- Adds bgemv T based off of sbgemv T kernel
- Adds bgemv N which is slightly alterated to not use Y as an
accumulator due to the output being bf16 which results in loss of
precision
- Enables BGEMM_GEMV_FORWARD to proxy BGEMM to BGEMV with new kernels

This also improves the testing and generic kernel by re-using the BF16
conversion functions.

Built on top of https://github.com/OpenMathLib/OpenBLAS/pull/5357 and derived from https://github.com/OpenMathLib/OpenBLAS/pull/5287

Co-authored-by: Ye Tao <ye.tao@arm.com>

No interface named {c/z}a{min/max}, keeping it would
cause ambiguity

to fix linking problems with (at least) the AMD version of flang that creates dependencies on more than just the fortran runtime.

This patch is to enable benchtest for half precision gemm
when BUILD_HALF is set during make.

[OpenBLAS]: benchmark for her/her2 LEVEL2 functions

[OpenBlas]: Add benchmark tpsv file and modify benchmark/Makefile

[description]: benchmark for her/her2
[solution]: added benchmark for her/her2, modified makefile in benchmark
[dts]:

Add benchmark for SPMV and fix segmentation fault when data size >= 50000

[Description]: Solve lack of tpsv benchmark.

[Description]:Solve the problem of missing tpmv.c benchmark file

add benchmark for axpby test

modified:   benchmark/Makefile
	modified:   benchmark/rot.c

benchmark/hpmv&hbmv: add benchmark/hpmv.c and benchmark/hbmv.c

Add benchmark file rotm.c and modify benchmark/Makefile to test s/drotm

add benchmark for trsv

Add syr2 benchmark

Signed-off-by: Xuqiang Chen chenxuqiang3@hisilicon.com

modified:   benchmark/Makefile
	new file:   benchmark/rotm.c

2.Fixed some errors

Also revert 0bbca5e

* Bench small matrices with multi-threading
* Close #727