Optimized daxpy2v implementation

- Optimized axpy2v implementation for double
  datatype by handling rows in mulitple of 4
  and store the final computed result at the
  end of computation, preventing unnecessary
  stores for improving the performance.

- Optimal and reuse of vector registers for
  faster computation.

AMD-Internal: [CPUPL-1973]
Change-Id: I7b8ef94d0f67c1c666fdce26e9b2b7291365d2e9

kernels/zen/1f/CMakeLists.txt

@@ -8,5 +8,8 @@ target_sources("${PROJECT_NAME}"
     ${CMAKE_CURRENT_SOURCE_DIR}/bli_axpyf_zen_int_4.c
     ${CMAKE_CURRENT_SOURCE_DIR}/bli_axpyf_zen_int_6.c
     ${CMAKE_CURRENT_SOURCE_DIR}/bli_axpy2v_zen_int.c
+<<<<<<< HEAD
     ${CMAKE_CURRENT_SOURCE_DIR}/bli_dotxaxpyf_zen_int_8.c           
+=======
+>>>>>>> 8b5b2707... Optimized daxpy2v implementation
     )

kernels/zen/1f/bli_axpy2v_zen_int.c

@@ -186,6 +186,7 @@ void bli_daxpy2v_zen_int
 		);
 	}
 }
+<<<<<<< HEAD
 
 /**
  * zaxpy2v kernel performs axpy2v operation.
@@ -718,4 +719,6 @@ void bli_zaxpy2v_zen_int
         }
     }
     AOCL_DTL_TRACE_EXIT(AOCL_DTL_LEVEL_TRACE_4)
-}
+}
+=======
+>>>>>>> 8b5b2707... Optimized daxpy2v implementation

kernels/zen/bli_kernels_zen.h

@@ -32,6 +32,19 @@
    OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
 */
+<<<<<<< HEAD
+=======
+// hemv helper function
+void bli_pre_hemv_8x8(double *a, double *x,
+               double *y, double *alpha,
+               dim_t cs_a, dim_t rs_a);
+
+void bli_post_hemv_8x8(double *a, double *x,
+               double *y, double *alpha,
+               dim_t cs_a, dim_t rs_a);
+
+
+>>>>>>> 8b5b2707... Optimized daxpy2v implementation
 // -- level-1m --
 PACKM_KER_PROT(double, d, packm_8xk_gen_zen)
 PACKM_KER_PROT(double, d, packm_6xk_gen_zen)
@@ -129,7 +142,10 @@ AXPYF_KER_PROT( dcomplex, z, axpyf_zen_int_5 )
 AXPYF_KER_PROT( dcomplex, z, axpyf_zen_int_4 )
 // axpy2v (intrinsics)
 AXPY2V_KER_PROT(double, d, axpy2v_zen_int )
+<<<<<<< HEAD
 AXPY2V_KER_PROT(dcomplex, z, axpy2v_zen_int )
+=======
+>>>>>>> 8b5b2707... Optimized daxpy2v implementation
 
 // dotxf (intrinsics)
 DOTXF_KER_PROT( float,    s, dotxf_zen_int_8 )