project

lec10/gptest.c

+#include <stdio.h>
+// #### #### #### #### .  #### #### #### #### 
+// S 15 . 16 
+#define FX_Q_NUM 16
+#define FX_2_MINUS_16  1.52587890625e-05F
+#define FX_2_PLUS_16   (1<<16)
+#define FX_S_15_16 11516
+#define FX_SYSTEM  FX_S_15_16
+
+typedef int fixed32; 
+
+fixed32 fromFloat(float fa)
+{
+	return (fixed32) (fa * FX_2_PLUS_16) ; 
+}
+
+float toFloat(fixed32 xa)
+{
+	return ((float) (xa)) * FX_2_MINUS_16 ; 
+}
+
+fixed32 fxAdd(fixed32 a, fixed32 b)
+{
+	return fromFloat((toFloat(a) + toFloat(b))); 
+}
+
+fixed32 fxAdd2(fixed32 a, fixed32 b)
+{
+	return a+b;
+}
+
+int main()
+{
+	long long i=0; 
+	int ia, ib, ic, ic2; 
+	float fa; 
+	// fscanf(stdin, "%d %d", &ia, &ib); 
+	for ( i = 0; i < (long long) 256*256*256*256; i+=(256*256))
+	{
+		ic = fxAdd(i, i); 
+		ic2 = fxAdd2(i, i); 
+		fprintf(stdout, "%f + %f : %f , %f diff = %d\n", toFloat(i), toFloat(i), toFloat(ic), toFloat(ic2), ic-ic2); 
+	}
+	//fprintf(stdout, "%d + %d : %d\n", ia, ib , ic) ; 
+	//fprintf(stdout, "%f + %f : %f\n", toFloat(ia), toFloat(ib), toFloat(ic)); 
+}

lec10/test.c

@@ -23,7 +23,7 @@ float toFloat_fn(fixed32 xa)
 
 fixed32 fxAdd(fixed32 a, fixed32 b)
 {
-	return fromFloat((toFloat(a) + toFloat(b))); 
+	return (fixed32)(FX_2_PLUS_16 * ((((float) (a)) * FX_2_MINUS_16)+(((float) (b)) * FX_2_MINUS_16))); 
 }
 
 fixed32 fxAdd2(fixed32 a, fixed32 b)

test/pjtest.c

+#include <stdio.h>
+int main(){
+	printf("%d \n", ((1<<16)|(17<<8)|(14)));
+}