From 1d0a78fa65835f24b42a8bbba89e65caef457de0 Mon Sep 17 00:00:00 2001
From: RI2757 <alisson_ricardo_gr@hotmail.com>
Date: Mon, 13 Apr 2015 20:11:59 -0300
Subject: [PATCH] =?UTF-8?q?Exerc=C3=ADcio=201?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 sse.c | 72 +++++++++++++++++++++++++----------------------------------
 1 file changed, 31 insertions(+), 41 deletions(-)

diff --git a/sse.c b/sse.c
index 211722f..16ac603 100644
--- a/sse.c
+++ b/sse.c
@@ -2,7 +2,7 @@
 Exemplo de código SSE utilizando Intel instrinsics e gcc built-in functions
 
 Compile usando:
-gcc sse.c -o sse -msse -msse4.2 -O3
+gcc sse.c -o sse -msse -msse4.2
 
 
 http://www.songho.ca/misc/sse/sse.html
@@ -24,7 +24,6 @@ AVX
 #include <pmmintrin.h> // SSE3
 #include <smmintrin.h> // SSE4.1
 
-
 #define VECTOR_SIZE         4
 typedef float v4sf __attribute__ ((vector_size(sizeof(float)*VECTOR_SIZE))); 
 
@@ -34,7 +33,6 @@ typedef union f4vector
     float   f[VECTOR_SIZE];
 } f4vector;
 
-
 void add_intel_intrinsics(float *a, float *b, float *c)
 {
   __m128 va = _mm_load_ps (a);
@@ -52,60 +50,52 @@ void add_intel_intrinsics(float *a, float *b, float *c)
   */
 }
 
-
-
-v4sf add_gcc_builtin(v4sf a, v4sf b)
-{
-  return __builtin_ia32_addps (a, b);
-}
-
-
-
 int main (int argc, char *argv[])
 {
   float *a __attribute__ ((aligned(16))) = (float*)malloc (sizeof(float) * 4); //aloca um vetor de 16bytes (128bits) alinhado em endereços múltiplos de 16bytes.
   float *b __attribute__ ((aligned(16))) = (float*)malloc (sizeof(float) * 4);
-  float *c __attribute__ ((aligned(16))) = (float*)malloc (sizeof(float) * 4);
-  
+
   int i = 0;
-  
+  int j = 10;
+  int x, y, w=2,h=2, c=0;
   for (i = 0; i < 4; ++i) {
-    a[i] = i;
-    b[i] = i;
+    a[i] = 200+j;
+    b[i] = 10;
+    j = j +10;
   }
-  
-  printf("Intel SSE\n");
-  
-  add_intel_intrinsics(a, b, c);
 
+  printf("vetor A \n");
   for (i = 0; i < 4; ++i) {
-    printf("%f\n", c[i]);
+    printf("%.2f\n", a[i]);
   }
-  
-  free(a);
-  free(b);
-  free(c);
-  
-  printf("\nGCC Built-in Functions\n");
-  
-  v4sf d, e, f;
-
-  d = (v4sf){0, 1, 2, 3};
-  e = (v4sf){0, 1, 2, 3};
-  f = add_gcc_builtin(d, e);  
 
+  printf("\nvetor B \n");
   for (i = 0; i < 4; ++i) {
-    printf("%f\n", f[i]);
+    printf("%.2f\n", b[i]);
   }
   
-  
-  printf("\nGCC implicity vectorization\n");
-  
-  f = d + e;  
+for(i=0;i<(h*w);i+=4){
+  	__m128 img= _mm_load_ps(a);
+  	__m128 intensidade= _mm_load_ps(b);
+  	__m128 r = _mm_add_ps(img,intensidade);
+   	_mm_store_ps(b,r);
+	c+=4;
+  }
 
+ 
+ printf("\nvetor Alterado\n");
+    for(y=0;y<h;++y){
+		for(x=0;x<w;++x){
+			printf("%.2f \n",b[y*w+x]);
+		}
+	}
+ /* printf("vetor Alterado \n");
   for (i = 0; i < 4; ++i) {
-    printf("%f\n", f[i]);
-  }
+    printf("%.2f\n", b[i]);
+  }*/
+  
+  free(a);
+  free(b);
 
   return 0;
 }