sl@0: /*
sl@0:  * LIBOIL - Library of Optimized Inner Loops
sl@0:  * Copyright (c) 2005 David A. Schleef <ds@schleef.org>
sl@0:  * All rights reserved.
sl@0:  *
sl@0:  * Redistribution and use in source and binary forms, with or without
sl@0:  * modification, are permitted provided that the following conditions
sl@0:  * are met:
sl@0:  * 1. Redistributions of source code must retain the above copyright
sl@0:  *    notice, this list of conditions and the following disclaimer.
sl@0:  * 2. Redistributions in binary form must reproduce the above copyright
sl@0:  *    notice, this list of conditions and the following disclaimer in the
sl@0:  *    documentation and/or other materials provided with the distribution.
sl@0:  * 
sl@0:  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
sl@0:  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
sl@0:  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
sl@0:  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT,
sl@0:  * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
sl@0:  * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
sl@0:  * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
sl@0:  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
sl@0:  * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
sl@0:  * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
sl@0:  * POSSIBILITY OF SUCH DAMAGE.
sl@0:  */
sl@0: //Portions Copyright (c)  2008-2009 Nokia Corporation and/or its subsidiary(-ies). All rights reserved. 
sl@0: 
sl@0: #ifdef HAVE_CONFIG_H
sl@0: #include "config.h"
sl@0: #endif
sl@0: 
sl@0: #include <math.h>
sl@0: 
sl@0: #include <liboil/liboil.h>
sl@0: #include <liboil/liboilfunction.h>
sl@0: #include <liboil/liboilclasses.h>
sl@0: 
sl@0: 
sl@0: #if 0
sl@0: static void
sl@0: add_f32_unroll4 (float *dest, float *src1, float *src2, int n)
sl@0: {
sl@0:   int i;
sl@0: 
sl@0:   for(i=0;i<(n&(~0x3));i+=4){
sl@0:     dest[i+0] = src1[i+0] + src2[i+0];
sl@0:     dest[i+1] = src1[i+1] + src2[i+1];
sl@0:     dest[i+2] = src1[i+2] + src2[i+2];
sl@0:     dest[i+3] = src1[i+3] + src2[i+3];
sl@0:   }
sl@0:   for(;i<n;i++){
sl@0:     dest[i] = src1[i] + src2[i];
sl@0:   }
sl@0: }
sl@0: OIL_DEFINE_IMPL (add_f32_unroll4, add_f32);
sl@0: #endif
sl@0: 
sl@0: #if 0
sl@0: static void
sl@0: add_f32_unroll4b (float *dest, float *src1, float *src2, int n)
sl@0: {
sl@0:   int i;
sl@0: 
sl@0:   for(i=0;i<(n&(~0x3));i+=4){
sl@0:     *dest++ = *src1++ + *src2++;
sl@0:     *dest++ = *src1++ + *src2++;
sl@0:     *dest++ = *src1++ + *src2++;
sl@0:     *dest++ = *src1++ + *src2++;
sl@0:   }
sl@0:   for(;i<n;i++){
sl@0:     *dest++ = *src1++ + *src2++;
sl@0:   }
sl@0: }
sl@0: OIL_DEFINE_IMPL (add_f32_unroll4b, add_f32);
sl@0: #endif
sl@0: 
sl@0: 
sl@0: #if 0
sl@0: static void
sl@0: multiply_f32_unroll4 (float *dest, float *src1, float *src2, int n)
sl@0: {
sl@0:   int i;
sl@0: 
sl@0:   for(i=0;i<(n&(~0x3));i+=4){
sl@0:     dest[i+0] = src1[i+0] * src2[i+0];
sl@0:     dest[i+1] = src1[i+1] * src2[i+1];
sl@0:     dest[i+2] = src1[i+2] * src2[i+2];
sl@0:     dest[i+3] = src1[i+3] * src2[i+3];
sl@0:   }
sl@0:   for(;i<n;i++){
sl@0:     dest[i] = src1[i] * src2[i];
sl@0:   }
sl@0: }
sl@0: OIL_DEFINE_IMPL (multiply_f32_unroll4, multiply_f32);
sl@0: #endif
sl@0: 
sl@0: static void
sl@0: scalaradd_f32_ns_unroll4 (float *dest, float *src1, float *src2, int n)
sl@0: {
sl@0:   int i;
sl@0: 
sl@0:   for(i=0;i<(n&(~0x3));i+=4){
sl@0:     dest[i+0] = src1[i+0] + src2[0];
sl@0:     dest[i+1] = src1[i+1] + src2[0];
sl@0:     dest[i+2] = src1[i+2] + src2[0];
sl@0:     dest[i+3] = src1[i+3] + src2[0];
sl@0:   }
sl@0:   for(;i<n;i++){
sl@0:     dest[i] = src1[i] + src2[0];
sl@0:   }
sl@0: }
sl@0: OIL_DEFINE_IMPL (scalaradd_f32_ns_unroll4, scalaradd_f32_ns);
sl@0: 
sl@0: static void
sl@0: scalarmultiply_f32_ns_unroll4 (float *dest, float *src1, float *src2, int n)
sl@0: {
sl@0:   int i;
sl@0: 
sl@0:   for(i=0;i<(n&(~0x3));i+=4){
sl@0:     dest[i+0] = src1[i+0] * src2[0];
sl@0:     dest[i+1] = src1[i+1] * src2[0];
sl@0:     dest[i+2] = src1[i+2] * src2[0];
sl@0:     dest[i+3] = src1[i+3] * src2[0];
sl@0:   }
sl@0:   for(;i<n;i++){
sl@0:     dest[i] = src1[i] * src2[0];
sl@0:   }
sl@0: }
sl@0: OIL_DEFINE_IMPL (scalarmultiply_f32_ns_unroll4, scalarmultiply_f32_ns);
sl@0: 
sl@0: static void
sl@0: scalarmultiply_f64_ns_unroll4 (double *dest, double *src1, double *src2, int n)
sl@0: {
sl@0:   int i;
sl@0: 
sl@0:   for(i=0;i<(n&(~0x3));i+=4){
sl@0:     dest[i+0] = src1[i+0] * src2[0];
sl@0:     dest[i+1] = src1[i+1] * src2[0];
sl@0:     dest[i+2] = src1[i+2] * src2[0];
sl@0:     dest[i+3] = src1[i+3] * src2[0];
sl@0:   }
sl@0:   for(;i<n;i++){
sl@0:     dest[i] = src1[i] * src2[0];
sl@0:   }
sl@0: }
sl@0: OIL_DEFINE_IMPL (scalarmultiply_f64_ns_unroll4, scalarmultiply_f64_ns);
sl@0: 
sl@0: #if 0
sl@0: 
sl@0: #ifdef	__SYMBIAN32__
sl@0:  
sl@0: OilFunctionImpl* __oil_function_impl_add_f32_unroll4() {
sl@0: 		return &_oil_function_impl_add_f32_unroll4;
sl@0: }
sl@0: #endif
sl@0: 
sl@0: #ifdef	__SYMBIAN32__
sl@0:  
sl@0: OilFunctionImpl* __oil_function_impl_add_f32_unroll4b() {
sl@0: 		return &_oil_function_impl_add_f32_unroll4b;
sl@0: }
sl@0: #endif
sl@0: 
sl@0: #ifdef	__SYMBIAN32__
sl@0:  
sl@0: OilFunctionImpl* __oil_function_impl_multiply_f32_unroll4() {
sl@0: 		return &_oil_function_impl_multiply_f32_unroll4;
sl@0: }
sl@0: #endif
sl@0: 
sl@0: #endif
sl@0: #ifdef	__SYMBIAN32__
sl@0:  
sl@0: OilFunctionImpl* __oil_function_impl_scalaradd_f32_ns_unroll4() {
sl@0: 		return &_oil_function_impl_scalaradd_f32_ns_unroll4;
sl@0: }
sl@0: #endif
sl@0: 
sl@0: #ifdef	__SYMBIAN32__
sl@0:  
sl@0: OilFunctionImpl* __oil_function_impl_scalarmultiply_f32_ns_unroll4() {
sl@0: 		return &_oil_function_impl_scalarmultiply_f32_ns_unroll4;
sl@0: }
sl@0: #endif
sl@0: 
sl@0: #ifdef	__SYMBIAN32__
sl@0:  
sl@0: OilFunctionImpl* __oil_function_impl_scalarmultiply_f64_ns_unroll4() {
sl@0: 		return &_oil_function_impl_scalarmultiply_f64_ns_unroll4;
sl@0: }
sl@0: #endif
sl@0: