/*************************************************************************** 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" ); }