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.

Cholesky Decomposition on C6472?

Has anyone ever implemented Cholesky Decomposition on the C6472?  I wasn't sure if there were any libraries that would improve the efficiency of this or if just straight C coding implementation would be the way to go.  Thanks.

 

  • Evan Haas,

    I am certain that you searched TI.com and these forums for Cholesky and did not find what you were looking for. The TI Wiki Pages is also a good place to search, and you would not have found anything there, either.

    So my assessment is that TI does not directly offer anything for you, although in the calculations you could take advantage of the DSPLIB functions that might be used in the whole Cholesky solution.

    Your best choice will be to search with your favorite internet search engine for anything about Cholesky and C6000. This has been around for many years and has been addressed by textbooks and third parties for many of our processors. If it has not yet been adapted to the C6472, then you might have a fun project to take on, perhaps migrating a solution from another C6000 family DSP to the C6472.

    Were you looking to parallel-ize the operations to use all 6 cores in parallel, or did you just want to run the solution on a single core while the others are doing other work?

    Regards,
    RandyP

  • We have a pure C and C with intrinsic versions of Cholesky for the Key Stone (C66XX) family.  Obviously the C version can be used on the C6472, but the optimized version may contain intrinsic  that are not supported by the 6472.

    I have not put it on the Wiki yet, so if you are interested, reply to this posing with your email address

     

    Ran

  • Is there a way to PM you on this forum?  I'd rather not post my email up here if possible...