arm-none-eabi-gcc : Printing float number using printf

Question

I am writing a C program for SAM3N arm cortex-M3 microcontroller. When I try to print float numbers, it just prints 'f'. Example: printf("%f",43.12); prints f only, not the 43.12.

But printing with integers works fine.

How to enable full printing of floats? I know that the compiler, by default, disabled float printing to reduce code size (i.e. seems, they linked cut-down version). Also please note that, there is no CFLAGS=-Dprintf=iprintf used in makefile.

Tools details:

ARM/GNU C Compiler : (crosstool-NG 1.13.1 - Atmel build: 13) 4.6.1
Above tool come with Atmel studio 6.0.

AFAIK this microcontroller is has no FPU. So you need to refer to compiler documentation to libc how to enable full printf capability. For example on AVR it was "-Wl,-u,vfprintf -lprintf_flt -lm". Another way, you should contact your tools vendor perhaps it is known issue. — Sergei Nikulov, Oct 03 '12 at 07:35
Unfortunately, this -lprintf works only for AVR-GCC, but NOT for ARM. — Prabhu U, Oct 03 '12 at 08:34

score 11 · Answer 1 · answered Sep 08 '14 at 12:12

11

Can you try, adding the below option in linker settings

-lc -lrdimon -u _printf_float

and it worked for me in ARM-CORTEXM0

answered Sep 08 '14 at 12:12

Shanmugasundaram Viswanathan

111
1
3

Works on an ARM Cortex-M4 as well. – Bence Kaulics Mar 07 '17 at 21:52
What is `-lrdimon` required for? It seems to cause undesired behavior in my application. – theorifice Sep 07 '18 at 18:55
`-lrdimon` is unnecessary and caused some strange behavior. I was never sure what the effect of this flag was ... – theorifice Dec 11 '18 at 23:45
Alternative data point: in my case the printf call crashed without the -lrdimon parameter. – KT. Jul 25 '20 at 00:00

score 4 · Answer 2 · edited May 12 '16 at 11:25

4

It can be so that your platform/libs does not support %f format specifier for printf/sprintf. As a first approach you can roll your own printf for floats/doubles:

void printDouble(double v, int decimalDigits)
{
  int i = 1;
  int intPart, fractPart;
  for (;decimalDigits!=0; i*=10, decimalDigits--);
  intPart = (int)v;
  fractPart = (int)((v-(double)(int)v)*i);
  if(fractPart < 0) fractPart *= -1;
  printf("%i.%i", intPart, fractPart);
}

edited May 12 '16 at 11:25

mrUnfused

45
1
7

answered Oct 03 '12 at 08:11

Agnius Vasiliauskas

10,935
5
50
70

Thanks for the idea, but I already implemented this as a work around (using modf, lround). And I wanted proper printf so that code will be cleaner. – Prabhu U Oct 03 '12 at 08:32

score 1 · Answer 3 · edited Feb 07 '23 at 17:54

The answer of @Agnius Vasiliauskas is partially correct; however, it doesn't consider the case when v is negative and greater than -1 (-1<v<0) in fact in that situation the intPart should have a minus sign ahead.

For example: if is v=-0.55 intPart will be 0, so you will print 0.55 and not -0.55! You have to add the correct sign manually, as shown below.

void printDouble(double v, int decimalDigits)
{
  int i = 1;
  int intPart, fractPart;
  for (;decimalDigits!=0; i*=10, decimalDigits--);
  intPart = (int)v;
  fractPart = (int)((v-(double)(int)v)*i);
  if(fractPart < 0) fractPart *= -1;

  if(v<0 && intPart=0) printf("-%i.%i", intPart, fractPart);
  else printf("%i.%i", intPart, fractPart);
}

score 0 · Answer 4 · answered Oct 03 '12 at 07:36

0

In compilers where floats are disabled by default, it is unusual that the default libraries can handle them. Look through your references and you will find information on how to recompile the appropriate libraries with float support, or (more likely) where to find a version that has already been built with it.

answered Oct 03 '12 at 07:36

Sniggerfardimungus

11,583
10
52
97

This is exactly where I am stuck. Seems, I need to replace libc by some other C library. Seems, default libc does not have printf that can printf floats. Any help/suggestion om how I can make it? – Prabhu U Oct 03 '12 at 08:44
It is the library that doesn't support printfing floats right - gcc shouldn't care about that? If you are looking for a different libc, try newlib, http://sourceware.org/newlib/ – auselen Oct 03 '12 at 11:33

arm-none-eabi-gcc : Printing float number using printf

4 Answers4

Linked