• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 82 subscribers
  • Views 228 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.

TMS320F2800132: C2000 Power Efficient Array/DotProduct Multiplication..

Newman
Expert 6005 points
Part Number: TMS320F2800132
Other Parts Discussed in Thread: C2000WARE

My customer wants to do the following..

"Suppose I have an array of N values, say they’re 16 bit, and I wish to find the dot product to another array which can be either 8 or 16 bit representations.  Normally, I would need to perform N multiply accumulates (MACs).
The question is how can I perform this operation while expending the least amount of energy?  "

Any advice on how to do this efficiently w/ C2000?  Any related app notes?

Thanks, Merril

  • 0
    TI__Mastermind 26915 points

    Merril,

    If you look in C2000Ware, we have some assembly optimized dot product operations. I don't know if there's one that exactly fits what you're describing i.e. 16b int with 8/16b int, but there are quite a few mac examples on vectors.

    C:\ti\c2000\C2000Ware_5_01_00_00\libraries\dsp\FPU\c28\examples\vector

    As a key, 

    SP = single precision

    DP = double precision

    RV = real vector

    CV = complex vector

    RS = real scalar

    CS = complex scalar

    Thanks,

    Sira