• State TI Thinks Resolved
  • Locked Locked
  • Replies 1 reply
  • Answers 1 answer
  • Subscribers 82 subscribers
  • Views 593 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.

Compiler/TMDSHVMTRPFCKIT: comparison between TMS320F28035's CLA unit and TMS320F28335 for execution speed and complexity of coding

ali karimi
Intellectual 360 points

Part Number: TMDSHVMTRPFCKIT

Tool/software: TI C/C++ Compiler

Hi Motor Drive Group,


would you please compare TMS320F28035's CLA unit with TMS320F28335 for execution speed and complexity of coding.
 
I want to implement an extra complicated math algorithm in CLA (besides FOC Sensorless in CPU). this algorithm needs high precise floating point and matrix calculations.

i am C language user. i want to make a decision between speed&code complexity&price of these two families.

Does CLA unit execute concurrent with main CUP ?(i have read "By using the CLA to service time-critical control loops, the main CPU
 is free to perform other system tasks.")

Best wishes,

Ali

  • 0
    TI__Mastermind 27200 points

    Ali,

    The simple answer is yes - the CLA does run concurrently with the main CPU, so of these two devices if you were to choose the F28035 you would be able to compute your math algorithm in parallel with whatever else you had the main CPU doing. However, there are details worth considering and your final choice will likely be compromise between execution time and cost.

    In terms of pure CPU cycle efficiency the F28335 would execute your algorithm more efficiently. It has a larger register set and as you know, runs at almost twice the speed (150 MHz vs. 80 MHz). You would also have much more memory available, and since the device is inherently floating point, you won't need to concern yourself about mixing fixed and floating point data types in your code. All these are important considerations if you have a complex algorithm and execution time is critical.

    My view is the F28335 would be a better choice from a technical perspective, however you will have to factor in the higher price of that device. For even better performance you might also like to consider the F28379D - this comprises two floating point CPUs, each with a CLA, and runs at 200 MHz; although again this performance is reflected in the price.

    Regards,

    Richard