diff options
Diffstat (limited to 'volk')
9 files changed, 12 insertions, 11 deletions
diff --git a/volk/include/volk/volk_32f_s32f_power_32f_a.h b/volk/include/volk/volk_32f_s32f_power_32f_a.h index c4fa31bd1..09c905961 100644 --- a/volk/include/volk/volk_32f_s32f_power_32f_a.h +++ b/volk/include/volk/volk_32f_s32f_power_32f_a.h @@ -21,12 +21,12 @@ */ static inline void volk_32f_s32f_power_32f_a_sse4_1(float* cVector, const float* aVector, const float power, unsigned int num_points){ unsigned int number = 0; - const unsigned int quarterPoints = num_points / 4; float* cPtr = cVector; const float* aPtr = aVector; #ifdef LV_HAVE_LIB_SIMDMATH + const unsigned int quarterPoints = num_points / 4; __m128 vPower = _mm_set_ps1(power); __m128 zeroValue = _mm_setzero_ps(); __m128 signMask; @@ -78,12 +78,12 @@ static inline void volk_32f_s32f_power_32f_a_sse4_1(float* cVector, const float* */ static inline void volk_32f_s32f_power_32f_a_sse(float* cVector, const float* aVector, const float power, unsigned int num_points){ unsigned int number = 0; - const unsigned int quarterPoints = num_points / 4; float* cPtr = cVector; const float* aPtr = aVector; #ifdef LV_HAVE_LIB_SIMDMATH + const unsigned int quarterPoints = num_points / 4; __m128 vPower = _mm_set_ps1(power); __m128 zeroValue = _mm_setzero_ps(); __m128 signMask; diff --git a/volk/include/volk/volk_32fc_s32f_atan2_32f_a.h b/volk/include/volk/volk_32fc_s32f_atan2_32f_a.h index 9304b0c28..7bd001aa0 100644 --- a/volk/include/volk/volk_32fc_s32f_atan2_32f_a.h +++ b/volk/include/volk/volk_32fc_s32f_atan2_32f_a.h @@ -24,10 +24,10 @@ static inline void volk_32fc_s32f_atan2_32f_a_sse4_1(float* outputVector, const float* outPtr = outputVector; unsigned int number = 0; - const unsigned int quarterPoints = num_points / 4; const float invNormalizeFactor = 1.0 / normalizeFactor; #ifdef LV_HAVE_LIB_SIMDMATH + const unsigned int quarterPoints = num_points / 4; __m128 testVector = _mm_set_ps1(2*M_PI); __m128 correctVector = _mm_set_ps1(M_PI); __m128 vNormalizeFactor = _mm_set_ps1(invNormalizeFactor); @@ -86,10 +86,10 @@ static inline void volk_32fc_s32f_atan2_32f_a_sse(float* outputVector, const lv float* outPtr = outputVector; unsigned int number = 0; - const unsigned int quarterPoints = num_points / 4; const float invNormalizeFactor = 1.0 / normalizeFactor; #ifdef LV_HAVE_LIB_SIMDMATH + const unsigned int quarterPoints = num_points / 4; __m128 testVector = _mm_set_ps1(2*M_PI); __m128 correctVector = _mm_set_ps1(M_PI); __m128 vNormalizeFactor = _mm_set_ps1(invNormalizeFactor); diff --git a/volk/include/volk/volk_32fc_s32f_power_32fc_a.h b/volk/include/volk/volk_32fc_s32f_power_32fc_a.h index ec1d7167f..588b532b4 100644 --- a/volk/include/volk/volk_32fc_s32f_power_32fc_a.h +++ b/volk/include/volk/volk_32fc_s32f_power_32fc_a.h @@ -28,12 +28,12 @@ static inline lv_32fc_t __volk_s32fc_s32f_power_s32fc_a(const lv_32fc_t exp, con */ static inline void volk_32fc_s32f_power_32fc_a_sse(lv_32fc_t* cVector, const lv_32fc_t* aVector, const float power, unsigned int num_points){ unsigned int number = 0; - const unsigned int quarterPoints = num_points / 4; lv_32fc_t* cPtr = cVector; const lv_32fc_t* aPtr = aVector; #ifdef LV_HAVE_LIB_SIMDMATH + const unsigned int quarterPoints = num_points / 4; __m128 vPower = _mm_set_ps1(power); __m128 cplxValue1, cplxValue2, magnitude, phase, iValue, qValue; diff --git a/volk/include/volk/volk_32fc_x2_conjugate_dot_prod_32fc_u.h b/volk/include/volk/volk_32fc_x2_conjugate_dot_prod_32fc_u.h index 02faf86c2..3ae7208a8 100644 --- a/volk/include/volk/volk_32fc_x2_conjugate_dot_prod_32fc_u.h +++ b/volk/include/volk/volk_32fc_x2_conjugate_dot_prod_32fc_u.h @@ -66,7 +66,8 @@ static inline void volk_32fc_x2_conjugate_dot_prod_32fc_u_generic(lv_32fc_t* res static inline void volk_32fc_x2_conjugate_dot_prod_32fc_u_sse3(lv_32fc_t* result, const lv_32fc_t* input, const lv_32fc_t* taps, unsigned int num_bytes) { - __VOLK_ATTR_ALIGNED(16) static const uint32_t conjugator[4]= {0x00000000, 0x80000000, 0x00000000, 0x80000000}; + // Variable never used? + //__VOLK_ATTR_ALIGNED(16) static const uint32_t conjugator[4]= {0x00000000, 0x80000000, 0x00000000, 0x80000000}; union HalfMask { uint32_t intRep[4]; diff --git a/volk/orc/volk_16ic_magnitude_16i_a_orc_impl.orc b/volk/orc/volk_16ic_magnitude_16i_a_orc_impl.orc index fbaebc46d..3966526ed 100644 --- a/volk/orc/volk_16ic_magnitude_16i_a_orc_impl.orc +++ b/volk/orc/volk_16ic_magnitude_16i_a_orc_impl.orc @@ -17,7 +17,7 @@ x2 divf iqf, iqf, scalar x2 mulf prodiqf, iqf, iqf splitql qf, if, prodiqf addf sumf, if, qf -sqrtf rootf, sumf +sqrt, sqrtf rootf, sumf mulf rootf, rootf, scalar convfl rootl, rootf convlw dst, rootl diff --git a/volk/orc/volk_16sc_magnitude_32f_aligned16_orc_impl.orc b/volk/orc/volk_16sc_magnitude_32f_aligned16_orc_impl.orc index 66fef7d2e..d5c8212c3 100644 --- a/volk/orc/volk_16sc_magnitude_32f_aligned16_orc_impl.orc +++ b/volk/orc/volk_16sc_magnitude_32f_aligned16_orc_impl.orc @@ -22,4 +22,4 @@ divf imagf, imagf, scalar mulf realf, realf, realf mulf imagf, imagf, imagf addf sumf, realf, imagf -sqrtf dst, sumf +sqrt, sqrtf dst, sumf diff --git a/volk/orc/volk_32f_sqrt_32f_a_orc_impl.orc b/volk/orc/volk_32f_sqrt_32f_a_orc_impl.orc index f339b1122..719cca037 100644 --- a/volk/orc/volk_32f_sqrt_32f_a_orc_impl.orc +++ b/volk/orc/volk_32f_sqrt_32f_a_orc_impl.orc @@ -1,4 +1,4 @@ .function volk_32f_sqrt_32f_a_orc_impl .source 4 src .dest 4 dst -sqrtf dst, src +sqrt, sqrtf dst, src diff --git a/volk/orc/volk_32fc_magnitude_32f_a_orc_impl.orc b/volk/orc/volk_32fc_magnitude_32f_a_orc_impl.orc index 032ab2b1b..4fc0642fc 100644 --- a/volk/orc/volk_32fc_magnitude_32f_a_orc_impl.orc +++ b/volk/orc/volk_32fc_magnitude_32f_a_orc_impl.orc @@ -10,4 +10,4 @@ x2 mulf prodiqf, src, src splitql qf, if, prodiqf addf sumf, if, qf -sqrtf dst, sumf +sqrt, sqrtf dst, sumf diff --git a/volk/orc/volk_32fc_s32f_magnitude_16i_a_orc_impl.orc b/volk/orc/volk_32fc_s32f_magnitude_16i_a_orc_impl.orc index d3bf78935..ed89c3d18 100644 --- a/volk/orc/volk_32fc_s32f_magnitude_16i_a_orc_impl.orc +++ b/volk/orc/volk_32fc_s32f_magnitude_16i_a_orc_impl.orc @@ -14,7 +14,7 @@ x2 mulf prodiqf, src, src splitql qf, if, prodiqf addf sumf, if, qf -sqrtf rootf, sumf +sqrt, sqrtf rootf, sumf mulf rootf, rootf, scalar #cmpltf maskl, 32768.0, rootf #andl maskl, maskl, 0x80000000 |