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.

C66 IQMath Library

Other Parts Discussed in Thread: SYSBIOS, MATHLIB

Is there the IQMath Librar for the C66 DSP?

Here is the IQMath Library of current version "v2.1.4.0".

http://www.ti.com/tool/sprc542

Is this version applicable for the C66 DSP?
To apply it, must the source codes be used?

Our customer used "v2.1.3" for C64x+ DSP.
What are the differences between "v2.1.3" and "v2.1.4"?

Best regards,

Daisuke

 

  • Hi Daisuke,

    The update for IQMATH 2.1.4 from 2.1.3 are as mentioned in the release notes :

    -Bug in inlined functions that are part of the source package is resolved
    -Global definition of some variables changed to local const to prevent redefinition errors reported in user code with similar variables.
    -Linux installer created
    -Included TI TSPA license
    -ELF version of IQMATH library included

    IQMATH is a fixed point library so it can be used as is on the C66x DSP as the new DSP core maintains binary compatability with previous generation DSP cores. The ELF version of the library allows users to use it with MCSDK software and SYSBIOS which supports ELF binary format as the default binary format.

    Regards,

    Rahul

  • Daisuke,

    The IQMATH library was developed for fixed point processors as a way to easily do fixed-point representations of floating point numbers for math operations. The C66x core has built-in floating point math support, so you may get better performance by using the native floating point instead of the IQMATH library's version of those. Please take a look at the MATHLIB functions and DSPLIB functions for the C66x as an alternative.

    Regards,
    RandyP

  • Hi Rahul and RandyP,

    Thank you for your reply.

    I just understood. The IQMath Library is called A Virtual Floating Point Engine.

    Best regards,

    Daisuke

     

  • I have been using an old IQmath library version where I build the source directly. I'm now trying to transition to the pre-built library with version 2.1.4, however, it seems like the big-endian ELF build (IQmath_c64x+e_elf.lib) is actually built in little-endian mode. My big-endian unit test CCS project fails to link to this library file. Can someone please confirm this and provide a fix? I am fine linking to 2.1.3, if that doesn't have this bug.

    Thanks,
    Manu