carray 3.0.0 → 3.0.1

data/ext/carray_kernels_binop.c

@@ -34,6 +34,7 @@ static inline VALUE BOOL2VAL (boolean8_t x) { return x ? Qtrue : Qfalse; }
 #endif
 
 #include "ca_op_powi.h"
+#include "ca_op_cmplx64.h"
 
 /* ===== binop pmax ============================================ */
 
@@ -3860,7 +3861,7 @@ ca_binop_mul_cmplx64_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = (*p1) * (*p2); }
+        { (*p3) = op_cmul_cmplx64((*p1), (*p2)); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -3874,13 +3875,13 @@ ca_binop_mul_cmplx64_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = (*p1) * (*p2); }
+      { (*p3) = op_cmul_cmplx64((*p1), (*p2)); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     cmplx64_t *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = (*p1) * (*p2); }
+      { (*p3) = op_cmul_cmplx64((*p1), (*p2)); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -4029,7 +4030,14 @@ ca_binop_div_int8_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+        { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int8_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -4043,13 +4051,27 @@ ca_binop_div_int8_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int8_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     int8_t *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int8_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -4119,7 +4141,14 @@ ca_binop_div_int16_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+        { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int16_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -4133,13 +4162,27 @@ ca_binop_div_int16_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int16_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     int16_t *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int16_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -4209,7 +4252,14 @@ ca_binop_div_int32_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+        { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int32_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -4223,13 +4273,27 @@ ca_binop_div_int32_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int32_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     int32_t *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int32_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -4299,7 +4363,14 @@ ca_binop_div_int64_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+        { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int64_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -4313,13 +4384,27 @@ ca_binop_div_int64_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int64_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     int64_t *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) / (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int64_t _q = (*p1) / (*p2);
+        if ( (*p1) % (*p2) != 0 && (((*p1) < 0) != ((*p2) < 0)) ) { _q -= 1; }
+        (*p3) = _q;
+      }
+     }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -4479,7 +4564,7 @@ ca_binop_div_cmplx64_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = (*p1) / (*p2); }
+        { (*p3) = op_cdiv_cmplx64((*p1), (*p2)); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -4493,13 +4578,13 @@ ca_binop_div_cmplx64_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = (*p1) / (*p2); }
+      { (*p3) = op_cdiv_cmplx64((*p1), (*p2)); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     cmplx64_t *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = (*p1) / (*p2); }
+      { (*p3) = op_cdiv_cmplx64((*p1), (*p2)); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -5177,7 +5262,7 @@ ca_binop_rcp_mul_cmplx64_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = (*p2) / (*p1); }
+        { (*p3) = op_cdiv_cmplx64((*p2), (*p1)); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -5191,13 +5276,13 @@ ca_binop_rcp_mul_cmplx64_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = (*p2) / (*p1); }
+      { (*p3) = op_cdiv_cmplx64((*p2), (*p1)); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     cmplx64_t *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = (*p2) / (*p1); }
+      { (*p3) = op_cdiv_cmplx64((*p2), (*p1)); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -5346,7 +5431,14 @@ ca_binop_mod_int8_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+        { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int8_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -5360,13 +5452,27 @@ ca_binop_mod_int8_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int8_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     int8_t *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int8_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -5436,7 +5542,14 @@ ca_binop_mod_int16_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+        { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int16_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -5450,13 +5563,27 @@ ca_binop_mod_int16_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int16_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     int16_t *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int16_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -5526,7 +5653,14 @@ ca_binop_mod_int32_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+        { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int32_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -5540,13 +5674,27 @@ ca_binop_mod_int32_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int32_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     int32_t *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int32_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -5616,7 +5764,14 @@ ca_binop_mod_int64_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+        { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int64_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -5630,13 +5785,27 @@ ca_binop_mod_int64_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int64_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     int64_t *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { if ((*p2)==0) {ca_zerodiv();}; (*p3) = (*p1) % (*p2); }
+      { 
+      if ((*p2)==0) {ca_zerodiv();};
+      {
+        int64_t _r = (*p1) % (*p2);
+        if ( _r != 0 && ((_r < 0) != ((*p2) < 0)) ) { _r += (*p2); }
+        (*p3) = _r;
+      }
+     }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -5706,7 +5875,16 @@ ca_binop_mod_float32_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = fmod(*p1, *p2); }
+        { 
+      {
+        float32_t _r = fmod((*p1), (*p2));
+        if ( _r != 0 ) {
+          if ( (_r < 0) != ((*p2) < 0) ) { _r += (*p2); }
+        }
+        else { _r = copysign((float32_t) 0, (*p2)); }
+        (*p3) = _r;
+      }
+     }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -5720,13 +5898,31 @@ ca_binop_mod_float32_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = fmod(*p1, *p2); }
+      { 
+      {
+        float32_t _r = fmod((*p1), (*p2));
+        if ( _r != 0 ) {
+          if ( (_r < 0) != ((*p2) < 0) ) { _r += (*p2); }
+        }
+        else { _r = copysign((float32_t) 0, (*p2)); }
+        (*p3) = _r;
+      }
+     }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     float *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = fmod(*p1, *p2); }
+      { 
+      {
+        float32_t _r = fmod((*p1), (*p2));
+        if ( _r != 0 ) {
+          if ( (_r < 0) != ((*p2) < 0) ) { _r += (*p2); }
+        }
+        else { _r = copysign((float32_t) 0, (*p2)); }
+        (*p3) = _r;
+      }
+     }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -5751,7 +5947,16 @@ ca_binop_mod_float64_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = fmod(*p1, *p2); }
+        { 
+      {
+        float64_t _r = fmod((*p1), (*p2));
+        if ( _r != 0 ) {
+          if ( (_r < 0) != ((*p2) < 0) ) { _r += (*p2); }
+        }
+        else { _r = copysign((float64_t) 0, (*p2)); }
+        (*p3) = _r;
+      }
+     }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -5765,13 +5970,31 @@ ca_binop_mod_float64_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = fmod(*p1, *p2); }
+      { 
+      {
+        float64_t _r = fmod((*p1), (*p2));
+        if ( _r != 0 ) {
+          if ( (_r < 0) != ((*p2) < 0) ) { _r += (*p2); }
+        }
+        else { _r = copysign((float64_t) 0, (*p2)); }
+        (*p3) = _r;
+      }
+     }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     double *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = fmod(*p1, *p2); }
+      { 
+      {
+        float64_t _r = fmod((*p1), (*p2));
+        if ( _r != 0 ) {
+          if ( (_r < 0) != ((*p2) < 0) ) { _r += (*p2); }
+        }
+        else { _r = copysign((float64_t) 0, (*p2)); }
+        (*p3) = _r;
+      }
+     }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -5854,10 +6077,10 @@ VALUE rb_ca_mod (VALUE self, VALUE other)
 VALUE rb_ca_mod_bang (VALUE self, VALUE other)
 { return rb_ca_call_binop_bang(self, other, ca_binop_mod); }
 
-/* ===== binop reminder ============================================ */
+/* ===== binop fmod ============================================ */
 
 static void
-ca_binop_reminder_int8_t (ca_size_t n, boolean8_t *m,
+ca_binop_fmod_int8_t (ca_size_t n, boolean8_t *m,
                                  char * __restrict__ ptr1, ca_size_t i1,
                                  char * __restrict__ ptr2, ca_size_t i2,
                                  char * __restrict__ ptr3, ca_size_t i3)
@@ -5902,7 +6125,7 @@ ca_binop_reminder_int8_t (ca_size_t n, boolean8_t *m,
 }
 
 static void
-ca_binop_reminder_uint8_t (ca_size_t n, boolean8_t *m,
+ca_binop_fmod_uint8_t (ca_size_t n, boolean8_t *m,
                                  char * __restrict__ ptr1, ca_size_t i1,
                                  char * __restrict__ ptr2, ca_size_t i2,
                                  char * __restrict__ ptr3, ca_size_t i3)
@@ -5947,7 +6170,7 @@ ca_binop_reminder_uint8_t (ca_size_t n, boolean8_t *m,
 }
 
 static void
-ca_binop_reminder_int16_t (ca_size_t n, boolean8_t *m,
+ca_binop_fmod_int16_t (ca_size_t n, boolean8_t *m,
                                  char * __restrict__ ptr1, ca_size_t i1,
                                  char * __restrict__ ptr2, ca_size_t i2,
                                  char * __restrict__ ptr3, ca_size_t i3)
@@ -5992,7 +6215,7 @@ ca_binop_reminder_int16_t (ca_size_t n, boolean8_t *m,
 }
 
 static void
-ca_binop_reminder_uint16_t (ca_size_t n, boolean8_t *m,
+ca_binop_fmod_uint16_t (ca_size_t n, boolean8_t *m,
                                  char * __restrict__ ptr1, ca_size_t i1,
                                  char * __restrict__ ptr2, ca_size_t i2,
                                  char * __restrict__ ptr3, ca_size_t i3)
@@ -6037,7 +6260,7 @@ ca_binop_reminder_uint16_t (ca_size_t n, boolean8_t *m,
 }
 
 static void
-ca_binop_reminder_int32_t (ca_size_t n, boolean8_t *m,
+ca_binop_fmod_int32_t (ca_size_t n, boolean8_t *m,
                                  char * __restrict__ ptr1, ca_size_t i1,
                                  char * __restrict__ ptr2, ca_size_t i2,
                                  char * __restrict__ ptr3, ca_size_t i3)
@@ -6082,7 +6305,7 @@ ca_binop_reminder_int32_t (ca_size_t n, boolean8_t *m,
 }
 
 static void
-ca_binop_reminder_uint32_t (ca_size_t n, boolean8_t *m,
+ca_binop_fmod_uint32_t (ca_size_t n, boolean8_t *m,
                                  char * __restrict__ ptr1, ca_size_t i1,
                                  char * __restrict__ ptr2, ca_size_t i2,
                                  char * __restrict__ ptr3, ca_size_t i3)
@@ -6127,7 +6350,7 @@ ca_binop_reminder_uint32_t (ca_size_t n, boolean8_t *m,
 }
 
 static void
-ca_binop_reminder_int64_t (ca_size_t n, boolean8_t *m,
+ca_binop_fmod_int64_t (ca_size_t n, boolean8_t *m,
                                  char * __restrict__ ptr1, ca_size_t i1,
                                  char * __restrict__ ptr2, ca_size_t i2,
                                  char * __restrict__ ptr3, ca_size_t i3)
@@ -6172,7 +6395,7 @@ ca_binop_reminder_int64_t (ca_size_t n, boolean8_t *m,
 }
 
 static void
-ca_binop_reminder_uint64_t (ca_size_t n, boolean8_t *m,
+ca_binop_fmod_uint64_t (ca_size_t n, boolean8_t *m,
                                  char * __restrict__ ptr1, ca_size_t i1,
                                  char * __restrict__ ptr2, ca_size_t i2,
                                  char * __restrict__ ptr3, ca_size_t i3)
@@ -6217,7 +6440,7 @@ ca_binop_reminder_uint64_t (ca_size_t n, boolean8_t *m,
 }
 
 static void
-ca_binop_reminder_float32_t (ca_size_t n, boolean8_t *m,
+ca_binop_fmod_float32_t (ca_size_t n, boolean8_t *m,
                                  char * __restrict__ ptr1, ca_size_t i1,
                                  char * __restrict__ ptr2, ca_size_t i2,
                                  char * __restrict__ ptr3, ca_size_t i3)
@@ -6235,7 +6458,7 @@ ca_binop_reminder_float32_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = remainder(*p1, *p2); }
+        { (*p3) = fmodf((*p1), (*p2)); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -6249,20 +6472,20 @@ ca_binop_reminder_float32_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = remainder(*p1, *p2); }
+      { (*p3) = fmodf((*p1), (*p2)); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     float *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = remainder(*p1, *p2); }
+      { (*p3) = fmodf((*p1), (*p2)); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
 }
 
 static void
-ca_binop_reminder_float64_t (ca_size_t n, boolean8_t *m,
+ca_binop_fmod_float64_t (ca_size_t n, boolean8_t *m,
                                  char * __restrict__ ptr1, ca_size_t i1,
                                  char * __restrict__ ptr2, ca_size_t i2,
                                  char * __restrict__ ptr3, ca_size_t i3)
@@ -6280,7 +6503,7 @@ ca_binop_reminder_float64_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = remainder(*p1, *p2); }
+        { (*p3) = fmod((*p1), (*p2)); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -6294,20 +6517,20 @@ ca_binop_reminder_float64_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = remainder(*p1, *p2); }
+      { (*p3) = fmod((*p1), (*p2)); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     double *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = remainder(*p1, *p2); }
+      { (*p3) = fmod((*p1), (*p2)); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
 }
 
 static void
-ca_binop_reminder_VALUE (ca_size_t n, boolean8_t *m,
+ca_binop_fmod_VALUE (ca_size_t n, boolean8_t *m,
                                  char * __restrict__ ptr1, ca_size_t i1,
                                  char * __restrict__ ptr2, ca_size_t i2,
                                  char * __restrict__ ptr3, ca_size_t i3)
@@ -6325,7 +6548,7 @@ ca_binop_reminder_VALUE (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = rb_funcall((*p1), rb_intern("%"), 1, (*p2)); }
+        { (*p3) = rb_funcall((*p1), rb_intern("remainder"), 1, (*p2)); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -6339,49 +6562,49 @@ ca_binop_reminder_VALUE (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = rb_funcall((*p1), rb_intern("%"), 1, (*p2)); }
+      { (*p3) = rb_funcall((*p1), rb_intern("remainder"), 1, (*p2)); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     VALUE *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = rb_funcall((*p1), rb_intern("%"), 1, (*p2)); }
+      { (*p3) = rb_funcall((*p1), rb_intern("remainder"), 1, (*p2)); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
 }
 
 ca_binop_func_t
-ca_binop_reminder[CA_NTYPE] = {
+ca_binop_fmod[CA_NTYPE] = {
   ca_binop_not_implement,
   ca_binop_not_implement,
-  ca_binop_reminder_int8_t,
-  ca_binop_reminder_uint8_t,
-  ca_binop_reminder_int16_t,
-  ca_binop_reminder_uint16_t,
-  ca_binop_reminder_int32_t,
-  ca_binop_reminder_uint32_t,
-  ca_binop_reminder_int64_t,
-  ca_binop_reminder_uint64_t,
-  ca_binop_reminder_float32_t,
-  ca_binop_reminder_float64_t,
+  ca_binop_fmod_int8_t,
+  ca_binop_fmod_uint8_t,
+  ca_binop_fmod_int16_t,
+  ca_binop_fmod_uint16_t,
+  ca_binop_fmod_int32_t,
+  ca_binop_fmod_uint32_t,
+  ca_binop_fmod_int64_t,
+  ca_binop_fmod_uint64_t,
+  ca_binop_fmod_float32_t,
+  ca_binop_fmod_float64_t,
   ca_binop_not_implement,
   ca_binop_not_implement,
   ca_binop_not_implement,
   ca_binop_not_implement,
-  ca_binop_reminder_VALUE,
+  ca_binop_fmod_VALUE,
 };
 
-VALUE rb_ca_reminder (VALUE self, VALUE other)
+VALUE rb_ca_fmod (VALUE self, VALUE other)
 {
   if ( ! rb_ca_test_castable(other) ) {
-    return rb_ca_binop_pass_to_other(self, other, rb_intern("reminder"));
+    return rb_ca_binop_pass_to_other(self, other, rb_intern("fmod"));
   }
-  return rb_ca_call_binop(self, other, ca_binop_reminder);
+  return rb_ca_call_binop(self, other, ca_binop_fmod);
 }
 
-VALUE rb_ca_reminder_bang (VALUE self, VALUE other)
-{ return rb_ca_call_binop_bang(self, other, ca_binop_reminder); }
+VALUE rb_ca_fmod_bang (VALUE self, VALUE other)
+{ return rb_ca_call_binop_bang(self, other, ca_binop_fmod); }
 
 /* ===== binop bit_and_i ============================================ */
 
@@ -9466,7 +9689,7 @@ ca_binop_power_float32_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = pow((*p1), (*p2)); }
+        { (*p3) = powf((*p1), (*p2)); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -9480,13 +9703,13 @@ ca_binop_power_float32_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = pow((*p1), (*p2)); }
+      { (*p3) = powf((*p1), (*p2)); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     float *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = pow((*p1), (*p2)); }
+      { (*p3) = powf((*p1), (*p2)); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -9725,7 +9948,7 @@ ca_binop_copysign_float32_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = copysign((*p1), (*p2)); }
+        { (*p3) = copysignf((*p1), (*p2)); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -9739,13 +9962,13 @@ ca_binop_copysign_float32_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = copysign((*p1), (*p2)); }
+      { (*p3) = copysignf((*p1), (*p2)); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     float *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = copysign((*p1), (*p2)); }
+      { (*p3) = copysignf((*p1), (*p2)); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -9924,7 +10147,7 @@ ca_binop_nextafter_float32_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = nextafter((*p1), (*p2)); }
+        { (*p3) = nextafterf((*p1), (*p2)); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -9938,13 +10161,13 @@ ca_binop_nextafter_float32_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = nextafter((*p1), (*p2)); }
+      { (*p3) = nextafterf((*p1), (*p2)); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     float *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = nextafter((*p1), (*p2)); }
+      { (*p3) = nextafterf((*p1), (*p2)); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -10102,205 +10325,6 @@ VALUE rb_ca_nextafter (VALUE self, VALUE other)
 VALUE rb_ca_nextafter_bang (VALUE self, VALUE other)
 { return rb_ca_call_binop_bang(self, other, ca_binop_nextafter); }
 
-/* ===== binop fmod ============================================ */
-
-static void
-ca_binop_fmod_float32_t (ca_size_t n, boolean8_t *m,
-                                 char * __restrict__ ptr1, ca_size_t i1,
-                                 char * __restrict__ ptr2, ca_size_t i2,
-                                 char * __restrict__ ptr3, ca_size_t i3)
-{
-  float * __restrict__ q1 = (float *) ptr1;
-  float * __restrict__ q2 = (float *) ptr2;
-  float * __restrict__ q3 = (float *) ptr3;
-  ca_size_t k;
-  if ( m ) {
-    float *p1 = q1, *p2 = q2, *p3 = q3;
-    boolean8_t *pm;
-    for (k=0; k<n; k++) {
-      pm = m + k;
-      if ( ! *pm ) {
-        p1 = q1 + k*i1;
-        p2 = q2 + k*i2;
-        p3 = q3 + k*i3;
-        { (*p3) = fmod((*p1), (*p2)); }
-      }
-    }
-  } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
-    /* contig branch: SIMD-friendly form.  Index-form `q[k]` exposes
-       the unit-stride access to the vectorizer; `__restrict__` on
-       q1/q2/q3 promises no aliasing so the omp simd pragma can
-       license lane-parallel loads/stores.  Element-wise binop has
-       no reassoc, so ε-close numeric contract is unaffected. */
-    float *p1, *p2, *p3;
-    (void)p1; (void)p2; (void)p3;
-    #pragma omp simd
-    for (k=0; k<n; k++) {
-      p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = fmod((*p1), (*p2)); }
-    }
-  } else {
-    /* strided branch: numo-shape `p += s` increment form. */
-    float *p1 = q1, *p2 = q2, *p3 = q3;
-    for (k=0; k<n; k++) {
-      { (*p3) = fmod((*p1), (*p2)); }
-      p1 += i1; p2 += i2; p3 += i3;
-    }
-  }
-}
-
-static void
-ca_binop_fmod_float64_t (ca_size_t n, boolean8_t *m,
-                                 char * __restrict__ ptr1, ca_size_t i1,
-                                 char * __restrict__ ptr2, ca_size_t i2,
-                                 char * __restrict__ ptr3, ca_size_t i3)
-{
-  double * __restrict__ q1 = (double *) ptr1;
-  double * __restrict__ q2 = (double *) ptr2;
-  double * __restrict__ q3 = (double *) ptr3;
-  ca_size_t k;
-  if ( m ) {
-    double *p1 = q1, *p2 = q2, *p3 = q3;
-    boolean8_t *pm;
-    for (k=0; k<n; k++) {
-      pm = m + k;
-      if ( ! *pm ) {
-        p1 = q1 + k*i1;
-        p2 = q2 + k*i2;
-        p3 = q3 + k*i3;
-        { (*p3) = fmod((*p1), (*p2)); }
-      }
-    }
-  } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
-    /* contig branch: SIMD-friendly form.  Index-form `q[k]` exposes
-       the unit-stride access to the vectorizer; `__restrict__` on
-       q1/q2/q3 promises no aliasing so the omp simd pragma can
-       license lane-parallel loads/stores.  Element-wise binop has
-       no reassoc, so ε-close numeric contract is unaffected. */
-    double *p1, *p2, *p3;
-    (void)p1; (void)p2; (void)p3;
-    #pragma omp simd
-    for (k=0; k<n; k++) {
-      p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = fmod((*p1), (*p2)); }
-    }
-  } else {
-    /* strided branch: numo-shape `p += s` increment form. */
-    double *p1 = q1, *p2 = q2, *p3 = q3;
-    for (k=0; k<n; k++) {
-      { (*p3) = fmod((*p1), (*p2)); }
-      p1 += i1; p2 += i2; p3 += i3;
-    }
-  }
-}
-
-static void
-ca_binop_fmod_VALUE (ca_size_t n, boolean8_t *m,
-                                 char * __restrict__ ptr1, ca_size_t i1,
-                                 char * __restrict__ ptr2, ca_size_t i2,
-                                 char * __restrict__ ptr3, ca_size_t i3)
-{
-  VALUE * __restrict__ q1 = (VALUE *) ptr1;
-  VALUE * __restrict__ q2 = (VALUE *) ptr2;
-  VALUE * __restrict__ q3 = (VALUE *) ptr3;
-  ca_size_t k;
-  if ( m ) {
-    VALUE *p1 = q1, *p2 = q2, *p3 = q3;
-    boolean8_t *pm;
-    for (k=0; k<n; k++) {
-      pm = m + k;
-      if ( ! *pm ) {
-        p1 = q1 + k*i1;
-        p2 = q2 + k*i2;
-        p3 = q3 + k*i3;
-        { {
-  VALUE _l = (*p1);
-  VALUE _r = (*p2);
-  if ((RB_FLOAT_TYPE_P(_l) || RB_INTEGER_TYPE_P(_l) || RB_TYPE_P(_l, T_RATIONAL)) &&
-      (RB_FLOAT_TYPE_P(_r) || RB_INTEGER_TYPE_P(_r) || RB_TYPE_P(_r, T_RATIONAL))) {
-    (*p3) = DBL2NUM(fmod(NUM2DBL(_l), NUM2DBL(_r)));
-  } else {
-    (*p3) = rb_funcall(_l, rb_intern("fmod"), 1, _r);
-  }
-}
- }
-      }
-    }
-  } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
-    /* contig branch: SIMD-friendly form.  Index-form `q[k]` exposes
-       the unit-stride access to the vectorizer; `__restrict__` on
-       q1/q2/q3 promises no aliasing so the omp simd pragma can
-       license lane-parallel loads/stores.  Element-wise binop has
-       no reassoc, so ε-close numeric contract is unaffected. */
-    VALUE *p1, *p2, *p3;
-    (void)p1; (void)p2; (void)p3;
-    #pragma omp simd
-    for (k=0; k<n; k++) {
-      p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { {
-  VALUE _l = (*p1);
-  VALUE _r = (*p2);
-  if ((RB_FLOAT_TYPE_P(_l) || RB_INTEGER_TYPE_P(_l) || RB_TYPE_P(_l, T_RATIONAL)) &&
-      (RB_FLOAT_TYPE_P(_r) || RB_INTEGER_TYPE_P(_r) || RB_TYPE_P(_r, T_RATIONAL))) {
-    (*p3) = DBL2NUM(fmod(NUM2DBL(_l), NUM2DBL(_r)));
-  } else {
-    (*p3) = rb_funcall(_l, rb_intern("fmod"), 1, _r);
-  }
-}
- }
-    }
-  } else {
-    /* strided branch: numo-shape `p += s` increment form. */
-    VALUE *p1 = q1, *p2 = q2, *p3 = q3;
-    for (k=0; k<n; k++) {
-      { {
-  VALUE _l = (*p1);
-  VALUE _r = (*p2);
-  if ((RB_FLOAT_TYPE_P(_l) || RB_INTEGER_TYPE_P(_l) || RB_TYPE_P(_l, T_RATIONAL)) &&
-      (RB_FLOAT_TYPE_P(_r) || RB_INTEGER_TYPE_P(_r) || RB_TYPE_P(_r, T_RATIONAL))) {
-    (*p3) = DBL2NUM(fmod(NUM2DBL(_l), NUM2DBL(_r)));
-  } else {
-    (*p3) = rb_funcall(_l, rb_intern("fmod"), 1, _r);
-  }
-}
- }
-      p1 += i1; p2 += i2; p3 += i3;
-    }
-  }
-}
-
-ca_binop_func_t
-ca_binop_fmod[CA_NTYPE] = {
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_fmod_float32_t,
-  ca_binop_fmod_float64_t,
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_not_implement,
-  ca_binop_fmod_VALUE,
-};
-
-VALUE rb_ca_fmod (VALUE self, VALUE other)
-{
-  if ( ! rb_ca_test_castable(other) ) {
-    return rb_ca_binop_pass_to_other(self, other, rb_intern("fmod"));
-  }
-  return rb_ca_call_binop(self, other, ca_binop_fmod);
-}
-
-VALUE rb_ca_fmod_bang (VALUE self, VALUE other)
-{ return rb_ca_call_binop_bang(self, other, ca_binop_fmod); }
-
 /* ===== binop atan2 ============================================ */
 
 static void
@@ -10322,7 +10346,7 @@ ca_binop_atan2_float32_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = atan2((*p1), (*p2)); }
+        { (*p3) = atan2f((*p1), (*p2)); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -10336,13 +10360,13 @@ ca_binop_atan2_float32_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = atan2((*p1), (*p2)); }
+      { (*p3) = atan2f((*p1), (*p2)); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     float *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = atan2((*p1), (*p2)); }
+      { (*p3) = atan2f((*p1), (*p2)); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -10521,7 +10545,7 @@ ca_binop_hypot_float32_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = hypot((*p1), (*p2)); }
+        { (*p3) = hypotf((*p1), (*p2)); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -10535,13 +10559,13 @@ ca_binop_hypot_float32_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = hypot((*p1), (*p2)); }
+      { (*p3) = hypotf((*p1), (*p2)); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     float *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = hypot((*p1), (*p2)); }
+      { (*p3) = hypotf((*p1), (*p2)); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -10720,7 +10744,7 @@ ca_binop_logaddexp_float32_t (ca_size_t n, boolean8_t *m,
         p1 = q1 + k*i1;
         p2 = q2 + k*i2;
         p3 = q3 + k*i3;
-        { (*p3) = fmax((*p1), (*p2)) + log1p(exp(-fabs((*p1) - (*p2)))); }
+        { (*p3) = fmaxf((*p1), (*p2)) + log1pf(expf(-fabsf((*p1) - (*p2)))); }
       }
     }
   } else if ( i1 == 1 && i2 == 1 && i3 == 1 ) {
@@ -10734,13 +10758,13 @@ ca_binop_logaddexp_float32_t (ca_size_t n, boolean8_t *m,
     #pragma omp simd
     for (k=0; k<n; k++) {
       p1 = q1 + k; p2 = q2 + k; p3 = q3 + k;
-      { (*p3) = fmax((*p1), (*p2)) + log1p(exp(-fabs((*p1) - (*p2)))); }
+      { (*p3) = fmaxf((*p1), (*p2)) + log1pf(expf(-fabsf((*p1) - (*p2)))); }
     }
   } else {
     /* strided branch: numo-shape `p += s` increment form. */
     float *p1 = q1, *p2 = q2, *p3 = q3;
     for (k=0; k<n; k++) {
-      { (*p3) = fmax((*p1), (*p2)) + log1p(exp(-fabs((*p1) - (*p2)))); }
+      { (*p3) = fmaxf((*p1), (*p2)) + log1pf(expf(-fabsf((*p1) - (*p2)))); }
       p1 += i1; p2 += i2; p3 += i3;
     }
   }
@@ -10932,8 +10956,8 @@ Init_carray_kernels_binop (void)
   rb_define_method(rb_cCArray, "rcp_mul!", rb_ca_rcp_mul_bang, 1);
   rb_define_method(rb_cCArray, "%", rb_ca_mod, 1);
   rb_define_method(rb_cCArray, "mod!", rb_ca_mod_bang, 1);
-  rb_define_method(rb_cCArray, "reminder", rb_ca_reminder, 1);
-  rb_define_method(rb_cCArray, "reminder!", rb_ca_reminder_bang, 1);
+  rb_define_method(rb_cCArray, "fmod", rb_ca_fmod, 1);
+  rb_define_method(rb_cCArray, "fmod!", rb_ca_fmod_bang, 1);
   rb_define_method(rb_cCArray, "&", rb_ca_bit_and_i, 1);
   rb_define_method(rb_cCArray, "bit_and_i!", rb_ca_bit_and_i_bang, 1);
   rb_define_method(rb_cCArray, "|", rb_ca_bit_or_i, 1);
@@ -10968,8 +10992,6 @@ Init_carray_kernels_binop (void)
   rb_define_method(rb_cCArray, "copysign!", rb_ca_copysign_bang, 1);
   rb_define_method(rb_cCArray, "nextafter", rb_ca_nextafter, 1);
   rb_define_method(rb_cCArray, "nextafter!", rb_ca_nextafter_bang, 1);
-  rb_define_method(rb_cCArray, "fmod", rb_ca_fmod, 1);
-  rb_define_method(rb_cCArray, "fmod!", rb_ca_fmod_bang, 1);
   rb_define_method(rb_cCArray, "atan2", rb_ca_atan2, 1);
   rb_define_method(rb_cCArray, "atan2!", rb_ca_atan2_bang, 1);
   rb_define_method(rb_cCArray, "hypot", rb_ca_hypot, 1);