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

[OpenBLAS]:fix the iamax benchmark error

Move declarations of lapack_complex_custom types outside the extern C

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

[OpenBlas]:Add benchmark tpmv.c and modify benchmark/Makefile

Prevent attempts to run ctest or test when fortran is not available

fixes #2510

The main Makefile takes care of this in the build process, but users or CI jobs may try to run this directly

[Description]:the result for i?amax is not MFlops, it is MBytes

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

Apply fix for LAPACK issue 394 (fixed-form code beyond column 72)

Fix INTERFACE64 not propagating to the fortran codes on ARMV8

Fix  pr #2487 error

reference to XERBLA extending beyond column 72, breaking builds with compilers that default to traditional punch card format

rebase

Add benchmark for ?amax, ?max, ?amin, ?min, i?max, i?amin and i?min.

update

add benchmark for axpby test

add benchmark for csrot and zdrot

Remove redundant code

Fix use of make vs $(MAKE) in building lapack-testing

Modify the main Makefile in OpenBLAS

update

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

[Description]: check the compiler version and show the detail info

Broke lapack-testing on BSD as their default "make" does not support GNU Makefile syntax

rebase

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 spr/spr2

add benchmark for trsv

Add syr2 benchmark

Use acq/rel semantics to pass flags/pointers in getrf_parallel.

The current implementation has locks, but the locks each only
have a critical section of one variable so atomic reads/writes
with barriers can be used to achieve the same behavior.

Like the previous patch, pthread_mutex_lock isn't fair, so in a
tight loop the previous thread that has the lock can keep it
starving another thread, even if that thread is about to write
the data that will stop the current thread from spinning.

On a 64c Arm system this improves performance by 20x on sgesv.goto.

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