CMake

*refactoring of SSE4.1 optimizations of SynetMergedConvolution32f. #647

Job	Run time
build_aarch64_cross (Debug)	3m 58s
build_and_test_clang (Debug)	15m 1s
build_aarch64_cross (Release)	4m 19s
build_and_test_gcc_new (13, Debug)	15m 51s
build_and_test_clang (Release)	16m 8s
build_and_test_gcc_old (10, Debug)	13m 57s
build_and_test_gcc_new (13, Release)	19m 31s
build_and_test_gcc_old (10, Release)	16m 2s
build_and_test_gcc_old (11, Debug)	14m 31s
build_and_test_mingw (Debug)	18m 46s
build_arm_cross (Debug)	4m 38s
build_and_test_gcc_old (11, Release)	16m 33s
build_and_test_mingw (Release)	18m 27s
build_arm_cross (Release)	3m 30s
	3h 1m 12s

Provide feedback