Fixes for AVXVNNI instruction set - Clang Compiler

[Srihari-mcw]

The PR is to fix issues with the AVX_VNNI instruction set with the clang compiler. The updates were built across compilers and was seen to be building fine post the changes

Error seen :

The performance was tested before and after changes and were found to be similar - Tested with Linux GCC 12.3

model	size	params	backend	threads	test	t/s	speedup	Commit id
llama 7B Q4_0	3.56 GiB	6.74 B	CPU	14	pp 512	52.69 ± 0.15	0.02%	2a4e792
llama 7B Q4_0	3.56 GiB	6.74 B	CPU	14	pp 512	52.68 ± 0.19		7909e858
llama 7B Q4_0	3.56 GiB	6.74 B	CPU	14	tg 128	19.91 ± 0.25	0.55%	2a4e792
llama 7B Q4_0	3.56 GiB	6.74 B	CPU	14	tg 128	19.80 ± 0.46		7909e858

The perplexity was tested for 32 chunks and was found to be the same for Q4_0 model before and after changes - 5.4993 +- 0.13676

Flags enabled - Flags:| CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX_VNNI = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | LLAMAFILE = 1 | OPENMP = 1 | AARCH64_REPACK = 1 |

Model - Meta LLama2 7B - https://huggingface.co/meta-llama/Llama-2-7b

[slaren]

Thanks, this also fixes AVX VNNI with MSVC, so I have enabled it for MSVC as well.