Add quadprecision math library

[shibatch]

This is a part of implementation of issue #233.
At this point, add, mul, div and sqrt are implemented. Remaining functions will be committed in the succeeding PRs.
As for vector extensions, SSE2, AVX, FMA4, AVX2, AV2_128, AVX512F and AdvSIMD are supported.

This quadprecision math library is built only if -DBUILD_QUAD option is given to cmake. For some time(1 year?), this subproject is positioned at alpha development stage.

[shibatch]

I forgot to mention that "-march=armv8-a+simd" option is removed as pointed out in issue #232.
The work around for ICE of armclang is also removed.

[fpetrogalli]

Is there any reason why the quad precision functions and tests live in a separate folder and not in src/libm?

[shibatch]

Is there any reason why the quad precision functions and tests live in a separate folder and not in src/libm?

There is so much difference in code maturity and public demand.
I think there is smaller number of people who are interested in the qp library.
I am going to add many qp functions in coming years.
If we keep qp functions in the same folder as libm, we will have much more hustle.

[fpetrogalli]

This look good to me now, I am happy to keep it separate from libsleef/libsleefgnuabi. Please resolve the conflicts and squash ll commits in one commit, with a sensible commit message of what you have added in this PR (please avoid a merge as it would introduce commits with empty message).