• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 53 subscribers
  • Views 2119 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Cast from float to int (TMS320C24xx ANSI C Compiler, Version 7.04)

Johannes von Löwis
Expert 1786 points
Other Parts Discussed in Thread: TMS320LF2407A

For the following code

  float x = -32768.0f;
  int   x_i   = (int) x;
  long  x_l_1 = (int) x;
  long  x_l_2 = (long) x;
  long  x_l_3 = (int) -32768.0f;

compiled with TMS320C24xx ANSI C Compiler, Version 7.04 I get on a TMS320LF2407A the following results (in the watch window from CCS3.3):

x:        -32768.0  
x_l_3:    -32768
x_i:      -29186  
x_l_1:    167422
x_l_2:    -32768 

The values of x_i and x_l_1 are unexpected. Can someone please comment on this?

Regards

Johannes

  • 0
    TI__Guru* 84225 points

    That is without a doubt a bug.  Please run this program to check whether it is the compiler or the watch window that is at fault (my bet is the compiler):

    #include <stdio.h>

int main()
{
    float x = -32768.0f;
    int   x_i   = (int) x;
    long  x_l_1 = (int) x;
    long  x_l_2 = (long) x;
    long  x_l_3 = (int) -32768.0f;

    printf("%f\n", x);
    printf("%d\n", x_i);
    printf("%ld\n", x_l_1);
    printf("%ld\n", x_l_2);
    printf("%ld\n", x_l_3);

    return 0;
}
  • 0 in reply to Archaeologist
    Expert 1786 points

    The output of this program in Stdout tab is consistent with the contents of the watch window.

    Regards, Johannes

  • 0 in reply to Johannes von Löwis
    TI__Guru* 84225 points

    It seems pretty likely that there is a bug in the library helper function which converts from float to int (F$$FTOI).  However, I am not at all familiar with C24x assembly code, so I can't analyze this any further; I can't even verify that the problem is actually in F$$FTOI rather than the compiler.  There is no more recent version of the C24x compiler available, and no updates will be made.  It should be possible to extract F$$FTOI from the library and make whatever fixes are necessary, and add that directly to your project, but it would probably be easier just to make sure you never attempt to cast from float to int; cast to long instead.  I'm sorry I cannot be of further help.

  • 0 in reply to Archaeologist
    Prodigy 60 points

    Hi,

    My code is:

    void AdcShow(char c, int value)
    {
    float temp1, temp2;
    uint32_t i, count;
    temp1 = (float)(value-2048);
    temp1 /= 410;
    intToString(temp1,sendbuf);
    mibspiSendData(c);
    wait(0x400);
    mibspiSendData(' ');
    wait(0x400);
    if(value < 0)
    count = 5;
    else
    count = 4;
    for(i = count; i > 0; i--)
    {
    mibspiSendData(sendbuf[i-1]);
    wait(0x400);

    }
    mibspiSendData('g');
    wait(0x400);

    }

    In debug mode,when excuting  temp1 = (float)(value-2048);   it goes wrong.  It come into sys_intvecs.asm.      Can you give me a hand? Thank you!

  • 0 in reply to Dem Devlin
    TI__Guru* 84225 points

    Are you using C24xx?