Improve Windows threading performance scaling

fix typo

Add 15 level1 optimizations for LoongArch.

Fix keyword used to count successful tests (Reference-LAPACK PR 954)

Fix memory leak in LAPACK testing framework (Reference-LAPACK PR 953)

Add LAPACK tests for the Dynamic Mode Decomposition functions from Reference-LAPACK PR 736

Use 64bit build on `CMAKE_SYSTEM_PROCESSOR=i386` on Darwin

Write xerbla.c.o into archive regardless of timestamp by using ar -rs
instead of ar -ru.

Here a bit tricky things.

A value `CMAKE_SYSTEM_PROCESSOR` is came from output of `uname -m` which
migth be 32bit with 64bit building applicaiton.

So, for that case use `CMAKE_SIZEOF_VOID_P` to detect the target.

See https://trac.macports.org/ticket/68488

Add some refines and optimizations for LoongArch.

Allow weak linking on old macOS

Fix CMAKE installation location of lapacke_mangling header

1. Use getauxval instead of cpucfg to test hardware capability.
2. Remove unnecessary code and option for compiler check in c_check.

Enable overstep of too long args without DYNAMIC_ARCH

Update version number and documentation of Reference-LAPACK to 3.12.0