denk]

In message <3CB71524.46DC9F87@esteem.com> you wrote:
>
> > You cannot mix FPU-code and  soft-float  in  one  application/library
> > setup.
...
> If my kernel has fpu-emulation, do I need to compile soft-float into my
> glibc?

You can compile either ALL libraries and ALL  applications  with,  or
ALL without FP.

> Does putting fpu-emulation in the kernel and not using soft-float in
> glibc
> save me RAM?

Not any significant amount, I think.

Also be awre the the  emulation  code  in  the  kernel  is  far  from
complete or perfect.

Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
The IQ of the group is the lowest IQ of a member of the group divided
by the number of people in the group.

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: MPC8xx fpu support and compiling glib.

[Dr. Craig Hollabaugh]

At 09:47 PM 4/12/2002 +0200, Wolfgang Denk wrote:
>Also be awre the the  emulation  code  in  the  kernel  is  far  from
>complete or perfect.

Does this mean you recommend no emulation code in the kernel and ALL libraries and ALL applications be compiled with -msoft-float?


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: MPC8xx fpu support and compiling glib.

[Wolfgang Denk]

In message <3.0.3.32.20020412141129.0070c384@spudcentral.com> you wrote:
>
> At 09:47 PM 4/12/2002 +0200, Wolfgang Denk wrote:
> >Also be awre the the  emulation  code  in  the  kernel  is  far  from
> >complete or perfect.
>
> Does this mean you recommend no emulation code in the kernel and ALL libraries and ALL applications be compiled with -msoft-float?

For FPU-less systems: yes.

Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
Every little picofarad has a nanohenry all its own.      - Don Vonada

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/