RE: [avr-gcc-list] Too much RAM used: __clz_tab[] linked

[Weddington, Eric]

> -----Original Message-----
> From: 
> address@hidden 
> [mailto:address@hidden
> org] On Behalf Of Thomas D. Dean
> Sent: Saturday, June 07, 2008 7:50 PM
> To: address@hidden
> Subject: [avr-gcc-list] Too much RAM used: __clz_tab[] linked
> 
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=29524 has been 
> put off to 4.2.5
> 
> # avr-gcc --version
> avr-gcc (GCC) 4.2.2
> ...
> 
> I need to use int to float conversion with an ATmega16.
> 
> Comment 17 of the bug report gives a work around, I have not tested.
> 
> However, my project consists of many files, all compiled with a -c
> option.  
> 
> avr-gcc  -mmcu=atmega16  -Wall -Wmissing-prototypes  -Os \
>      -I. -I../../../../asus-avr-1.0/include   -c -o xxx.o 
> ../common/xxx.c
> 
> several of these files use int to float conversions.
> 
> The final link
> 
> avr-gcc main.o adc.o audio.o behaviors.o bin_ascii.o 
> boundary.o cli.o cruise.o datalog.o \
>  display.o distance.o encoder.o escape.o global_var.o 
> health.o location.o motor.o navigate.o \
>  odometer.o passive.o prowl.o range.o servo.o sonar.o spi.o 
> steps.o timer.o tilt.o trace.o xlate.o \
>  -mmcu=atmega16 -L../../lib -lavrx2.6 -lm -o main.elf
> 
> has no effect in the placement of the -lm.
> 
> Is there another work around?

Could you send me your map file and your Makefile? If you're ok with it,
you could just send me your whole project privately.

Eric Weddington
Atmel