diff --git a/src/arch/helperavx.h b/src/arch/helperavx.h
index e33a0e78..effdfa4c 100644
--- a/src/arch/helperavx.h
+++ b/src/arch/helperavx.h
@@ -128,20 +128,20 @@ static INLINE vdouble vreinterpret_vd_vm(vmask vm) { return _mm256_castsi256_pd(
 
 //
 
-static vint2 vloadu_vi2_p(int32_t *p) {
+static INLINE vint2 vloadu_vi2_p(int32_t *p) {
   vint2 r;
   r.x = _mm_loadu_si128((__m128i *) p     );
   r.y = _mm_loadu_si128((__m128i *)(p + 4));
   return r;
 }
 
-static void vstoreu_v_p_vi2(int32_t *p, vint2 v) {
+static INLINE void vstoreu_v_p_vi2(int32_t *p, vint2 v) {
   _mm_storeu_si128((__m128i *) p     , v.x);
   _mm_storeu_si128((__m128i *)(p + 4), v.y);  
 }
 
-static vint vloadu_vi_p(int32_t *p) { return _mm_loadu_si128((__m128i *)p); }
-static void vstoreu_v_p_vi(int32_t *p, vint v) { _mm_storeu_si128((__m128i *)p, v); }
+static INLINE vint vloadu_vi_p(int32_t *p) { return _mm_loadu_si128((__m128i *)p); }
+static INLINE void vstoreu_v_p_vi(int32_t *p, vint v) { _mm_storeu_si128((__m128i *)p, v); }
 
 //
 
@@ -570,7 +570,7 @@ static INLINE void vsscatter2_v_p_i_i_vf(float *ptr, int offset, int step, vfloa
 
 //
 
-static vquad loadu_vq_p(void *p) {
+static INLINE vquad loadu_vq_p(void *p) {
   vquad vq;
   memcpy(&vq, p, VECTLENDP * 16);
   return vq;
diff --git a/src/arch/helperavx2.h b/src/arch/helperavx2.h
index 47fcdc60..81e1f027 100644
--- a/src/arch/helperavx2.h
+++ b/src/arch/helperavx2.h
@@ -114,10 +114,10 @@ static INLINE vdouble vreinterpret_vd_vm(vmask vm) { return _mm256_castsi256_pd(
 
 //
 
-static vint2 vloadu_vi2_p(int32_t *p) { return _mm256_loadu_si256((__m256i const *)p); }
-static void vstoreu_v_p_vi2(int32_t *p, vint2 v) { _mm256_storeu_si256((__m256i *)p, v); }
-static vint vloadu_vi_p(int32_t *p) { return _mm_loadu_si128((__m128i *)p); }
-static void vstoreu_v_p_vi(int32_t *p, vint v) { _mm_storeu_si128((__m128i *)p, v); }
+static INLINE vint2 vloadu_vi2_p(int32_t *p) { return _mm256_loadu_si256((__m256i const *)p); }
+static INLINE void vstoreu_v_p_vi2(int32_t *p, vint2 v) { _mm256_storeu_si256((__m256i *)p, v); }
+static INLINE vint vloadu_vi_p(int32_t *p) { return _mm_loadu_si128((__m128i *)p); }
+static INLINE void vstoreu_v_p_vi(int32_t *p, vint v) { _mm_storeu_si128((__m128i *)p, v); }
 
 //
 
@@ -440,7 +440,7 @@ static INLINE void vsscatter2_v_p_i_i_vf(float *ptr, int offset, int step, vfloa
 
 //
 
-static vquad loadu_vq_p(void *p) {
+static INLINE vquad loadu_vq_p(void *p) {
   vquad vq;
   memcpy(&vq, p, VECTLENDP * 16);
   return vq;
diff --git a/src/arch/helperavx2_128.h b/src/arch/helperavx2_128.h
index 5233db1b..be7c56a2 100644
--- a/src/arch/helperavx2_128.h
+++ b/src/arch/helperavx2_128.h
@@ -109,10 +109,10 @@ static INLINE vdouble vreinterpret_vd_vm(vmask vm) { return _mm_castsi128_pd(vm)
 
 //
 
-static vint2 vloadu_vi2_p(int32_t *p) { return _mm_loadu_si128((__m128i const *)p); }
-static void vstoreu_v_p_vi2(int32_t *p, vint2 v) { _mm_storeu_si128((__m128i *)p, v); }
-static vint vloadu_vi_p(int32_t *p) { return _mm_loadu_si128((__m128i *)p); }
-static void vstoreu_v_p_vi(int32_t *p, vint v) { _mm_storeu_si128((__m128i *)p, v); }
+static INLINE vint2 vloadu_vi2_p(int32_t *p) { return _mm_loadu_si128((__m128i const *)p); }
+static INLINE void vstoreu_v_p_vi2(int32_t *p, vint2 v) { _mm_storeu_si128((__m128i *)p, v); }
+static INLINE vint vloadu_vi_p(int32_t *p) { return _mm_loadu_si128((__m128i *)p); }
+static INLINE void vstoreu_v_p_vi(int32_t *p, vint v) { _mm_storeu_si128((__m128i *)p, v); }
 
 //
 
@@ -413,7 +413,7 @@ static INLINE void vsscatter2_v_p_i_i_vf(float *ptr, int offset, int step, vfloa
 
 //
 
-static vquad loadu_vq_p(void *p) {
+static INLINE vquad loadu_vq_p(void *p) {
   vquad vq = {
     vloadu_vi2_p((int32_t *)p),
     vloadu_vi2_p((int32_t *)((uint8_t *)p + sizeof(vmask)))
@@ -433,7 +433,7 @@ static INLINE vargquad cast_aq_vq(vquad vq) {
   return aq;
 }
 
-static void vstoreu_v_p_vq(void *p, vquad vq) {
+static INLINE void vstoreu_v_p_vq(void *p, vquad vq) {
   vstoreu_v_p_vi2((int32_t *)p, vcast_vi2_vm(vq.x));
   vstoreu_v_p_vi2((int32_t *)((uint8_t *)p + sizeof(vmask)), vcast_vi2_vm(vq.y));
 }
diff --git a/src/arch/helpersse2.h b/src/arch/helpersse2.h
index 833f5f9b..141603a5 100644
--- a/src/arch/helpersse2.h
+++ b/src/arch/helpersse2.h
@@ -131,11 +131,11 @@ static INLINE int vtestallones_i_vo64(vopmask g) { return _mm_movemask_epi8(g) =
 
 //
 
-static vint2 vloadu_vi2_p(int32_t *p) { return _mm_loadu_si128((__m128i *)p); }
-static void vstoreu_v_p_vi2(int32_t *p, vint2 v) { _mm_storeu_si128((__m128i *)p, v); }
+static INLINE vint2 vloadu_vi2_p(int32_t *p) { return _mm_loadu_si128((__m128i *)p); }
+static INLINE void vstoreu_v_p_vi2(int32_t *p, vint2 v) { _mm_storeu_si128((__m128i *)p, v); }
 
-static vint vloadu_vi_p(int32_t *p) { return _mm_loadu_si128((__m128i *)p); }
-static void vstoreu_v_p_vi(int32_t *p, vint v) { _mm_storeu_si128((__m128i *)p, v); }
+static INLINE vint vloadu_vi_p(int32_t *p) { return _mm_loadu_si128((__m128i *)p); }
+static INLINE void vstoreu_v_p_vi(int32_t *p, vint v) { _mm_storeu_si128((__m128i *)p, v); }
 
 //
 
@@ -466,7 +466,7 @@ static INLINE void vsscatter2_v_p_i_i_vf(float *ptr, int offset, int step, vfloa
 
 //
 
-static vquad loadu_vq_p(void *p) {
+static INLINE vquad loadu_vq_p(void *p) {
   vquad vq;
   memcpy(&vq, p, VECTLENDP * 16);
   return vq;
diff --git a/src/libm/sleefinline_header.h.org b/src/libm/sleefinline_header.h.org
index df0c8215..d0cb3617 100644
--- a/src/libm/sleefinline_header.h.org
+++ b/src/libm/sleefinline_header.h.org
@@ -20,7 +20,9 @@
 #pragma fp_contract (off)
 #endif
 
+#if !(defined(__GNUC__) && !defined (__clang__))
 #pragma STDC FP_CONTRACT OFF
+#endif
 
 #ifndef SLEEF_FP_ILOGB0
 #define SLEEF_FP_ILOGB0 ((int)0x80000000)