Merge pull request #2 in PROC_LIBS/mathlib from CATREQ-2702 to master

* commit '3e54f23f82d27fb53301507eb5f66b7a4020fa64':
  Descoping c66 big endian
  CATREQ-2702: Descoping big-endian target

Descoping c66 big endian

CATREQ-2702: Descoping big-endian target

Update release notes.

Signed-off-by: Jianzhong Xu <xuj@ti.com>

Update MATHLIB benchmarking results based on C6000 8.3.2 compiler.

Signed-off-by: Jianzhong Xu <xuj@ti.com>

Ignore testing results for special case when input is NaN for acoshdp,
log10sp, log2sp, logsp.

Update release notes.

Signed-off-by: Jianzhong Xu <xuj@ti.com>

Renamed package information in tirex to be compatible with 3.1.2.1.

Signed-off-by: Jianzhong Xu <xuj@ti.com>

PRSDK-4876: fixed atan2dp overwriting unknown memory.

Signed-off-by: Jianzhong Xu <xuj@ti.com>

Changed package name for tirex. Separated include header files into groups.

Signed-off-by: Jianzhong Xu <xuj@ti.com>

PRSDK-4874: adding documentation in release notes on how to rebuild MATHLIB.

Signed-off-by: Jianzhong Xu <xuj@ti.com>

PRSDK-5011: fixing TIREX metadata typos.

Signed-off-by: Jianzhong Xu <xuj@ti.com>

Explicitly specify header files to be included in zip file.

Added readme.txt about how to build MATHLIB from git.ti.com source.

Updated release notes.

Added release notes.

MATHLIB: optimized floating-point math functions for TI DSPs.