gcc/testsuite/gcc.dg/vect/slp-33.c

   1 /* { dg-require-effective-target vect_int } */
   2 /* { dg-require-effective-target vect_float } */
   3
   4 #include <stdarg.h>
   5 #include "tree-vect.h"
   6
   7 #define N 8
   8
   9 int
  10 main1 ()
  11 {
  12   int i;
  13   unsigned int out[N*8], a0, a1, a2, a3, a4, a5, a6, a7, b1, b0, b2, b3, b4, b5, b6, b7;
  14   unsigned int in[N*8] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63};
  15   float out2[N*8];
  16
  17   /* SLP with unrolling by 4.  */
  18   for (i = 0; i < N; i++)
  19     {
  20       a0 = in[i*7] + 5;
  21       a1 = in[i*7 + 1] + 6;
  22       a2 = in[i*7 + 2] + 7;
  23       a3 = in[i*7 + 3] + 8;
  24       a4 = in[i*7 + 4] + 9;
  25       a5 = in[i*7 + 5] + 10;
  26       a6 = in[i*7 + 6] + 11;
  27
  28       b0 = a0 * 3;
  29       b1 = a1 * 2;
  30       b2 = a2 * 12;
  31       b3 = a3 * 5;
  32       b4 = a4 * 8;
  33       b5 = a5 * 4;
  34       b6 = a6 * 3;
  35
  36       out[i*7] = b0 - 2;
  37       out[i*7 + 1] = b1 - 3;
  38       out[i*7 + 2] = b2 - 2;
  39       out[i*7 + 3] = b3 - 1;
  40       out[i*7 + 4] = b4 - 8;
  41       out[i*7 + 5] = b5 - 7;
  42       out[i*7 + 6] = b6 - 3;
  43     }
  44
  45   /* check results:  */
  46 #pragma GCC novector
  47   for (i = 0; i < N; i++)
  48     {
  49       if (out[i*7] !=  (in[i*7] + 5) * 3 - 2
  50          || out[i*7 + 1] != (in[i*7 + 1] + 6) * 2 - 3
  51          || out[i*7 + 2] != (in[i*7 + 2] + 7) * 12 - 2
  52          || out[i*7 + 3] != (in[i*7 + 3] + 8) * 5 - 1
  53          || out[i*7 + 4] != (in[i*7 + 4] + 9) * 8 - 8
  54          || out[i*7 + 5] != (in[i*7 + 5] + 10) * 4 - 7
  55          || out[i*7 + 6] != (in[i*7 + 6] + 11) * 3 - 3)
  56         abort ();
  57     }
  58
  59   /* SLP with unrolling by 4.  */
  60   for (i = 0; i < N*2; i++)
  61     {
  62       out[i*3] = (in[i*3] + 2) * 3;
  63       out[i*3 + 1] = (in[i*3 + 1] + 2) * 7;
  64       out[i*3 + 2] = (in[i*3 + 2] + 7) * 3;
  65     }
  66
  67   /* check results:  */
  68 #pragma GCC novector
  69   for (i = 0; i < N*2; i++)
  70     {
  71       if (out[i*3] !=  (in[i*3] + 2) * 3
  72          || out[i*3 + 1] != (in[i*3 + 1] + 2) * 7
  73          || out[i*3 + 2] != (in[i*3 + 2] + 7) * 3)
  74         abort ();
  75     }
  76
  77   /* SLP with unrolling by 4.  */
  78   for (i = 0; i < N*2; i++)
  79     {
  80       out2[i*3] = (float) (in[i*3] * 2 + 5) ;
  81       out2[i*3 + 1] = (float) (in[i*3 + 1] * 3 + 7);
  82       out2[i*3 + 2] = (float) (in[i*3 + 2] * 5 + 4);
  83     }
  84
  85   /* check results:  */
  86 #pragma GCC novector
  87   for (i = 0; i < N*2; i++)
  88     {
  89       if (out2[i*3] !=  (float) (in[i*3] * 2 + 5)
  90          || out2[i*3 + 1] != (float) (in[i*3 + 1] * 3 + 7)
  91          || out2[i*3 + 2] != (float) (in[i*3 + 2] * 5 + 4))
  92         abort ();
  93     }
  94
  95
  96   return 0;
  97 }
  98
  99 int main (void)
 100 {
 101   check_vect ();
 102
 103   main1 ();
 104
 105   return 0;
 106 }
 107
 108 /* { dg-final { scan-tree-dump-times "vectorized 3 loops" 1 "vect"  {target {vect_uintfloat_cvt && vect_int_mult} } } } */
 109 /* { dg-final { scan-tree-dump-times "vectorized 2 loops" 1 "vect"  {target {{! { vect_uintfloat_cvt}} && vect_int_mult} } } } */
 110 /* { dg-final { scan-tree-dump-times "vectorized 0 loops" 1 "vect"  {target {{! { vect_uintfloat_cvt}} && {! {vect_int_mult}}} } } } */
 111 /* { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 3 "vect" {target {vect_uintfloat_cvt && vect_int_mult} xfail { vect_variable_length && vect_load_lanes } } } } */
 112 /* { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 2 "vect"  {target {{! { vect_uintfloat_cvt}} && vect_int_mult} } } } */
 113 /* { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 0 "vect"  {target {{! { vect_uintfloat_cvt}} && {! {vect_int_mult}}} } } } */
 114