• State Resolved
  • Locked Locked
  • Replies 1 reply
  • Subscribers 98 subscribers
  • Views 329 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

C6748 FPU instructions

David Luca
Expert 2795 points
Other Parts Discussed in Thread: MATHLIB

Hello,

I was wondering since I didn't have time to check assembly instructions, does the CCSv5.2 with compiler v7.3.7 generates FPU instructions for C6748 FPU co-processor when we are using float type variables or is it emulated (or fixed-point)?  Why is then Mathlib provided to optimize some of the functions?

Best regards,

David.

  • 0
    TI__Guru** 116170 points

    David,

    David Luca said:
    does the CCSv5.2 with compiler v7.3.7 generates FPU instructions for C6748 FPU co-processor when we are using float type variables or is it emulated (or fixed-point)?

    The Compiler v7.3.7 will generate FPU instructions as long as you have specified the target platform as C674x and not as C6400.

    David Luca said:
    Why is then Mathlib provided to optimize some of the functions?

    The compiler comes with  RTS libraries for the C6740 target which includes math functions for computing single and double precision trignometric, arithmetic and log/exponential functions.The value of each of these functions are computed using expansion series.The default implementation of these functions in compiler RTS libraries use a larger set of terms from the expansion series of these functions to achieve greater precision and also additional control codes while evaluating the functions. The MATHLIB reduces the number of terms which have neglible impact in the computation and eliminates some of the control code to achieve higher performance with very little impact on the accuracy.

    Best Regards,

    Rahul