diff --git a/kernel/power/KERNEL.POWER8 b/kernel/power/KERNEL.POWER8
index 0f4516812..99bb38096 100644
--- a/kernel/power/KERNEL.POWER8
+++ b/kernel/power/KERNEL.POWER8
@@ -130,8 +130,8 @@ ZDOTKERNEL   =  zdot.c
 #CNRM2KERNEL  = ../arm/znrm2.c
 #ZNRM2KERNEL  = ../arm/znrm2.c
 #
-#SROTKERNEL   = ../arm/rot.c
-#DROTKERNEL   = ../arm/rot.c
+SROTKERNEL   = srot.c
+DROTKERNEL   = drot.c
 #CROTKERNEL   = ../arm/zrot.c
 #ZROTKERNEL   = ../arm/zrot.c
 #
diff --git a/kernel/power/drot.c b/kernel/power/drot.c
new file mode 100644
index 000000000..c93f69b12
--- /dev/null
+++ b/kernel/power/drot.c
@@ -0,0 +1,167 @@
+/***************************************************************************
+Copyright (c) 2013-2016, The OpenBLAS Project
+All rights reserved.
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+1. Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+2. Redistributions in binary form must reproduce the above copyright
+notice, this list of conditions and the following disclaimer in
+the documentation and/or other materials provided with the
+distribution.
+3. Neither the name of the OpenBLAS project nor the names of
+its contributors may be used to endorse or promote products
+derived from this software without specific prior written permission.
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ARE DISCLAIMED. IN NO EVENT SHALL THE OPENBLAS PROJECT OR CONTRIBUTORS BE
+LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
+USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*****************************************************************************/
+
+/**************************************************************************************
+* 2016/03/27 Werner Saar (wernsaar@googlemail.com)
+* 	 BLASTEST 		: OK
+* 	 CTEST			: OK
+* 	 TEST			: OK
+*	 LAPACK-TEST		: OK
+**************************************************************************************/
+
+
+
+#include "common.h"
+
+#pragma GCC optimize "O1"
+
+#if defined(POWER8)
+#include "drot_microk_power8.c"
+#endif
+
+
+#ifndef HAVE_KERNEL_16
+
+static void drot_kernel_16(BLASLONG n, FLOAT *x, FLOAT *y, FLOAT *c, FLOAT *s)
+{
+
+	BLASLONG i=0;
+	FLOAT f0, f1, f2, f3;
+	FLOAT x00, x01, x02, x03;
+	FLOAT g0, g1, g2, g3;
+	FLOAT y00, y01, y02, y03;
+	FLOAT *x1=x;
+	FLOAT *y1=y;
+	FLOAT c1=*c;
+	FLOAT s1=*s;
+
+	while ( i<n )
+	{
+
+		x00 = x1[0];
+		y00 = y1[0];
+		x01 = x1[1];
+		y01 = y1[1];
+		x02 = x1[2];
+		y02 = y1[2];
+		x03 = x1[3];
+		y03 = y1[3];
+
+		f0 = c1*x00 + s1*y00;
+		g0 = c1*y00 - s1*x00;
+		f1 = c1*x01 + s1*y01;
+		g1 = c1*y01 - s1*x01;
+		f2 = c1*x02 + s1*y02;
+		g2 = c1*y02 - s1*x02;
+		f3 = c1*x03 + s1*y03;
+		g3 = c1*y03 - s1*x03;
+
+		x1[0] = f0;
+		y1[0] = g0;
+		x1[1] = f1;
+		y1[1] = g1;
+		x1[2] = f2;
+		y1[2] = g2;
+		x1[3] = f3;
+		y1[3] = g3;
+
+		x1 += 4;
+		y1 += 4;
+
+		i+=4;
+	}
+	return;
+
+}
+
+
+#endif
+
+
+int CNAME(BLASLONG n, FLOAT *x, BLASLONG inc_x, FLOAT *y, BLASLONG inc_y, FLOAT c, FLOAT s)
+{
+	BLASLONG i=0;
+	BLASLONG ix=0,iy=0;
+	FLOAT c1[4] __attribute__ ((aligned (16)));;
+	FLOAT s1[4] __attribute__ ((aligned (16)));;
+	FLOAT *x1=x;
+	FLOAT *y1=y;
+	FLOAT temp;
+
+	if ( n <= 0     )  return(0);
+
+	if ( (inc_x == 1) && (inc_y == 1) )
+	{
+
+		BLASLONG n1 = n & -16;
+		if ( n1 > 0 )
+		{
+			c1[0]=c;
+			c1[1]=c;
+			c1[2]=c;
+			c1[3]=c;
+			s1[0]=s;
+			s1[1]=s;
+			s1[2]=s;
+			s1[3]=s;
+			drot_kernel_16(n1, x1, y1, c1, s1);
+			i=n1;
+		}
+
+		while(i < n)
+		{
+			temp  = c*x[i] + s*y[i] ;
+			y[i]  = c*y[i] - s*x[i] ;
+			x[i]  = temp ;
+
+			i++ ;
+
+		}
+
+
+	}
+	else
+	{
+
+		while(i < n)
+		{
+			temp   = c*x[ix] + s*y[iy] ;
+			y[iy]  = c*y[iy] - s*x[ix] ;
+			x[ix]  = temp ;
+
+			ix += inc_x ;
+			iy += inc_y ;
+			i++ ;
+
+		}
+
+	}
+	return(0);
+
+}
+
+
diff --git a/kernel/power/drot_microk_power8.c b/kernel/power/drot_microk_power8.c
new file mode 100644
index 000000000..4444ac7eb
--- /dev/null
+++ b/kernel/power/drot_microk_power8.c
@@ -0,0 +1,211 @@
+/***************************************************************************
+Copyright (c) 2013-2016, The OpenBLAS Project
+All rights reserved.
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+1. Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+2. Redistributions in binary form must reproduce the above copyright
+notice, this list of conditions and the following disclaimer in
+the documentation and/or other materials provided with the
+distribution.
+3. Neither the name of the OpenBLAS project nor the names of
+its contributors may be used to endorse or promote products
+derived from this software without specific prior written permission.
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ARE DISCLAIMED. IN NO EVENT SHALL THE OPENBLAS PROJECT OR CONTRIBUTORS BE
+LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
+USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*****************************************************************************/
+
+/**************************************************************************************
+* 2016/03/27 Werner Saar (wernsaar@googlemail.com)
+*
+* I don't use fused multiply-add ( precision problems with lapack )
+*
+* 	 BLASTEST 		: OK
+* 	 CTEST			: OK
+* 	 TEST			: OK
+*	 LAPACK-TEST		: OK
+**************************************************************************************/
+
+#define HAVE_KERNEL_16 1
+
+static void drot_kernel_16( BLASLONG n, FLOAT *x, FLOAT *y, FLOAT *c, FLOAT *s) __attribute__ ((noinline));
+
+static void drot_kernel_16( BLASLONG n, FLOAT *x, FLOAT *y, FLOAT *c, FLOAT *s)
+{
+
+
+	BLASLONG i = n;
+	BLASLONG o16 = 16;
+	BLASLONG o32 = 32;
+	BLASLONG o48 = 48;
+	FLOAT *x1=x;
+	FLOAT *y1=y;
+	FLOAT *x2=x+1;
+	FLOAT *y2=y+1;
+
+	__asm__  __volatile__
+	(
+
+        "lxsdx          36 , %5, %3                          \n\t"	// load c
+        "lxsdx          37 , %5, %4                          \n\t"	// load s
+	"addi		%8 , %8, -8			     \n\t"
+	"addi		%9 , %9, -8			     \n\t"
+
+        "xxspltd        36 , 36, 0                           \n\t"
+        "xxspltd        37 , 37, 0                           \n\t"
+
+	"lxvd2x		32, 0, %1			    \n\t"	// load x
+	"lxvd2x		33, %5, %1			    \n\t"
+	"lxvd2x		34, %6, %1			    \n\t"
+	"lxvd2x		35, %7, %1			    \n\t"
+
+	"lxvd2x		40, 0, %2			    \n\t"	// load y
+	"lxvd2x		41, %5, %2			    \n\t"
+	"lxvd2x		42, %6, %2			    \n\t"
+	"lxvd2x		43, %7, %2			    \n\t"
+
+	"addi		%1, %1, 64			    \n\t"
+	"addi		%2, %2, 64			    \n\t"
+
+	"addic.		%0 , %0	, -8  	 	             \n\t"
+	"ble		2f		             	     \n\t"
+
+	".align 5				            \n\t"
+	"1:				                    \n\t"
+
+	"xvmuldp	48, 32, 36		    	    \n\t"	// c * x
+	"xvmuldp	49, 33, 36		    	    \n\t"
+	"xvmuldp	50, 34, 36		    	    \n\t"
+	"xvmuldp	51, 35, 36		    	    \n\t"
+
+	"xvmuldp	56, 40, 36		    	    \n\t"	// c * y
+	"xvmuldp	57, 41, 36		    	    \n\t"
+	"xvmuldp	58, 42, 36		    	    \n\t"
+	"xvmuldp	59, 43, 36		    	    \n\t"
+
+	"xvmuldp	52, 32, 37		    	    \n\t"	// s * x
+	"xvmuldp	53, 33, 37		    	    \n\t"
+
+	"lxvd2x		32, 0, %1			    \n\t"	// load x
+	"lxvd2x		33, %5, %1			    \n\t"
+
+	"xvmuldp	54, 34, 37		    	    \n\t"
+	"xvmuldp	55, 35, 37		    	    \n\t"
+
+	"lxvd2x		34, %6, %1			    \n\t"
+	"lxvd2x		35, %7, %1			    \n\t"
+
+	"xvmuldp	60, 40, 37		    	    \n\t"	// s * y
+	"xvmuldp	61, 41, 37		    	    \n\t"
+
+	"lxvd2x		40, 0, %2			    \n\t"	// load y
+	"lxvd2x		41, %5, %2			    \n\t"
+
+	"xvmuldp	62, 42, 37		    	    \n\t"
+	"xvmuldp	63, 43, 37		    	    \n\t"
+
+	"lxvd2x		42, %6, %2			    \n\t"
+	"lxvd2x		43, %7, %2			    \n\t"
+
+	"xvadddp	48, 48 , 60			    \n\t"	// c * x + s * y 
+	"xvadddp	49, 49 , 61			    \n\t"	// c * x + s * y 
+
+	"addi		%1, %1, 64			    \n\t"
+	"addi		%2, %2, 64			    \n\t"
+
+	"xvadddp	50, 50 , 62			    \n\t"	// c * x + s * y 
+	"xvadddp	51, 51 , 63			    \n\t"	// c * x + s * y 
+
+	"xvsubdp	56, 56 , 52			    \n\t"	// c * y - s * x
+	"xvsubdp	57, 57 , 53			    \n\t"	// c * y - s * x
+	"xvsubdp	58, 58 , 54			    \n\t"	// c * y - s * x
+	"xvsubdp	59, 59 , 55			    \n\t"	// c * y - s * x
+
+	"stxvd2x	48, 0, %8			    \n\t"	// store x
+	"stxvd2x	49, %5, %8			    \n\t"
+	"stxvd2x	50, %6, %8			    \n\t"
+	"stxvd2x	51, %7, %8			    \n\t"
+
+	"stxvd2x	56, 0, %9			    \n\t"	// store y
+	"stxvd2x	57, %5, %9			    \n\t"
+	"stxvd2x	58, %6, %9			    \n\t"
+	"stxvd2x	59, %7, %9			    \n\t"
+
+	"addi		%8, %8, 64			    \n\t"
+	"addi		%9, %9, 64			    \n\t"
+
+	"addic.		%0 , %0	, -8  	 	             \n\t"
+	"bgt		1b		             	     \n\t"
+
+	"2:						     \n\t"
+
+	"xvmuldp	48, 32, 36		    	    \n\t"	// c * x
+	"xvmuldp	49, 33, 36		    	    \n\t"
+	"xvmuldp	50, 34, 36		    	    \n\t"
+	"xvmuldp	51, 35, 36		    	    \n\t"
+
+	"xvmuldp	56, 40, 36		    	    \n\t"	// c * y
+	"xvmuldp	57, 41, 36		    	    \n\t"
+	"xvmuldp	58, 42, 36		    	    \n\t"
+	"xvmuldp	59, 43, 36		    	    \n\t"
+
+	"xvmuldp	52, 32, 37		    	    \n\t"	// s * x
+	"xvmuldp	53, 33, 37		    	    \n\t"
+	"xvmuldp	54, 34, 37		    	    \n\t"
+	"xvmuldp	55, 35, 37		    	    \n\t"
+
+	"xvmuldp	60, 40, 37		    	    \n\t"	// s * y
+	"xvmuldp	61, 41, 37		    	    \n\t"
+	"xvmuldp	62, 42, 37		    	    \n\t"
+	"xvmuldp	63, 43, 37		    	    \n\t"
+
+	"xvadddp	48, 48 , 60			    \n\t"	// c * x + s * y 
+	"xvadddp	49, 49 , 61			    \n\t"	// c * x + s * y 
+	"xvadddp	50, 50 , 62			    \n\t"	// c * x + s * y 
+	"xvadddp	51, 51 , 63			    \n\t"	// c * x + s * y 
+
+	"xvsubdp	56, 56 , 52			    \n\t"	// c * y - s * x
+	"xvsubdp	57, 57 , 53			    \n\t"	// c * y - s * x
+	"xvsubdp	58, 58 , 54			    \n\t"	// c * y - s * x
+	"xvsubdp	59, 59 , 55			    \n\t"	// c * y - s * x
+
+	"stxvd2x	48, 0, %8			    \n\t"	// store x
+	"stxvd2x	49, %5, %8			    \n\t"
+	"stxvd2x	50, %6, %8			    \n\t"
+	"stxvd2x	51, %7, %8			    \n\t"
+
+	"stxvd2x	56, 0, %9			    \n\t"	// store y
+	"stxvd2x	57, %5, %9			    \n\t"
+	"stxvd2x	58, %6, %9			    \n\t"
+	"stxvd2x	59, %7, %9			    \n\t"
+
+
+
+	:
+        : 
+          "r" (i),	// 0	
+	  "r" (x1),  	// 1
+          "r" (y1),     // 2
+          "r" (c),      // 3
+          "r" (s),      // 4
+	  "r" (o16),	// 5
+	  "r" (o32),	// 6
+	  "r" (o48),    // 7
+	  "r" (x2),     // 8
+	  "r" (y2)      // 9
+	: "cr0", "%0", "%1" , "%2", "%8", "%9", "memory"
+	);
+
+} 
+
+
diff --git a/kernel/power/srot.c b/kernel/power/srot.c
new file mode 100644
index 000000000..d464846a4
--- /dev/null
+++ b/kernel/power/srot.c
@@ -0,0 +1,167 @@
+/***************************************************************************
+Copyright (c) 2013-2016, The OpenBLAS Project
+All rights reserved.
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+1. Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+2. Redistributions in binary form must reproduce the above copyright
+notice, this list of conditions and the following disclaimer in
+the documentation and/or other materials provided with the
+distribution.
+3. Neither the name of the OpenBLAS project nor the names of
+its contributors may be used to endorse or promote products
+derived from this software without specific prior written permission.
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ARE DISCLAIMED. IN NO EVENT SHALL THE OPENBLAS PROJECT OR CONTRIBUTORS BE
+LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
+USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*****************************************************************************/
+
+/**************************************************************************************
+* 2016/03/26 Werner Saar (wernsaar@googlemail.com)
+* 	 BLASTEST 		: OK
+* 	 CTEST			: OK
+* 	 TEST			: OK
+*	 LAPACK-TEST		: OK
+**************************************************************************************/
+
+
+
+#include "common.h"
+
+#pragma GCC optimize "O1"
+
+#if defined(POWER8)
+#include "srot_microk_power8.c"
+#endif
+
+
+#ifndef HAVE_KERNEL_16
+
+static void srot_kernel_16(BLASLONG n, FLOAT *x, FLOAT *y, FLOAT *c, FLOAT *s)
+{
+
+	BLASLONG i=0;
+	FLOAT f0, f1, f2, f3;
+	FLOAT x00, x01, x02, x03;
+	FLOAT g0, g1, g2, g3;
+	FLOAT y00, y01, y02, y03;
+	FLOAT *x1=x;
+	FLOAT *y1=y;
+	FLOAT c1=*c;
+	FLOAT s1=*s;
+
+	while ( i<n )
+	{
+
+		x00 = x1[0];
+		y00 = y1[0];
+		x01 = x1[1];
+		y01 = y1[1];
+		x02 = x1[2];
+		y02 = y1[2];
+		x03 = x1[3];
+		y03 = y1[3];
+
+		f0 = c1*x00 + s1*y00;
+		g0 = c1*y00 - s1*x00;
+		f1 = c1*x01 + s1*y01;
+		g1 = c1*y01 - s1*x01;
+		f2 = c1*x02 + s1*y02;
+		g2 = c1*y02 - s1*x02;
+		f3 = c1*x03 + s1*y03;
+		g3 = c1*y03 - s1*x03;
+
+		x1[0] = f0;
+		y1[0] = g0;
+		x1[1] = f1;
+		y1[1] = g1;
+		x1[2] = f2;
+		y1[2] = g2;
+		x1[3] = f3;
+		y1[3] = g3;
+
+		x1 += 4;
+		y1 += 4;
+
+		i+=4;
+	}
+	return;
+
+}
+
+
+#endif
+
+
+int CNAME(BLASLONG n, FLOAT *x, BLASLONG inc_x, FLOAT *y, BLASLONG inc_y, FLOAT c, FLOAT s)
+{
+	BLASLONG i=0;
+	BLASLONG ix=0,iy=0;
+	FLOAT c1[4] __attribute__ ((aligned (16)));;
+	FLOAT s1[4] __attribute__ ((aligned (16)));;
+	FLOAT *x1=x;
+	FLOAT *y1=y;
+	FLOAT temp;
+
+	if ( n <= 0     )  return(0);
+
+	if ( (inc_x == 1) && (inc_y == 1) )
+	{
+
+		BLASLONG n1 = n & -16;
+		if ( n1 > 0 )
+		{
+			c1[0]=c;
+			c1[1]=c;
+			c1[2]=c;
+			c1[3]=c;
+			s1[0]=s;
+			s1[1]=s;
+			s1[2]=s;
+			s1[3]=s;
+			srot_kernel_16(n1, x1, y1, c1, s1);
+			i=n1;
+		}
+
+		while(i < n)
+		{
+			temp  = c*x[i] + s*y[i] ;
+			y[i]  = c*y[i] - s*x[i] ;
+			x[i]  = temp ;
+
+			i++ ;
+
+		}
+
+
+	}
+	else
+	{
+
+		while(i < n)
+		{
+			temp   = c*x[ix] + s*y[iy] ;
+			y[iy]  = c*y[iy] - s*x[ix] ;
+			x[ix]  = temp ;
+
+			ix += inc_x ;
+			iy += inc_y ;
+			i++ ;
+
+		}
+
+	}
+	return(0);
+
+}
+
+
diff --git a/kernel/power/srot_microk_power8.c b/kernel/power/srot_microk_power8.c
new file mode 100644
index 000000000..ade65500f
--- /dev/null
+++ b/kernel/power/srot_microk_power8.c
@@ -0,0 +1,208 @@
+/***************************************************************************
+Copyright (c) 2013-2016, The OpenBLAS Project
+All rights reserved.
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are
+met:
+1. Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+2. Redistributions in binary form must reproduce the above copyright
+notice, this list of conditions and the following disclaimer in
+the documentation and/or other materials provided with the
+distribution.
+3. Neither the name of the OpenBLAS project nor the names of
+its contributors may be used to endorse or promote products
+derived from this software without specific prior written permission.
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ARE DISCLAIMED. IN NO EVENT SHALL THE OPENBLAS PROJECT OR CONTRIBUTORS BE
+LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
+USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*****************************************************************************/
+
+/**************************************************************************************
+* 2016/03/27 Werner Saar (wernsaar@googlemail.com)
+*
+* I don't use fused multiply-add ( precision problems with lapack )
+*
+* 	 BLASTEST 		: OK
+* 	 CTEST			: OK
+* 	 TEST			: OK
+*	 LAPACK-TEST		: OK
+**************************************************************************************/
+
+#define HAVE_KERNEL_16 1
+
+static void srot_kernel_16( BLASLONG n, FLOAT *x, FLOAT *y, FLOAT *c, FLOAT *s) __attribute__ ((noinline));
+
+static void srot_kernel_16( BLASLONG n, FLOAT *x, FLOAT *y, FLOAT *c, FLOAT *s)
+{
+
+
+	BLASLONG i = n;
+	BLASLONG o16 = 16;
+	BLASLONG o32 = 32;
+	BLASLONG o48 = 48;
+	FLOAT *x1=x;
+	FLOAT *y1=y;
+	FLOAT *x2=x+1;
+	FLOAT *y2=y+1;
+
+	__asm__  __volatile__
+	(
+
+        "lxvw4x         36 , 0, %3                          \n\t"	// load c
+        "lxvw4x         37 , 0, %4                          \n\t"	// load s
+	"addi		%8 , %8, -4			     \n\t"
+	"addi		%9 , %9, -4			     \n\t"
+
+	"lxvw4x		32, 0, %1			    \n\t"	// load x
+	"lxvw4x		33, %5, %1			    \n\t"
+	"lxvw4x		34, %6, %1			    \n\t"
+	"lxvw4x		35, %7, %1			    \n\t"
+
+	"lxvw4x		40, 0, %2			    \n\t"	// load y
+	"lxvw4x		41, %5, %2			    \n\t"
+	"lxvw4x		42, %6, %2			    \n\t"
+	"lxvw4x		43, %7, %2			    \n\t"
+
+	"addi		%1, %1, 64			    \n\t"
+	"addi		%2, %2, 64			    \n\t"
+
+	"addic.		%0 , %0	, -16  	 	             \n\t"
+	"ble		2f		             	     \n\t"
+
+	".align 5				            \n\t"
+	"1:				                    \n\t"
+
+	"xvmulsp	48, 32, 36		    	    \n\t"	// c * x
+	"xvmulsp	49, 33, 36		    	    \n\t"
+	"xvmulsp	50, 34, 36		    	    \n\t"
+	"xvmulsp	51, 35, 36		    	    \n\t"
+
+	"xvmulsp	56, 40, 36		    	    \n\t"	// c * y
+	"xvmulsp	57, 41, 36		    	    \n\t"
+	"xvmulsp	58, 42, 36		    	    \n\t"
+	"xvmulsp	59, 43, 36		    	    \n\t"
+
+	"xvmulsp	52, 32, 37		    	    \n\t"	// s * x
+	"xvmulsp	53, 33, 37		    	    \n\t"
+
+	"lxvw4x		32, 0, %1			    \n\t"	// load x
+	"lxvw4x		33, %5, %1			    \n\t"
+
+	"xvmulsp	54, 34, 37		    	    \n\t"
+	"xvmulsp	55, 35, 37		    	    \n\t"
+
+	"lxvw4x		34, %6, %1			    \n\t"
+	"lxvw4x		35, %7, %1			    \n\t"
+
+	"xvmulsp	60, 40, 37		    	    \n\t"	// s * y
+	"xvmulsp	61, 41, 37		    	    \n\t"
+
+	"lxvw4x		40, 0, %2			    \n\t"	// load y
+	"lxvw4x		41, %5, %2			    \n\t"
+
+	"xvmulsp	62, 42, 37		    	    \n\t"
+	"xvmulsp	63, 43, 37		    	    \n\t"
+
+	"lxvw4x		42, %6, %2			    \n\t"
+	"lxvw4x		43, %7, %2			    \n\t"
+
+	"xvaddsp	48, 48 , 60			    \n\t"	// c * x + s * y 
+	"xvaddsp	49, 49 , 61			    \n\t"	// c * x + s * y 
+
+	"addi		%1, %1, 64			    \n\t"
+	"addi		%2, %2, 64			    \n\t"
+
+	"xvaddsp	50, 50 , 62			    \n\t"	// c * x + s * y 
+	"xvaddsp	51, 51 , 63			    \n\t"	// c * x + s * y 
+
+	"xvsubsp	56, 56 , 52			    \n\t"	// c * y - s * x
+	"xvsubsp	57, 57 , 53			    \n\t"	// c * y - s * x
+	"xvsubsp	58, 58 , 54			    \n\t"	// c * y - s * x
+	"xvsubsp	59, 59 , 55			    \n\t"	// c * y - s * x
+
+	"stxvw4x	48, 0, %8			    \n\t"	// store x
+	"stxvw4x	49, %5, %8			    \n\t"
+	"stxvw4x	50, %6, %8			    \n\t"
+	"stxvw4x	51, %7, %8			    \n\t"
+
+	"stxvw4x	56, 0, %9			    \n\t"	// store y
+	"stxvw4x	57, %5, %9			    \n\t"
+	"stxvw4x	58, %6, %9			    \n\t"
+	"stxvw4x	59, %7, %9			    \n\t"
+
+	"addi		%8, %8, 64			    \n\t"
+	"addi		%9, %9, 64			    \n\t"
+
+	"addic.		%0 , %0	, -16  	 	             \n\t"
+	"bgt		1b		             	     \n\t"
+
+	"2:						     \n\t"
+
+	"xvmulsp	48, 32, 36		    	    \n\t"	// c * x
+	"xvmulsp	49, 33, 36		    	    \n\t"
+	"xvmulsp	50, 34, 36		    	    \n\t"
+	"xvmulsp	51, 35, 36		    	    \n\t"
+
+	"xvmulsp	56, 40, 36		    	    \n\t"	// c * y
+	"xvmulsp	57, 41, 36		    	    \n\t"
+	"xvmulsp	58, 42, 36		    	    \n\t"
+	"xvmulsp	59, 43, 36		    	    \n\t"
+
+	"xvmulsp	52, 32, 37		    	    \n\t"	// s * x
+	"xvmulsp	53, 33, 37		    	    \n\t"
+	"xvmulsp	54, 34, 37		    	    \n\t"
+	"xvmulsp	55, 35, 37		    	    \n\t"
+
+	"xvmulsp	60, 40, 37		    	    \n\t"	// s * y
+	"xvmulsp	61, 41, 37		    	    \n\t"
+	"xvmulsp	62, 42, 37		    	    \n\t"
+	"xvmulsp	63, 43, 37		    	    \n\t"
+
+	"xvaddsp	48, 48 , 60			    \n\t"	// c * x + s * y 
+	"xvaddsp	49, 49 , 61			    \n\t"	// c * x + s * y 
+	"xvaddsp	50, 50 , 62			    \n\t"	// c * x + s * y 
+	"xvaddsp	51, 51 , 63			    \n\t"	// c * x + s * y 
+
+	"xvsubsp	56, 56 , 52			    \n\t"	// c * y - s * x
+	"xvsubsp	57, 57 , 53			    \n\t"	// c * y - s * x
+	"xvsubsp	58, 58 , 54			    \n\t"	// c * y - s * x
+	"xvsubsp	59, 59 , 55			    \n\t"	// c * y - s * x
+
+	"stxvw4x	48, 0, %8			    \n\t"	// store x
+	"stxvw4x	49, %5, %8			    \n\t"
+	"stxvw4x	50, %6, %8			    \n\t"
+	"stxvw4x	51, %7, %8			    \n\t"
+
+	"stxvw4x	56, 0, %9			    \n\t"	// store y
+	"stxvw4x	57, %5, %9			    \n\t"
+	"stxvw4x	58, %6, %9			    \n\t"
+	"stxvw4x	59, %7, %9			    \n\t"
+
+
+
+	:
+        : 
+          "r" (i),	// 0	
+	  "r" (x1),  	// 1
+          "r" (y1),     // 2
+          "r" (c),      // 3
+          "r" (s),      // 4
+	  "r" (o16),	// 5
+	  "r" (o32),	// 6
+	  "r" (o48),    // 7
+	  "r" (x2),     // 8
+	  "r" (y2)      // 9
+	: "cr0", "%0", "%1" , "%2", "%8", "%9", "memory"
+	);
+
+} 
+
+