• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 98 subscribers
  • Views 857 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.

Performance of a C674x DSP core calculating Double-Precision floating point values

Daisuke Maeda
Guru 20215 points


Single-Precision floating point calculation performance as follows:

  - Floating point MACs/cycle: 2

  - Arithmetic Floating point operations/cycle: 6

Please let me know Double-Precision floating point calculation performance.

Best regards,

Daisuke

 

  • 0
    TI__Mastermind 20790 points

    Daisuke,

    Double precision floating point goes to 2 MACs/cycle.
    Two Multiplies and Four Arithmetic operations for Double-Precission Floating point.
    INT -> DP conversion operations (.L units)
    Conversion to/from integer and single-precision values can be done on both .L and .S units on the C674x.

  • 0 in reply to Pubesh
    Guru 20215 points

    Hi Pubesh,

    Thank you for your reply.
    Sorry for my late reply.

    I understand that C674x Double-Precision Floating point calculations performance as follows:

      - Floating point MACs/cycle: 2

      - Arithmetic Floating point operations/cycle: 6 (2 Multiplies and 4 Arithmetic functions, except Load/Stores and DP conversions)

    In addition, here is performance for C66x.

    http://e2e.ti.com/support/dsp/c6000_multi-core_dsps/f/639/t/281161.aspx

    Therefore, I understand that a C674x core is equivalent to a C66x core in Double-Precision Floating point calculations performance.

    Best regards,

    Daisuke