Changed the type of variable *r* from int to size_t.
This change makes sure that a valid result of std::max(r + delta,
(r*3+1)/2) can be passed into the reserve function.

Squashed : ReLU operand order tests.

* 1. changed static const __m128/256 to const __m128/256 to avoid wierd instructions and calls inserted by compiler.
2. added universal intrinsics that wrap MOVNTPS and other such (non-temporary or "no cache" store) instructions. v_store_interleave() and v_store() got respective flags/overloaded variants
3. rewrote split & merge to use the "no cache" store instructions. It resulted in dramatic performance improvement when processing big arrays

* hopefully, fixed some test failures where 4-channel v_store_interleave() is used

* added missing implementation of the new universal intrinsics (v_store_aligned_nocache() etc.)

* fixed silly typo in the new intrinsics in intrin_vsx.hpp

* still trying to fix VSX compiler errors

* still trying to fix VSX compiler errors

* still trying to fix VSX compiler errors

* still trying to fix VSX compiler errors

Enable Torch layers tests with Intel's Inference Engine backend.

but execute tested functions in Layer_Test_Halide/Convolution.Accuracy

instead of 'int'

* fixed/updated v_load_deinterleave and v_store_interleave intrinsics; modified split() and merge() functions to use those intrinsics

* fixed a few compile errors and bug in v_load_deinterleave(ptr, v_uint32x4& a, v_uint32x4& b)

* fixed few more compile errors

improve error checks