Because of the holidays, TI E2E™ design support forum responses will be delayed from Dec. 25 through Jan. 2. Thank you for your patience.

  • State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 31 subscribers
  • Views 862 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.

Cortex R4F - Floating point and integer math.

Manoj Raut
Prodigy 40 points

HI,

I want to know the functional difference between the way Cortex R4F treats the math calculations of the floating and int. numbers.

Also cortex R4F has FPU unit for floating math Does it mean it uses separate dedicated hardware for it?

What is the advantage & difference over the normal math calculation hardware? and does it affects the timing(CPU cycles)  for the calcuation.

Thank you in advance.

  • 0
    TI__Mastermind 23000 points

    Hello,

    The R4F includes a dedicated hardware floating point unit (FPU). The FPU has its own register bank and a dedicated execute pipeline in order to improve performance and minimize impact between the integer and floating point calculations.  Note that it is possible to issue instructions to the integer and FP pipe in parallel.

    The primary advantage to a hardware floating point unit is improved performance.  Using dedicated floating point hardware vs. integer emulation of floating point can result in calculations which complete 50-300% faster. 

     

    Regards,

    Karl

  • 0
    TI__Guru**** 192226 points

    Manoj,

    FPU is a separate hardware which provides hw sipport for floating point operations in single, and double precision floating point arithmetic. There are 2 ways to do floating point arithmetic:

    1. HW: Use float instructions of FPU. (fast)

    2. SW: Have your compiler translate floating point arithmetic to integer arithmetic. (slow)

    Please refer to Chapter 14 of Cortex-R4F TRM (Rev r1p3) for cycles used by FPU instructions.

    http://infocenter.arm.com/help/topic/com.arm.doc.ddi0363e/DDI0363E_cortexr4_r1p3_trm.pdf

    Regards,

    QJ