|
| 1 | +// NOTE: Assertions have been autogenerated by utils/update_mc_test_checks.py UTC_ARGS: --unique --version 5 |
| 2 | +// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1100 -mattr=-real-true16 -show-encoding %s | FileCheck %s |
| 3 | +// RUN: llvm-mc -triple=amdgcn -mcpu=gfx1200 -mattr=-real-true16 -show-encoding %s | FileCheck %s |
| 4 | + |
| 5 | +v_dot2_bf16_bf16 v5, v1, v2, 100.0 |
| 6 | +// CHECK: v_dot2_bf16_bf16 v5, v1, v2, 0x42c8 ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xfe,0x03,0xc8,0x42,0x00,0x00] |
| 7 | + |
| 8 | +v_dot2_bf16_bf16 v2, v0, 1.0, v2 |
| 9 | +// CHECK: v_dot2_bf16_bf16 v2, v0, 1.0, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe5,0x09,0x04] |
| 10 | + |
| 11 | +v_dot2_bf16_bf16 v2, 1.0, v0, v2 |
| 12 | +// CHECK: v_dot2_bf16_bf16 v2, 1.0, v0, v2 ; encoding: [0x02,0x00,0x67,0xd6,0xf2,0x00,0x0a,0x04] |
| 13 | + |
| 14 | +v_dot2_bf16_bf16 v5, v1, v2, 1.0 |
| 15 | +// CHECK: v_dot2_bf16_bf16 v5, v1, v2, 1.0 ; encoding: [0x05,0x00,0x67,0xd6,0x01,0x05,0xca,0x03] |
| 16 | + |
| 17 | +v_dot2_bf16_bf16 v2, v0, -1.0, v2 |
| 18 | +// CHECK: v_dot2_bf16_bf16 v2, v0, -1.0, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe7,0x09,0x04] |
| 19 | + |
| 20 | +v_dot2_bf16_bf16 v2, v0, 0.5, v2 |
| 21 | +// CHECK: v_dot2_bf16_bf16 v2, v0, 0.5, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe1,0x09,0x04] |
| 22 | + |
| 23 | +v_dot2_bf16_bf16 v2, v0, -0.5, v2 |
| 24 | +// CHECK: v_dot2_bf16_bf16 v2, v0, -0.5, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe3,0x09,0x04] |
| 25 | + |
| 26 | +v_dot2_bf16_bf16 v2, v0, 2.0, v2 |
| 27 | +// CHECK: v_dot2_bf16_bf16 v2, v0, 2.0, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xe9,0x09,0x04] |
| 28 | + |
| 29 | +v_dot2_bf16_bf16 v2, v0, -2.0, v2 |
| 30 | +// CHECK: v_dot2_bf16_bf16 v2, v0, -2.0, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xeb,0x09,0x04] |
| 31 | + |
| 32 | +v_dot2_bf16_bf16 v2, v0, 4.0, v2 |
| 33 | +// CHECK: v_dot2_bf16_bf16 v2, v0, 4.0, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xed,0x09,0x04] |
| 34 | + |
| 35 | +v_dot2_bf16_bf16 v2, v0, -4.0, v2 |
| 36 | +// CHECK: v_dot2_bf16_bf16 v2, v0, -4.0, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xef,0x09,0x04] |
| 37 | + |
| 38 | +// Check 1/(2*pi) rounded value and ideomatic fp32 0.15915494 value |
| 39 | +// which cannot be accurately represented in bf16. |
| 40 | + |
| 41 | +v_dot2_bf16_bf16 v2, v0, 0.158203125, v2 |
| 42 | +// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04] |
| 43 | + |
| 44 | +v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 |
| 45 | +// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04] |
| 46 | + |
| 47 | +v_dot2_bf16_bf16 v2, v0, 0x3e22, v2 |
| 48 | +// CHECK: v_dot2_bf16_bf16 v2, v0, 0.15915494, v2 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0xf1,0x09,0x04] |
| 49 | + |
| 50 | +v_dot2_bf16_bf16 v2, v0, v2, 0.15915494 |
| 51 | +// CHECK: v_dot2_bf16_bf16 v2, v0, v2, 0.15915494 ; encoding: [0x02,0x00,0x67,0xd6,0x00,0x05,0xe2,0x03] |
| 52 | + |
| 53 | +v_dot2_f32_bf16 v2, v1, 0, v2 |
| 54 | +// CHECK: v_dot2_f32_bf16 v2, v1, 0, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0x01,0x09,0x1c] |
| 55 | + |
| 56 | +v_dot2_f32_bf16 v2, v1, 0.5, v2 |
| 57 | +// CHECK: v_dot2_f32_bf16 v2, v1, 0.5, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe1,0x09,0x1c] |
| 58 | + |
| 59 | +v_dot2_f32_bf16 v2, v1, -0.5, v2 |
| 60 | +// CHECK: v_dot2_f32_bf16 v2, v1, -0.5, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe3,0x09,0x1c] |
| 61 | + |
| 62 | +v_dot2_f32_bf16 v2, v1, 1.0, v2 |
| 63 | +// CHECK: v_dot2_f32_bf16 v2, v1, 1.0, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe5,0x09,0x1c] |
| 64 | + |
| 65 | +v_dot2_f32_bf16 v2, v1, -1.0, v2 |
| 66 | +// CHECK: v_dot2_f32_bf16 v2, v1, -1.0, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe7,0x09,0x1c] |
| 67 | + |
| 68 | +v_dot2_f32_bf16 v2, v1, 2.0, v2 |
| 69 | +// CHECK: v_dot2_f32_bf16 v2, v1, 2.0, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xe9,0x09,0x1c] |
| 70 | + |
| 71 | +v_dot2_f32_bf16 v2, v1, -2.0, v2 |
| 72 | +// CHECK: v_dot2_f32_bf16 v2, v1, -2.0, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xeb,0x09,0x1c] |
| 73 | + |
| 74 | +v_dot2_f32_bf16 v2, v1, 4.0, v2 |
| 75 | +// CHECK: v_dot2_f32_bf16 v2, v1, 4.0, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xed,0x09,0x1c] |
| 76 | + |
| 77 | +v_dot2_f32_bf16 v2, v1, -4.0, v2 |
| 78 | +// CHECK: v_dot2_f32_bf16 v2, v1, -4.0, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xef,0x09,0x1c] |
| 79 | + |
| 80 | +v_dot2_f32_bf16 v2, v1, 0.15915494, v2 |
| 81 | +// CHECK: v_dot2_f32_bf16 v2, v1, 0.15915494, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xf1,0x09,0x1c] |
| 82 | + |
| 83 | +v_dot2_f32_bf16 v2, v1, 0x3e22, v2 |
| 84 | +// CHECK: v_dot2_f32_bf16 v2, v1, 0.15915494, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xf1,0x09,0x1c] |
| 85 | + |
| 86 | +v_dot2_f32_bf16 v2, 0.5, v1, v2 |
| 87 | +// CHECK: v_dot2_f32_bf16 v2, 0.5, v1, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0xf0,0x02,0x0a,0x1c] |
| 88 | + |
| 89 | +v_dot2_f32_bf16 v2, -0.5, v1, v2 |
| 90 | +// CHECK: v_dot2_f32_bf16 v2, -0.5, v1, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0xf1,0x02,0x0a,0x1c] |
| 91 | + |
| 92 | +v_dot2_f32_bf16 v2, 1.0, v1, v2 |
| 93 | +// CHECK: v_dot2_f32_bf16 v2, 1.0, v1, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0xf2,0x02,0x0a,0x1c] |
| 94 | + |
| 95 | +v_dot2_f32_bf16 v2, -1.0, v1, v2 |
| 96 | +// CHECK: v_dot2_f32_bf16 v2, -1.0, v1, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0xf3,0x02,0x0a,0x1c] |
| 97 | + |
| 98 | +v_dot2_f32_bf16 v2, 2.0, v1, v2 |
| 99 | +// CHECK: v_dot2_f32_bf16 v2, 2.0, v1, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0xf4,0x02,0x0a,0x1c] |
| 100 | + |
| 101 | +v_dot2_f32_bf16 v2, -2.0, v1, v2 |
| 102 | +// CHECK: v_dot2_f32_bf16 v2, -2.0, v1, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0xf5,0x02,0x0a,0x1c] |
| 103 | + |
| 104 | +v_dot2_f32_bf16 v2, 4.0, v1, v2 |
| 105 | +// CHECK: v_dot2_f32_bf16 v2, 4.0, v1, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0xf6,0x02,0x0a,0x1c] |
| 106 | + |
| 107 | +v_dot2_f32_bf16 v2, -4.0, v1, v2 |
| 108 | +// CHECK: v_dot2_f32_bf16 v2, -4.0, v1, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0xf7,0x02,0x0a,0x1c] |
| 109 | + |
| 110 | +v_dot2_f32_bf16 v2, 100.0, v1, v2 |
| 111 | +// CHECK: v_dot2_f32_bf16 v2, 0x42c8, v1, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0xff,0x02,0x0a,0x1c,0xc8,0x42,0x00,0x00] |
| 112 | + |
| 113 | +v_dot2_f32_bf16 v2, v1, 100.0, v2 |
| 114 | +// CHECK: v_dot2_f32_bf16 v2, v1, 0x42c8, v2 ; encoding: [0x02,0x40,0x1a,0xcc,0x01,0xff,0x09,0x1c,0xc8,0x42,0x00,0x00] |
0 commit comments