• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 81 subscribers
  • Views 1644 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.

TMS 320F28335 IQ format Vs Floating point Unit

Narasimha Yalamala
Prodigy 185 points
Other Parts Discussed in Thread: TMS320F28335

Hi,

I am using TMS320F28335 processor which has 32 bit Floating point CPU. I have the requirement to have minimum precision to be set. I have the following doubts.

  1. Do we still need to use IQ format for handling floating point arithmetic? if so, what are the scenarios that IQ format should be used ? what are the advantages of using IQ format in floating point processor?
  2. Is there an option to set the precision for the floating point operations/variables ?
  3. what is the default precision value for TMS320F28335  FPU?

Thank in advance.

Narasimha

  • 0
    TI__Mastermind 20850 points

    Hi Narasimha,


    1. If you are using the FPU, there is no need to use IQMath. I dont see any advantage to using IQMath if you a floating-point engine available. The FPU uses 32-bit single precision floating point numbers.

    2. No, its single precision only

    3.Single precision float as per IEEE 754. You have 23 bits of precision in the mantissa.

    The following app note is a primer to the FPU on all our devices:

    http://focus.ti.com/lit/an/spraan9a/spraan9a.pdf

     

  • 0 in reply to Vishal_Coelho
    Prodigy 185 points

    Hi Vishal

    Thanks for your response.

    If i want to perform trigonometric functions(sin,cos), how can i handle these with out IQMath. May be i can do by adding math.h header, which i can not do it explicitly.

    Does "rts2800_fpu32.lib" also includes math.h to handle any arithematic.

    Narasimha

  • 0 in reply to Narasimha Yalamala
    TI__Mastermind 20850 points

    The rts2800_fpu32 library does support math.h. Be sure to turn on fpu32 support in the runtime build options of your project, CCS will automatically pull in the appropriate run time support library, in this case, rts2800_fpu32.lib.