sl@0: /*
sl@0:  * LIBOIL - Library of Optimized Inner Loops
sl@0:  * Copyright (c) 2003,2004 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 <liboil/liboilfunction.h>
sl@0: #include "liboil/simdpack/simdpack.h"
sl@0: 
sl@0: #include <math.h>
sl@0: 
sl@0: static void
sl@0: average2_u8_trick (uint8_t * dest, int dstr, uint8_t *src1, int sstr1,
sl@0:     uint8_t *src2, int sstr2, int n)
sl@0: {
sl@0:   unsigned int x, y, d;
sl@0: 
sl@0:   while (n&3) {
sl@0:     *dest = (*src1 + *src2) >> 1;
sl@0:     src1 += sstr1;
sl@0:     src2 += sstr2;
sl@0:     dest += dstr;
sl@0:     n--;
sl@0:   }
sl@0:   while (n > 0) {
sl@0:     x = (src1[0] << 24) | (src1[sstr1] << 16) | (src1[2 *
sl@0:             sstr1] << 8) | (src1[3 * sstr1]);
sl@0:     y = (src2[0] << 24) | (src2[sstr2] << 16) | (src2[2 *
sl@0:             sstr2] << 8) | (src2[3 * sstr2]);
sl@0:     d = (((x ^ y) & 0xfefefefe) >> 1) + (x & y);
sl@0:     dest[0] = (d >> 24);
sl@0:     dest[1*dstr] = (d >> 16);
sl@0:     dest[2*dstr] = (d >> 8);
sl@0:     dest[3*dstr] = (d >> 0);
sl@0:     src1 += 4 * sstr1;
sl@0:     src2 += 4 * sstr2;
sl@0:     dest += 4 * dstr;
sl@0:     n -= 4;
sl@0:   }
sl@0: }
sl@0: 
sl@0: OIL_DEFINE_IMPL (average2_u8_trick, average2_u8);
sl@0: 
sl@0: static void
sl@0: average2_u8_fast (uint8_t * dest, int dstr, uint8_t *src1, int sstr1,
sl@0:     uint8_t *src2, int sstr2, int n)
sl@0: {
sl@0:   while (n > 0) {
sl@0:     *dest = (*src1 + *src2) >> 1;
sl@0:     src1 += sstr1;
sl@0:     src2 += sstr2;
sl@0:     dest += dstr;
sl@0:     n--;
sl@0:   }
sl@0: }
sl@0: 
sl@0: OIL_DEFINE_IMPL (average2_u8_fast, average2_u8);
sl@0: 
sl@0: static void
sl@0: average2_u8_unroll4 (uint8_t * dest, int dstr, uint8_t *src1, int sstr1,
sl@0:     uint8_t *src2, int sstr2, int n)
sl@0: {
sl@0:   while (n & 0x3) {
sl@0:     *dest = (*src1 + *src2) >> 1;
sl@0:     dest += dstr;
sl@0:     src1 += sstr1;
sl@0:     src2 += sstr2;
sl@0:     n--;
sl@0:   }
sl@0:   while (n > 0) {
sl@0:     *dest = (*src1 + *src2) >> 1;
sl@0:     dest += dstr;
sl@0:     src1 += sstr1;
sl@0:     src2 += sstr2;
sl@0:     *dest = (*src1 + *src2) >> 1;
sl@0:     dest += dstr;
sl@0:     src1 += sstr1;
sl@0:     src2 += sstr2;
sl@0:     *dest = (*src1 + *src2) >> 1;
sl@0:     dest += dstr;
sl@0:     src1 += sstr1;
sl@0:     src2 += sstr2;
sl@0:     *dest = (*src1 + *src2) >> 1;
sl@0:     dest += dstr;
sl@0:     src1 += sstr1;
sl@0:     src2 += sstr2;
sl@0:     n -= 4;
sl@0:   }
sl@0: }
sl@0: 
sl@0: OIL_DEFINE_IMPL (average2_u8_unroll4, average2_u8);
sl@0: 
sl@0: 
sl@0: 
sl@0: #ifdef	__SYMBIAN32__
sl@0:  
sl@0: OilFunctionImpl* __oil_function_impl_average2_u8_trick() {
sl@0: 		return &_oil_function_impl_average2_u8_trick;
sl@0: }
sl@0: #endif
sl@0: 
sl@0: #ifdef	__SYMBIAN32__
sl@0:  
sl@0: OilFunctionImpl* __oil_function_impl_average2_u8_fast() {
sl@0: 		return &_oil_function_impl_average2_u8_fast;
sl@0: }
sl@0: #endif
sl@0: 
sl@0: #ifdef	__SYMBIAN32__
sl@0:  
sl@0: OilFunctionImpl* __oil_function_impl_average2_u8_unroll4() {
sl@0: 		return &_oil_function_impl_average2_u8_unroll4;
sl@0: }
sl@0: #endif
sl@0: