• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 53 subscribers
  • Views 752 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.

DINT for SPLOOP in C64+?

James Gort
Intellectual 590 points

Hi all-

We are using code generation tools v6.1.13, for a C64+ core (DM648), and are shocked to find that, though the compiled (and optimized for speed) assembly shows liberal use of SPLOOP, it takes the extra step of DINT before each SPLOOP. My understanding is that, in the C64+, SPLOOP *is* interruptable, so confused why DINT/RINT coming from compiler around the loops. Actually, more than confused, totally annoyed, as we have some large image loops that take a few millisecs to process, and we miss data from the UART!

Is 6.1.13 way out of date? (we are using CCS 3.3). We tried -mixxxxx, but it caused way too many problems. What I want is for the codegen tools to utilize what the processor is capable of--SPLOOPs that are interruptable!

Please Advise,

Jim Gort

  • 0
    TI__Expert 3455 points

    Jim,

     

    The compiler shouldn't be using DINT/RINT around loops that use the SPLOOP feature.  Please post a compilable testcase with options used to compile.  We can take a look.

     

    You may want to try a more recent patch version of 6.1.x, or even move to 7.2.x or 7.3.x.  6.1.13 is pretty old, the current patch release on that branch is 6.1.19.

    Download Code Generation Tools

     

    Thanks,

    -Todd

  • 0 in reply to Todd Hahn
    Intellectual 590 points

    Hi Todd-

    Thank you for your quick response.

    Yes, you are correct--the compiler is correctly not using DINT/RINT on the loops that it is able to use SPLOOP. It was the *other* loops, the manual software pipelined loops that, for various reasons, it could not use SPLOOP, that it was protecting with DINT/RINT. We are currently playing around with -mixxx to get compiler to optimize as much as possible, subject to the constraints of our maximum latency to interrupt services.

    Jim

  • 0 in reply to James Gort
    TI__Guru**** 244930 points

    This wiki article may also be useful.

    Thanks and regards,

    -George