diff --git a/lib/include/srslte/phy/utils/simd.h b/lib/include/srslte/phy/utils/simd.h
index e7820c307..0c378591e 100644
--- a/lib/include/srslte/phy/utils/simd.h
+++ b/lib/include/srslte/phy/utils/simd.h
@@ -310,10 +310,10 @@ static inline simd_f_t srslte_simd_f_addsub(simd_f_t a, simd_f_t b) {
   return _mm_addsub_ps(a, b);
 #else /* LV_HAVE_SSE */
 #ifdef HAVE_NEON // CURRENTLY USES GENERIC IMPLEMENTATION FOR NEON
- float* a_ptr = &a;
- float* b_ptr = &b;
+ float* a_ptr = (float*) &a;
+ float* b_ptr = (float*) &b;
  simd_f_t ret;
- float* c_ptr = &ret;
+ float* c_ptr = (float*) &ret;
  for(int i = 0; i<4;i++){
    if(i%2==0){
      c_ptr[i] = a_ptr[i] - b_ptr[i];
@@ -1028,12 +1028,12 @@ static inline simd_i_t srslte_simd_i_select(simd_i_t a, simd_i_t b, simd_sel_t s
 #else /* LV_HAVE_SSE */
 #ifdef HAVE_NEON // CURRENTLY USES GENERIC IMPLEMENTATION FOR NEON
   
- int* a_ptr = &a;
- int* b_ptr = &b;
+ int* a_ptr = (int*) &a;
+ int* b_ptr = (int*) &b;
  simd_i_t ret;
- int* sel = &selector;
+ int* sel = (int*) &selector;
  
- int* c_ptr = &ret;
+ int* c_ptr = (int*) &ret;
  for(int i = 0;i<4;i++)
  { 
    if(sel[i] == -1){
@@ -1097,7 +1097,7 @@ static inline simd_s_t srslte_simd_s_loadu(int16_t *ptr) {
   return _mm256_loadu_si256((__m256i*) ptr);
 #else /* LV_HAVE_AVX2 */
 #ifdef LV_HAVE_SSE
-  return _mm_loadu_si128((__m128i*) ptr)
+  return _mm_loadu_si128((__m128i*) ptr);
 #else /* LV_HAVE_SSE */
 #ifdef HAVE_NEON
   return vld1q_s16(ptr);
diff --git a/lib/src/phy/utils/test/vector_test.c b/lib/src/phy/utils/test/vector_test.c
index 8d5b9f2d6..4ebed9862 100644
--- a/lib/src/phy/utils/test/vector_test.c
+++ b/lib/src/phy/utils/test/vector_test.c
@@ -63,12 +63,12 @@ bool verbose = false;
     strncpy(func_name, #X, 32);\
     CODE;\
     passed = (mse < MAX_MSE);\
-    printf("%32s (%5d) ... %7.1f MSamp/s ... %3s Passed\n", func_name, block_size, \
-    (double) block_size*NOF_REPETITIONS/ *timing, passed?"":"Not");\
+    printf("%32s (%5d) ... %7.1f MSamp/s ... %3s Passed (%.6f)\n", func_name, block_size, \
+    (double) block_size*NOF_REPETITIONS/ *timing, passed?"":"Not", mse);\
     return passed;\
 }
 
-#define MALLOC(TYPE, NAME) TYPE *NAME = malloc(sizeof(TYPE)*block_size)
+#define MALLOC(TYPE, NAME) TYPE *NAME = srslte_vec_malloc(sizeof(TYPE)*block_size)
 
 
 static double elapsed_us(struct timeval *ts_start, struct timeval *ts_end) {
@@ -507,7 +507,7 @@ TEST(srslte_vec_abs_cf,
 
   for (int i = 0; i < block_size; i++) {
     gold = sqrtf(crealf(x[i]) * crealf(x[i]) + cimagf(x[i])*cimagf(x[i]));
-    mse += cabsf(gold - z[i]);
+    mse += cabsf(gold - z[i])/block_size;
   }
 
   free(x);
@@ -771,12 +771,27 @@ int main(int argc, char **argv) {
     size_count++;
   }
 
+  char fname[68];
+  FILE *f = NULL;
+  void * p = popen("(date +%g%m%d && hostname) | tr '\\r\\n' '__'", "r");
+  if (p) {
+    fgets(fname, 64, p);
+    strncpy(fname + strnlen(fname, 64) - 1, ".tsv", 4);
+    f = fopen(fname, "w");
+    if (f) printf("Saving benchmark results in '%s'\n", fname);
+  }
+  pclose(p);
+
+
   printf("\n");
   printf("%32s |", "Subroutine/MSps");
+  if (f) fprintf(f, "Subroutine/MSps Vs Vector size\t");
   for (int i = 0; i < size_count; i++) {
     printf(" %7d", sizes[i]);
+    if (f) fprintf(f, "%d\t", sizes[i]);
   }
   printf("  |\n");
+  if (f) fprintf(f, "\n");
 
   for (int j = 0; j < 32; j++) {
     printf("-");
@@ -789,12 +804,19 @@ int main(int argc, char **argv) {
 
   for (int i = 0; i < func_count; i++) {
     printf("%32s | ", func_names[i]);
+    if (f) fprintf(f, "%s\t", func_names[i]);
+
     for (int j = 0; j < size_count; j++) {
       printf(" %s%7.1f\x1b[0m", (passed[i][j])?"":"\x1B[31m", (double) NOF_REPETITIONS*(double)sizes[j]/timmings[i][j]);
+      if (f) fprintf(f, "%.1f\t", (double) NOF_REPETITIONS*(double)sizes[j]/timmings[i][j]);
+
       all_passed &= passed[i][j];
     }
     printf(" |\n");
+    if (f) fprintf(f, "\n");
   }
 
+  if (f) fclose(f);
+
   return (all_passed)?SRSLTE_SUCCESS:SRSLTE_ERROR;
 }