sse2 version of vec4 dot product

* use this for normalizing vector

include/cglm/vec4.h

@@ -122,7 +122,14 @@ glm_vec4_copy(vec4 v, vec4 dest) {
 CGLM_INLINE
 float
 glm_vec4_dot(vec4 a, vec4 b) {
+#if defined( __SSE__ ) || defined( __SSE2__ )
+  __m128 x0;
+  x0 = _mm_mul_ps(_mm_load_ps(a), _mm_load_ps(b));
+  x0 = _mm_add_ps(x0, _mm_shuffle1_ps(x0, 1, 0, 3, 2));
+  return _mm_cvtss_f32(_mm_add_ss(x0, _mm_shuffle1_ps(x0, 0, 1, 0, 1)));
+#else
   return a[0] * b[0] + a[1] * b[1] + a[2] * b[2] + a[3] * b[3];
+#endif
 }
 
 /*!
@@ -139,7 +146,7 @@ glm_vec4_dot(vec4 a, vec4 b) {
 CGLM_INLINE
 float
 glm_vec4_norm2(vec4 v) {
-  return v[0] * v[0] + v[1] * v[1] + v[2] * v[2] + v[3] * v[3];
+  return glm_vec_dot(v, v);
 }
 
 /*!

makefile.am

@@ -109,6 +109,8 @@ test_tests_SOURCES=\
     test/src/test_project.c \
     test/src/test_clamp.c \
     test/src/test_euler.c \
-    test/src/test_quat.c
+    test/src/test_quat.c \
+    test/src/test_vec4.c
+
 all-local:
 	sh ./post-build.sh

test/src/test_common.c

@@ -36,6 +36,16 @@ test_rand_vec3(vec3 dest) {
   dest[2] = drand48();
 }
 
+void
+test_rand_vec4(vec4 dest) {
+  srand((unsigned int)time(NULL));
+
+  dest[0] = drand48();
+  dest[1] = drand48();
+  dest[2] = drand48();
+  dest[3] = drand48();
+}
+
 float
 test_rand_angle(void) {
   srand((unsigned int)time(NULL));

test/src/test_common.h

@@ -40,6 +40,9 @@ test_assert_quat_eq(versor v1, versor v2);
 void
 test_rand_vec3(vec3 dest);
 
+void
+test_rand_vec4(vec4 dest) ;
+
 float
 test_rand_angle(void);
 

test/src/test_main.c

@@ -26,7 +26,10 @@ main(int argc, const char * argv[]) {
     cmocka_unit_test(test_euler),
 
     /* quaternion */
-    cmocka_unit_test(test_quat)
+    cmocka_unit_test(test_quat),
+
+    /* vec4 */
+    cmocka_unit_test(test_vec4)
   };
 
   return cmocka_run_group_tests(tests, NULL, NULL);

test/src/test_tests.h

@@ -28,4 +28,7 @@ test_euler(void **state);
 void
 test_quat(void **state);
 
+void
+test_vec4(void **state);
+
 #endif /* test_tests_h */

test/src/test_vec4.c

@@ -0,0 +1,30 @@
+/*
+ * Copyright (c), Recep Aslantas.
+ *
+ * MIT License (MIT), http://opensource.org/licenses/MIT
+ * Full license can be found in the LICENSE file
+ */
+
+#include "test_common.h"
+
+CGLM_INLINE
+float
+test_vec4_dot(vec4 a, vec4 b) {
+  return a[0] * b[0] + a[1] * b[1] + a[2] * b[2] + a[3] * b[3];
+}
+
+void
+test_vec4(void **state) {
+  vec4  v;
+  int   i;
+  float d1, d2;
+
+  /* test SSE/SIMD dot product */
+  for (i = 0; i < 100; i++) {
+    test_rand_vec4(v);
+    d1 = glm_vec4_dot(v, v);
+    d2 = test_vec4_dot(v, v);
+
+    assert_true(fabsf(d1 - d2) <= 0.000009);
+  }
+}