Errors with expsp() and MATHLIB on C674x

William List

Other Parts Discussed in Thread: MATHLIB

Hi, I have a customer that is using mathlib c674x 3_1_1_0 with C674x and cannot get any correct output for expsp(). The EVM he is using is the TMDX0SKL138BET (Spectrum Digital).

Some examples provided below of the output he is seeing:

This code:

printf ("expsp(%f) = %f\n", -1.0f, expsp (-1.0f));

printf ("expsp(%f) = %f\n", -0.5f, expsp (-0.5f));

printf ("expsp(%f) = %f\n", 0.0f, expsp (0.0f));

printf ("expsp(%f) = %f\n", 0.5f, expsp (0.5f));

printf ("expsp(%f) = %f\n", 1.0f, expsp (1.0f));

printf ("expsp_c(%f) = %f\n", -1.0f, expsp_c (-1.0f));

printf ("expsp_c(%f) = %f\n", -0.5f, expsp_c (-0.5f));

printf ("expsp_c(%f) = %f\n", 0.0f, expsp_c (0.0f));

printf ("expsp_c(%f) = %f\n", 0.5f, expsp_c (0.5f));

printf ("expsp_c(%f) = %f\n", 1.0f, expsp_c (1.0f));

Gives these answers:

expsp(-1.000000) = -inf

expsp(-0.500000) = -inf

expsp(0.000000) = 1.000000

expsp(0.500000) = 0.000000

expsp(1.000000) = 0.000000

expsp_c(-1.000000) = 0.367879

expsp_c(-0.500000) = 0.606531

expsp_c(0.000000) = 1.000000

expsp_c(0.500000) = 1.648721

expsp_c(1.000000) = 2.718282

Anyone seen anything similar using the assembly functions? Could it be a setup issue? Thanks in advance.

Bill

over 9 years ago

0 ToddMullanix over 9 years ago

TI__Guru* 96960 points

I moved this thread to the device forum.

Todd

0 Shankari G over 9 years ago

TI__Mastermind 43955 points

HI Bill,

I will have a look at it and get back to you.

Thanks for your patience.

0 Shankari G over 9 years ago

TI__Mastermind 43955 points

HI Bill,

I have just experimented the functions, "expsp" and the "expsp_c" of mathlib c674x 3_1_1_0 with C674x

and got the results. Please compare it with yours.

0 ran35366 over 9 years ago in reply to Shankari G

TI__Genius 12805 points

To add to Shankari

I did the same (in parallel with Shankari) and got the same results

Here is the code and the results:

{
      float x_in, y_ass, y_c   ;

      x_in = -1.0   ;
      y_ass = expsp(x_in) ;
      y_c = expsp_c(x_in) ;
      printf("x_in = %f y_ass = %f y_c = %f \n", x_in, y_ass, y_c) ;

      x_in = -0.5   ;
      y_ass = expsp(x_in) ;
      y_c = expsp_c(x_in) ;
      printf("x_in = %f y_ass = %f y_c = %f \n", x_in, y_ass, y_c) ;

      x_in = 0.0   ;
      y_ass = expsp(x_in) ;
      y_c = expsp_c(x_in) ;
      printf("x_in = %f y_ass = %f y_c = %f \n", x_in, y_ass, y_c) ;

      x_in = 0.5   ;
      y_ass = expsp(x_in) ;
      y_c = expsp_c(x_in) ;
      printf("x_in = %f y_ass = %f y_c = %f \n", x_in, y_ass, y_c) ;

      x_in = 1.0   ;
      y_ass = expsp(x_in) ;
      y_c = expsp_c(x_in) ;
      printf("x_in = %f y_ass = %f y_c = %f \n", x_in, y_ass, y_c) ;

      printf ("expsp(%f) = %f\n", -1.0f, expsp (-1.0f));

        printf ("expsp(%f) = %f\n", -0.5f, expsp (-0.5f));

        printf ("expsp(%f) = %f\n", 0.0f, expsp (0.0f));

        printf ("expsp(%f) = %f\n", 0.5f, expsp (0.5f));

        printf ("expsp(%f) = %f\n", 1.0f, expsp (1.0f));

        printf ("expsp_c(%f) = %f\n", -1.0f, expsp_c (-1.0f));

        printf ("expsp_c(%f) = %f\n", -0.5f, expsp_c (-0.5f));

        printf ("expsp_c(%f) = %f\n", 0.0f, expsp_c (0.0f));

        printf ("expsp_c(%f) = %f\n", 0.5f, expsp_c (0.5f));

        printf ("expsp_c(%f) = %f\n", 1.0f, expsp_c (1.0f));

}

And the results are correct

x_in = -0.500000 y_ass = 0.606531 y_c = 0.606531
x_in = 0.000000 y_ass = 1.000000 y_c = 1.000000
x_in = 0.500000 y_ass = 1.648721 y_c = 1.648721
x_in = 1.000000 y_ass = 2.718282 y_c = 2.718282
expsp(-1.000000) = 0.367879
expsp(-0.500000) = 0.606531
expsp(0.000000) = 1.000000
expsp(0.500000) = 1.648721
expsp(1.000000) = 2.718282
expsp_c(-1.000000) = 0.367879
expsp_c(-0.500000) = 0.606531
expsp_c(0.000000) = 1.000000
expsp_c(0.500000) = 1.648721
expsp_c(1.000000) = 2.718282

Ran