Mike Klein | 267f507 | 2019-06-03 16:27:46 -0500 | [diff] [blame] | 1 | A8 over A8 |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 2 | 12 values: |
| 3 | v0 = load8 arg(0) |
| 4 | v1 = to_f32 v0 |
| 5 | v2 = mul_f32 v1 3B808081 (0.0039215689) |
| 6 | v3 = load8 arg(1) |
| 7 | v4 = to_f32 v3 |
| 8 | v5 = mul_f32 v4 3B808081 (0.0039215689) |
| 9 | ↑ v6 = splat 3F800000 (1) |
| 10 | v7 = sub_f32 v6 v2 |
| 11 | v8 = mad_f32 v5 v7 v2 |
| 12 | v9 = mul_f32 v8 437F0000 (255) |
| 13 | v10 = round v9 |
| 14 | store8 arg(1) v10 |
Mike Klein | 22ea7e9 | 2019-06-10 12:05:48 -0500 | [diff] [blame] | 15 | |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 16 | 4 registers, 12 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 17 | 0 r0 = splat 3F800000 (1) |
Mike Klein | 754bad3 | 2019-06-05 10:47:46 -0500 | [diff] [blame] | 18 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 19 | 1 r1 = load8 arg(0) |
| 20 | 2 r1 = to_f32 r1 |
| 21 | 3 r1 = mul_f32 r1 3B808081 (0.0039215689) |
| 22 | 4 r2 = load8 arg(1) |
| 23 | 5 r2 = to_f32 r2 |
| 24 | 6 r2 = mul_f32 r2 3B808081 (0.0039215689) |
| 25 | 7 r3 = sub_f32 r0 r1 |
| 26 | 8 r1 = mad_f32 r2 r3 r1 |
| 27 | 9 r1 = mul_f32 r1 437F0000 (255) |
| 28 | 10 r1 = round r1 |
| 29 | 11 store8 arg(1) r1 |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 30 | |
| 31 | A8 over G8 |
| 32 | 17 values: |
| 33 | v0 = load8 arg(1) |
| 34 | v1 = to_f32 v0 |
| 35 | v2 = mul_f32 v1 3B808081 (0.0039215689) |
| 36 | v3 = load8 arg(0) |
| 37 | v4 = to_f32 v3 |
| 38 | v5 = mul_f32 v4 3B808081 (0.0039215689) |
| 39 | ↑ v6 = splat 3F800000 (1) |
| 40 | v7 = sub_f32 v6 v5 |
| 41 | v8 = mul_f32 v2 v7 |
| 42 | ↑ v9 = splat 3E59B3D0 (0.21259999) |
| 43 | ↑ v10 = splat 3F371759 (0.71520001) |
| 44 | v11 = mul_f32 v8 3D93DD98 (0.0722) |
| 45 | v12 = mad_f32 v8 v10 v11 |
| 46 | v13 = mad_f32 v8 v9 v12 |
| 47 | v14 = mul_f32 v13 437F0000 (255) |
| 48 | v15 = round v14 |
| 49 | store8 arg(1) v15 |
| 50 | |
| 51 | 5 registers, 17 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 52 | 0 r0 = splat 3F800000 (1) |
| 53 | 1 r1 = splat 3E59B3D0 (0.21259999) |
| 54 | 2 r2 = splat 3F371759 (0.71520001) |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 55 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 56 | 3 r3 = load8 arg(1) |
| 57 | 4 r3 = to_f32 r3 |
| 58 | 5 r3 = mul_f32 r3 3B808081 (0.0039215689) |
| 59 | 6 r4 = load8 arg(0) |
| 60 | 7 r4 = to_f32 r4 |
| 61 | 8 r4 = mul_f32 r4 3B808081 (0.0039215689) |
| 62 | 9 r4 = sub_f32 r0 r4 |
| 63 | 10 r4 = mul_f32 r3 r4 |
| 64 | 11 r3 = mul_f32 r4 3D93DD98 (0.0722) |
| 65 | 12 r3 = mad_f32 r4 r2 r3 |
| 66 | 13 r3 = mad_f32 r4 r1 r3 |
| 67 | 14 r3 = mul_f32 r3 437F0000 (255) |
| 68 | 15 r3 = round r3 |
| 69 | 16 store8 arg(1) r3 |
Mike Klein | 267f507 | 2019-06-03 16:27:46 -0500 | [diff] [blame] | 70 | |
Mike Klein | 754bad3 | 2019-06-05 10:47:46 -0500 | [diff] [blame] | 71 | A8 over RGBA_8888 |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 72 | 35 values: |
| 73 | v0 = load32 arg(1) |
| 74 | ↑ v1 = splat FF (3.5733111e-43) |
| 75 | v2 = extract v0 0 v1 |
| 76 | v3 = to_f32 v2 |
| 77 | v4 = mul_f32 v3 3B808081 (0.0039215689) |
| 78 | v5 = load8 arg(0) |
| 79 | v6 = to_f32 v5 |
| 80 | v7 = mul_f32 v6 3B808081 (0.0039215689) |
| 81 | ↑ v8 = splat 3F800000 (1) |
| 82 | v9 = sub_f32 v8 v7 |
| 83 | v10 = mul_f32 v4 v9 |
| 84 | v11 = mul_f32 v10 437F0000 (255) |
| 85 | v12 = round v11 |
| 86 | v13 = extract v0 8 v1 |
| 87 | v14 = to_f32 v13 |
| 88 | v15 = mul_f32 v14 3B808081 (0.0039215689) |
| 89 | v16 = mul_f32 v15 v9 |
| 90 | v17 = mul_f32 v16 437F0000 (255) |
| 91 | v18 = round v17 |
| 92 | v19 = pack v12 v18 8 |
| 93 | v20 = extract v0 16 v1 |
| 94 | v21 = to_f32 v20 |
| 95 | v22 = mul_f32 v21 3B808081 (0.0039215689) |
| 96 | v23 = mul_f32 v22 v9 |
| 97 | v24 = mul_f32 v23 437F0000 (255) |
| 98 | v25 = round v24 |
| 99 | v26 = extract v0 24 v1 |
| 100 | v27 = to_f32 v26 |
| 101 | v28 = mul_f32 v27 3B808081 (0.0039215689) |
| 102 | v29 = mad_f32 v28 v9 v7 |
| 103 | v30 = mul_f32 v29 437F0000 (255) |
| 104 | v31 = round v30 |
| 105 | v32 = pack v25 v31 8 |
| 106 | v33 = pack v19 v32 16 |
| 107 | store32 arg(1) v33 |
Mike Klein | 22ea7e9 | 2019-06-10 12:05:48 -0500 | [diff] [blame] | 108 | |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 109 | 7 registers, 35 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 110 | 0 r0 = splat FF (3.5733111e-43) |
| 111 | 1 r1 = splat 3F800000 (1) |
Mike Klein | 754bad3 | 2019-06-05 10:47:46 -0500 | [diff] [blame] | 112 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 113 | 2 r2 = load32 arg(1) |
| 114 | 3 r3 = extract r2 0 r0 |
| 115 | 4 r3 = to_f32 r3 |
| 116 | 5 r3 = mul_f32 r3 3B808081 (0.0039215689) |
| 117 | 6 r4 = load8 arg(0) |
| 118 | 7 r4 = to_f32 r4 |
| 119 | 8 r4 = mul_f32 r4 3B808081 (0.0039215689) |
| 120 | 9 r5 = sub_f32 r1 r4 |
| 121 | 10 r3 = mul_f32 r3 r5 |
| 122 | 11 r3 = mul_f32 r3 437F0000 (255) |
| 123 | 12 r3 = round r3 |
| 124 | 13 r6 = extract r2 8 r0 |
| 125 | 14 r6 = to_f32 r6 |
| 126 | 15 r6 = mul_f32 r6 3B808081 (0.0039215689) |
| 127 | 16 r6 = mul_f32 r6 r5 |
| 128 | 17 r6 = mul_f32 r6 437F0000 (255) |
| 129 | 18 r6 = round r6 |
| 130 | 19 r6 = pack r3 r6 8 |
| 131 | 20 r3 = extract r2 16 r0 |
| 132 | 21 r3 = to_f32 r3 |
| 133 | 22 r3 = mul_f32 r3 3B808081 (0.0039215689) |
| 134 | 23 r3 = mul_f32 r3 r5 |
| 135 | 24 r3 = mul_f32 r3 437F0000 (255) |
| 136 | 25 r3 = round r3 |
| 137 | 26 r2 = extract r2 24 r0 |
| 138 | 27 r2 = to_f32 r2 |
| 139 | 28 r2 = mul_f32 r2 3B808081 (0.0039215689) |
| 140 | 29 r4 = mad_f32 r2 r5 r4 |
| 141 | 30 r4 = mul_f32 r4 437F0000 (255) |
| 142 | 31 r4 = round r4 |
| 143 | 32 r4 = pack r3 r4 8 |
| 144 | 33 r4 = pack r6 r4 16 |
| 145 | 34 store32 arg(1) r4 |
Mike Klein | 754bad3 | 2019-06-05 10:47:46 -0500 | [diff] [blame] | 146 | |
| 147 | G8 over A8 |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 148 | 9 values: |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 149 | ↑ v0 = splat 3F800000 (1) |
Mike Klein | 1cb0599 | 2019-11-21 11:46:58 -0600 | [diff] [blame] | 150 | ↑ v1 = splat 0 (0) |
| 151 | v2 = load8 arg(1) |
| 152 | v3 = to_f32 v2 |
| 153 | v4 = mul_f32 v3 3B808081 (0.0039215689) |
| 154 | v5 = mad_f32 v4 v1 v0 |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 155 | v6 = mul_f32 v5 437F0000 (255) |
| 156 | v7 = round v6 |
| 157 | store8 arg(1) v7 |
Mike Klein | 22ea7e9 | 2019-06-10 12:05:48 -0500 | [diff] [blame] | 158 | |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 159 | 3 registers, 9 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 160 | 0 r0 = splat 3F800000 (1) |
| 161 | 1 r1 = splat 0 (0) |
Mike Klein | 754bad3 | 2019-06-05 10:47:46 -0500 | [diff] [blame] | 162 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 163 | 2 r2 = load8 arg(1) |
| 164 | 3 r2 = to_f32 r2 |
| 165 | 4 r2 = mul_f32 r2 3B808081 (0.0039215689) |
| 166 | 5 r2 = mad_f32 r2 r1 r0 |
| 167 | 6 r2 = mul_f32 r2 437F0000 (255) |
| 168 | 7 r2 = round r2 |
| 169 | 8 store8 arg(1) r2 |
Mike Klein | 754bad3 | 2019-06-05 10:47:46 -0500 | [diff] [blame] | 170 | |
| 171 | G8 over G8 |
Mike Klein | 1cb0599 | 2019-11-21 11:46:58 -0600 | [diff] [blame] | 172 | 16 values: |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 173 | v0 = load8 arg(0) |
| 174 | v1 = to_f32 v0 |
| 175 | v2 = mul_f32 v1 3B808081 (0.0039215689) |
| 176 | v3 = load8 arg(1) |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 177 | v4 = to_f32 v3 |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 178 | v5 = mul_f32 v4 3B808081 (0.0039215689) |
Mike Klein | 1cb0599 | 2019-11-21 11:46:58 -0600 | [diff] [blame] | 179 | ↑ v6 = splat 0 (0) |
| 180 | v7 = mad_f32 v5 v6 v2 |
| 181 | ↑ v8 = splat 3E59B3D0 (0.21259999) |
| 182 | ↑ v9 = splat 3F371759 (0.71520001) |
| 183 | v10 = mul_f32 v7 3D93DD98 (0.0722) |
| 184 | v11 = mad_f32 v7 v9 v10 |
| 185 | v12 = mad_f32 v7 v8 v11 |
| 186 | v13 = mul_f32 v12 437F0000 (255) |
| 187 | v14 = round v13 |
| 188 | store8 arg(1) v14 |
Mike Klein | 22ea7e9 | 2019-06-10 12:05:48 -0500 | [diff] [blame] | 189 | |
Mike Klein | 1cb0599 | 2019-11-21 11:46:58 -0600 | [diff] [blame] | 190 | 5 registers, 16 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 191 | 0 r0 = splat 0 (0) |
| 192 | 1 r1 = splat 3E59B3D0 (0.21259999) |
| 193 | 2 r2 = splat 3F371759 (0.71520001) |
Mike Klein | 754bad3 | 2019-06-05 10:47:46 -0500 | [diff] [blame] | 194 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 195 | 3 r3 = load8 arg(0) |
| 196 | 4 r3 = to_f32 r3 |
| 197 | 5 r3 = mul_f32 r3 3B808081 (0.0039215689) |
| 198 | 6 r4 = load8 arg(1) |
| 199 | 7 r4 = to_f32 r4 |
| 200 | 8 r4 = mul_f32 r4 3B808081 (0.0039215689) |
| 201 | 9 r3 = mad_f32 r4 r0 r3 |
| 202 | 10 r4 = mul_f32 r3 3D93DD98 (0.0722) |
| 203 | 11 r4 = mad_f32 r3 r2 r4 |
| 204 | 12 r4 = mad_f32 r3 r1 r4 |
| 205 | 13 r4 = mul_f32 r4 437F0000 (255) |
| 206 | 14 r4 = round r4 |
| 207 | 15 store8 arg(1) r4 |
Mike Klein | 754bad3 | 2019-06-05 10:47:46 -0500 | [diff] [blame] | 208 | |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 209 | G8 over RGBA_8888 |
| 210 | 35 values: |
| 211 | v0 = load8 arg(0) |
| 212 | v1 = to_f32 v0 |
| 213 | v2 = mul_f32 v1 3B808081 (0.0039215689) |
| 214 | v3 = load32 arg(1) |
| 215 | ↑ v4 = splat FF (3.5733111e-43) |
| 216 | v5 = extract v3 0 v4 |
| 217 | v6 = to_f32 v5 |
| 218 | v7 = mul_f32 v6 3B808081 (0.0039215689) |
Mike Klein | 1cb0599 | 2019-11-21 11:46:58 -0600 | [diff] [blame] | 219 | ↑ v8 = splat 0 (0) |
| 220 | v9 = mad_f32 v7 v8 v2 |
| 221 | v10 = mul_f32 v9 437F0000 (255) |
| 222 | v11 = round v10 |
| 223 | v12 = extract v3 8 v4 |
| 224 | v13 = to_f32 v12 |
| 225 | v14 = mul_f32 v13 3B808081 (0.0039215689) |
| 226 | v15 = mad_f32 v14 v8 v2 |
| 227 | v16 = mul_f32 v15 437F0000 (255) |
| 228 | v17 = round v16 |
| 229 | v18 = pack v11 v17 8 |
| 230 | v19 = extract v3 16 v4 |
| 231 | v20 = to_f32 v19 |
| 232 | v21 = mul_f32 v20 3B808081 (0.0039215689) |
| 233 | v22 = mad_f32 v21 v8 v2 |
| 234 | v23 = mul_f32 v22 437F0000 (255) |
| 235 | v24 = round v23 |
| 236 | ↑ v25 = splat 3F800000 (1) |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 237 | v26 = extract v3 24 v4 |
| 238 | v27 = to_f32 v26 |
| 239 | v28 = mul_f32 v27 3B808081 (0.0039215689) |
Mike Klein | 1cb0599 | 2019-11-21 11:46:58 -0600 | [diff] [blame] | 240 | v29 = mad_f32 v28 v8 v25 |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 241 | v30 = mul_f32 v29 437F0000 (255) |
| 242 | v31 = round v30 |
Mike Klein | 1cb0599 | 2019-11-21 11:46:58 -0600 | [diff] [blame] | 243 | v32 = pack v24 v31 8 |
| 244 | v33 = pack v18 v32 16 |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 245 | store32 arg(1) v33 |
Mike Klein | 22ea7e9 | 2019-06-10 12:05:48 -0500 | [diff] [blame] | 246 | |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 247 | 7 registers, 35 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 248 | 0 r0 = splat FF (3.5733111e-43) |
| 249 | 1 r1 = splat 0 (0) |
| 250 | 2 r2 = splat 3F800000 (1) |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 251 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 252 | 3 r3 = load8 arg(0) |
| 253 | 4 r3 = to_f32 r3 |
| 254 | 5 r3 = mul_f32 r3 3B808081 (0.0039215689) |
| 255 | 6 r4 = load32 arg(1) |
| 256 | 7 r5 = extract r4 0 r0 |
| 257 | 8 r5 = to_f32 r5 |
| 258 | 9 r5 = mul_f32 r5 3B808081 (0.0039215689) |
| 259 | 10 r5 = mad_f32 r5 r1 r3 |
| 260 | 11 r5 = mul_f32 r5 437F0000 (255) |
| 261 | 12 r5 = round r5 |
| 262 | 13 r6 = extract r4 8 r0 |
| 263 | 14 r6 = to_f32 r6 |
| 264 | 15 r6 = mul_f32 r6 3B808081 (0.0039215689) |
| 265 | 16 r6 = mad_f32 r6 r1 r3 |
| 266 | 17 r6 = mul_f32 r6 437F0000 (255) |
| 267 | 18 r6 = round r6 |
| 268 | 19 r6 = pack r5 r6 8 |
| 269 | 20 r5 = extract r4 16 r0 |
| 270 | 21 r5 = to_f32 r5 |
| 271 | 22 r5 = mul_f32 r5 3B808081 (0.0039215689) |
| 272 | 23 r3 = mad_f32 r5 r1 r3 |
| 273 | 24 r3 = mul_f32 r3 437F0000 (255) |
| 274 | 25 r3 = round r3 |
| 275 | 26 r4 = extract r4 24 r0 |
| 276 | 27 r4 = to_f32 r4 |
| 277 | 28 r4 = mul_f32 r4 3B808081 (0.0039215689) |
| 278 | 29 r4 = mad_f32 r4 r1 r2 |
| 279 | 30 r4 = mul_f32 r4 437F0000 (255) |
| 280 | 31 r4 = round r4 |
| 281 | 32 r4 = pack r3 r4 8 |
| 282 | 33 r4 = pack r6 r4 16 |
| 283 | 34 store32 arg(1) r4 |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 284 | |
| 285 | RGBA_8888 over A8 |
| 286 | 14 values: |
| 287 | v0 = load32 arg(0) |
| 288 | ↑ v1 = splat FF (3.5733111e-43) |
| 289 | v2 = extract v0 24 v1 |
| 290 | v3 = to_f32 v2 |
| 291 | v4 = mul_f32 v3 3B808081 (0.0039215689) |
| 292 | v5 = load8 arg(1) |
| 293 | v6 = to_f32 v5 |
| 294 | v7 = mul_f32 v6 3B808081 (0.0039215689) |
| 295 | ↑ v8 = splat 3F800000 (1) |
| 296 | v9 = sub_f32 v8 v4 |
| 297 | v10 = mad_f32 v7 v9 v4 |
| 298 | v11 = mul_f32 v10 437F0000 (255) |
| 299 | v12 = round v11 |
| 300 | store8 arg(1) v12 |
| 301 | |
| 302 | 5 registers, 14 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 303 | 0 r0 = splat FF (3.5733111e-43) |
| 304 | 1 r1 = splat 3F800000 (1) |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 305 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 306 | 2 r2 = load32 arg(0) |
| 307 | 3 r2 = extract r2 24 r0 |
| 308 | 4 r2 = to_f32 r2 |
| 309 | 5 r2 = mul_f32 r2 3B808081 (0.0039215689) |
| 310 | 6 r3 = load8 arg(1) |
| 311 | 7 r3 = to_f32 r3 |
| 312 | 8 r3 = mul_f32 r3 3B808081 (0.0039215689) |
| 313 | 9 r4 = sub_f32 r1 r2 |
| 314 | 10 r2 = mad_f32 r3 r4 r2 |
| 315 | 11 r2 = mul_f32 r2 437F0000 (255) |
| 316 | 12 r2 = round r2 |
| 317 | 13 store8 arg(1) r2 |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 318 | |
| 319 | RGBA_8888 over G8 |
| 320 | 30 values: |
| 321 | v0 = load32 arg(0) |
| 322 | ↑ v1 = splat FF (3.5733111e-43) |
| 323 | v2 = extract v0 0 v1 |
| 324 | v3 = to_f32 v2 |
| 325 | v4 = mul_f32 v3 3B808081 (0.0039215689) |
| 326 | v5 = load8 arg(1) |
| 327 | v6 = to_f32 v5 |
| 328 | v7 = mul_f32 v6 3B808081 (0.0039215689) |
| 329 | v8 = extract v0 24 v1 |
| 330 | v9 = to_f32 v8 |
| 331 | v10 = mul_f32 v9 3B808081 (0.0039215689) |
| 332 | ↑ v11 = splat 3F800000 (1) |
| 333 | v12 = sub_f32 v11 v10 |
| 334 | v13 = mad_f32 v7 v12 v4 |
| 335 | ↑ v14 = splat 3E59B3D0 (0.21259999) |
| 336 | v15 = extract v0 8 v1 |
| 337 | v16 = to_f32 v15 |
| 338 | v17 = mul_f32 v16 3B808081 (0.0039215689) |
| 339 | v18 = mad_f32 v7 v12 v17 |
| 340 | ↑ v19 = splat 3F371759 (0.71520001) |
| 341 | v20 = extract v0 16 v1 |
| 342 | v21 = to_f32 v20 |
| 343 | v22 = mul_f32 v21 3B808081 (0.0039215689) |
| 344 | v23 = mad_f32 v7 v12 v22 |
| 345 | v24 = mul_f32 v23 3D93DD98 (0.0722) |
| 346 | v25 = mad_f32 v18 v19 v24 |
| 347 | v26 = mad_f32 v13 v14 v25 |
| 348 | v27 = mul_f32 v26 437F0000 (255) |
| 349 | v28 = round v27 |
| 350 | store8 arg(1) v28 |
| 351 | |
| 352 | 9 registers, 30 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 353 | 0 r0 = splat FF (3.5733111e-43) |
| 354 | 1 r1 = splat 3F800000 (1) |
| 355 | 2 r2 = splat 3E59B3D0 (0.21259999) |
| 356 | 3 r3 = splat 3F371759 (0.71520001) |
Mike Klein | a630732 | 2019-06-07 15:44:26 -0500 | [diff] [blame] | 357 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 358 | 4 r4 = load32 arg(0) |
| 359 | 5 r5 = extract r4 0 r0 |
| 360 | 6 r5 = to_f32 r5 |
| 361 | 7 r5 = mul_f32 r5 3B808081 (0.0039215689) |
| 362 | 8 r6 = load8 arg(1) |
| 363 | 9 r6 = to_f32 r6 |
| 364 | 10 r6 = mul_f32 r6 3B808081 (0.0039215689) |
| 365 | 11 r7 = extract r4 24 r0 |
| 366 | 12 r7 = to_f32 r7 |
| 367 | 13 r7 = mul_f32 r7 3B808081 (0.0039215689) |
| 368 | 14 r7 = sub_f32 r1 r7 |
| 369 | 15 r5 = mad_f32 r6 r7 r5 |
| 370 | 16 r8 = extract r4 8 r0 |
| 371 | 17 r8 = to_f32 r8 |
| 372 | 18 r8 = mul_f32 r8 3B808081 (0.0039215689) |
| 373 | 19 r8 = mad_f32 r6 r7 r8 |
| 374 | 20 r4 = extract r4 16 r0 |
| 375 | 21 r4 = to_f32 r4 |
| 376 | 22 r4 = mul_f32 r4 3B808081 (0.0039215689) |
| 377 | 23 r4 = mad_f32 r6 r7 r4 |
| 378 | 24 r4 = mul_f32 r4 3D93DD98 (0.0722) |
| 379 | 25 r4 = mad_f32 r8 r3 r4 |
| 380 | 26 r4 = mad_f32 r5 r2 r4 |
| 381 | 27 r4 = mul_f32 r4 437F0000 (255) |
| 382 | 28 r4 = round r4 |
| 383 | 29 store8 arg(1) r4 |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 384 | |
| 385 | RGBA_8888 over RGBA_8888 |
| 386 | 45 values: |
| 387 | v0 = load32 arg(0) |
| 388 | ↑ v1 = splat FF (3.5733111e-43) |
| 389 | v2 = extract v0 0 v1 |
| 390 | v3 = to_f32 v2 |
| 391 | v4 = mul_f32 v3 3B808081 (0.0039215689) |
| 392 | v5 = load32 arg(1) |
| 393 | v6 = extract v5 0 v1 |
| 394 | v7 = to_f32 v6 |
| 395 | v8 = mul_f32 v7 3B808081 (0.0039215689) |
| 396 | v9 = extract v0 24 v1 |
| 397 | v10 = to_f32 v9 |
| 398 | v11 = mul_f32 v10 3B808081 (0.0039215689) |
| 399 | ↑ v12 = splat 3F800000 (1) |
| 400 | v13 = sub_f32 v12 v11 |
| 401 | v14 = mad_f32 v8 v13 v4 |
| 402 | v15 = mul_f32 v14 437F0000 (255) |
| 403 | v16 = round v15 |
| 404 | v17 = extract v0 8 v1 |
| 405 | v18 = to_f32 v17 |
| 406 | v19 = mul_f32 v18 3B808081 (0.0039215689) |
| 407 | v20 = extract v5 8 v1 |
| 408 | v21 = to_f32 v20 |
| 409 | v22 = mul_f32 v21 3B808081 (0.0039215689) |
| 410 | v23 = mad_f32 v22 v13 v19 |
| 411 | v24 = mul_f32 v23 437F0000 (255) |
| 412 | v25 = round v24 |
| 413 | v26 = pack v16 v25 8 |
| 414 | v27 = extract v0 16 v1 |
| 415 | v28 = to_f32 v27 |
| 416 | v29 = mul_f32 v28 3B808081 (0.0039215689) |
| 417 | v30 = extract v5 16 v1 |
| 418 | v31 = to_f32 v30 |
| 419 | v32 = mul_f32 v31 3B808081 (0.0039215689) |
| 420 | v33 = mad_f32 v32 v13 v29 |
| 421 | v34 = mul_f32 v33 437F0000 (255) |
| 422 | v35 = round v34 |
| 423 | v36 = extract v5 24 v1 |
| 424 | v37 = to_f32 v36 |
| 425 | v38 = mul_f32 v37 3B808081 (0.0039215689) |
| 426 | v39 = mad_f32 v38 v13 v11 |
| 427 | v40 = mul_f32 v39 437F0000 (255) |
| 428 | v41 = round v40 |
| 429 | v42 = pack v35 v41 8 |
| 430 | v43 = pack v26 v42 16 |
| 431 | store32 arg(1) v43 |
| 432 | |
| 433 | 9 registers, 45 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 434 | 0 r0 = splat FF (3.5733111e-43) |
| 435 | 1 r1 = splat 3F800000 (1) |
Mike Klein | 8c1e0ef | 2019-11-12 09:07:23 -0600 | [diff] [blame] | 436 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 437 | 2 r2 = load32 arg(0) |
| 438 | 3 r3 = extract r2 0 r0 |
| 439 | 4 r3 = to_f32 r3 |
| 440 | 5 r3 = mul_f32 r3 3B808081 (0.0039215689) |
| 441 | 6 r4 = load32 arg(1) |
| 442 | 7 r5 = extract r4 0 r0 |
| 443 | 8 r5 = to_f32 r5 |
| 444 | 9 r5 = mul_f32 r5 3B808081 (0.0039215689) |
| 445 | 10 r6 = extract r2 24 r0 |
| 446 | 11 r6 = to_f32 r6 |
| 447 | 12 r6 = mul_f32 r6 3B808081 (0.0039215689) |
| 448 | 13 r7 = sub_f32 r1 r6 |
| 449 | 14 r3 = mad_f32 r5 r7 r3 |
| 450 | 15 r3 = mul_f32 r3 437F0000 (255) |
| 451 | 16 r3 = round r3 |
| 452 | 17 r5 = extract r2 8 r0 |
| 453 | 18 r5 = to_f32 r5 |
| 454 | 19 r5 = mul_f32 r5 3B808081 (0.0039215689) |
| 455 | 20 r8 = extract r4 8 r0 |
| 456 | 21 r8 = to_f32 r8 |
| 457 | 22 r8 = mul_f32 r8 3B808081 (0.0039215689) |
| 458 | 23 r5 = mad_f32 r8 r7 r5 |
| 459 | 24 r5 = mul_f32 r5 437F0000 (255) |
| 460 | 25 r5 = round r5 |
| 461 | 26 r5 = pack r3 r5 8 |
| 462 | 27 r2 = extract r2 16 r0 |
| 463 | 28 r2 = to_f32 r2 |
| 464 | 29 r2 = mul_f32 r2 3B808081 (0.0039215689) |
| 465 | 30 r3 = extract r4 16 r0 |
| 466 | 31 r3 = to_f32 r3 |
| 467 | 32 r3 = mul_f32 r3 3B808081 (0.0039215689) |
| 468 | 33 r2 = mad_f32 r3 r7 r2 |
| 469 | 34 r2 = mul_f32 r2 437F0000 (255) |
| 470 | 35 r2 = round r2 |
| 471 | 36 r4 = extract r4 24 r0 |
| 472 | 37 r4 = to_f32 r4 |
| 473 | 38 r4 = mul_f32 r4 3B808081 (0.0039215689) |
| 474 | 39 r6 = mad_f32 r4 r7 r6 |
| 475 | 40 r6 = mul_f32 r6 437F0000 (255) |
| 476 | 41 r6 = round r6 |
| 477 | 42 r6 = pack r2 r6 8 |
| 478 | 43 r6 = pack r5 r6 16 |
| 479 | 44 store32 arg(1) r6 |
Mike Klein | 267f507 | 2019-06-03 16:27:46 -0500 | [diff] [blame] | 480 | |
Mike Klein | 397fc88 | 2019-06-20 11:37:10 -0500 | [diff] [blame] | 481 | I32 (Naive) 8888 over 8888 |
Mike Klein | aab45b5 | 2019-07-02 15:39:23 -0500 | [diff] [blame] | 482 | 29 values: |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 483 | v0 = load32 arg(0) |
Mike Klein | 5e533c9 | 2019-07-22 13:44:54 -0500 | [diff] [blame] | 484 | ↑ v1 = splat FF (3.5733111e-43) |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 485 | v2 = extract v0 0 v1 |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 486 | v3 = load32 arg(1) |
| 487 | v4 = extract v3 0 v1 |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 488 | v5 = extract v0 24 v1 |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 489 | ↑ v6 = splat 100 (3.5873241e-43) |
| 490 | v7 = sub_i32 v6 v5 |
| 491 | v8 = mul_i32 v4 v7 |
| 492 | v9 = shr_i32 v8 8 |
| 493 | v10 = add_i32 v2 v9 |
| 494 | v11 = extract v0 8 v1 |
| 495 | v12 = extract v3 8 v1 |
| 496 | v13 = mul_i32 v12 v7 |
Mike Klein | 8ac9f4e | 2019-07-25 14:32:19 -0500 | [diff] [blame] | 497 | v14 = shr_i32 v13 8 |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 498 | v15 = add_i32 v11 v14 |
| 499 | v16 = pack v10 v15 8 |
| 500 | v17 = extract v0 16 v1 |
| 501 | v18 = extract v3 16 v1 |
| 502 | v19 = mul_i32 v18 v7 |
Mike Klein | 8ac9f4e | 2019-07-25 14:32:19 -0500 | [diff] [blame] | 503 | v20 = shr_i32 v19 8 |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 504 | v21 = add_i32 v17 v20 |
| 505 | v22 = extract v3 24 v1 |
| 506 | v23 = mul_i32 v22 v7 |
| 507 | v24 = shr_i32 v23 8 |
| 508 | v25 = add_i32 v5 v24 |
| 509 | v26 = pack v21 v25 8 |
| 510 | v27 = pack v16 v26 16 |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 511 | store32 arg(1) v27 |
Mike Klein | aab45b5 | 2019-07-02 15:39:23 -0500 | [diff] [blame] | 512 | |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 513 | 9 registers, 29 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 514 | 0 r0 = splat FF (3.5733111e-43) |
| 515 | 1 r1 = splat 100 (3.5873241e-43) |
Mike Klein | 397fc88 | 2019-06-20 11:37:10 -0500 | [diff] [blame] | 516 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 517 | 2 r2 = load32 arg(0) |
| 518 | 3 r3 = extract r2 0 r0 |
| 519 | 4 r4 = load32 arg(1) |
| 520 | 5 r5 = extract r4 0 r0 |
| 521 | 6 r6 = extract r2 24 r0 |
| 522 | 7 r7 = sub_i32 r1 r6 |
| 523 | 8 r5 = mul_i32 r5 r7 |
| 524 | 9 r5 = shr_i32 r5 8 |
| 525 | 10 r5 = add_i32 r3 r5 |
| 526 | 11 r3 = extract r2 8 r0 |
| 527 | 12 r8 = extract r4 8 r0 |
| 528 | 13 r8 = mul_i32 r8 r7 |
| 529 | 14 r8 = shr_i32 r8 8 |
| 530 | 15 r8 = add_i32 r3 r8 |
| 531 | 16 r8 = pack r5 r8 8 |
| 532 | 17 r2 = extract r2 16 r0 |
| 533 | 18 r5 = extract r4 16 r0 |
| 534 | 19 r5 = mul_i32 r5 r7 |
| 535 | 20 r5 = shr_i32 r5 8 |
| 536 | 21 r5 = add_i32 r2 r5 |
| 537 | 22 r4 = extract r4 24 r0 |
| 538 | 23 r7 = mul_i32 r4 r7 |
| 539 | 24 r7 = shr_i32 r7 8 |
| 540 | 25 r7 = add_i32 r6 r7 |
| 541 | 26 r7 = pack r5 r7 8 |
| 542 | 27 r7 = pack r8 r7 16 |
| 543 | 28 store32 arg(1) r7 |
Mike Klein | 397fc88 | 2019-06-20 11:37:10 -0500 | [diff] [blame] | 544 | |
Mike Klein | 7b7077c | 2019-06-03 17:10:59 -0500 | [diff] [blame] | 545 | I32 8888 over 8888 |
Mike Klein | aab45b5 | 2019-07-02 15:39:23 -0500 | [diff] [blame] | 546 | 29 values: |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 547 | v0 = load32 arg(0) |
Mike Klein | 5e533c9 | 2019-07-22 13:44:54 -0500 | [diff] [blame] | 548 | ↑ v1 = splat FF (3.5733111e-43) |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 549 | v2 = bit_and v0 v1 |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 550 | v3 = load32 arg(1) |
| 551 | v4 = bit_and v3 v1 |
Mike Klein | 8ac9f4e | 2019-07-25 14:32:19 -0500 | [diff] [blame] | 552 | v5 = shr_i32 v0 24 |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 553 | ↑ v6 = splat 100 (3.5873241e-43) |
| 554 | v7 = sub_i32 v6 v5 |
| 555 | v8 = mul_i16x2 v4 v7 |
| 556 | v9 = shr_i32 v8 8 |
| 557 | v10 = add_i32 v2 v9 |
| 558 | v11 = bytes v0 2 |
| 559 | v12 = bytes v3 2 |
| 560 | v13 = mul_i16x2 v12 v7 |
Mike Klein | 8ac9f4e | 2019-07-25 14:32:19 -0500 | [diff] [blame] | 561 | v14 = shr_i32 v13 8 |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 562 | v15 = add_i32 v11 v14 |
| 563 | v16 = pack v10 v15 8 |
| 564 | v17 = bytes v0 3 |
| 565 | v18 = bytes v3 3 |
| 566 | v19 = mul_i16x2 v18 v7 |
Mike Klein | 8ac9f4e | 2019-07-25 14:32:19 -0500 | [diff] [blame] | 567 | v20 = shr_i32 v19 8 |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 568 | v21 = add_i32 v17 v20 |
| 569 | v22 = shr_i32 v3 24 |
| 570 | v23 = mul_i16x2 v22 v7 |
| 571 | v24 = shr_i32 v23 8 |
| 572 | v25 = add_i32 v5 v24 |
| 573 | v26 = pack v21 v25 8 |
| 574 | v27 = pack v16 v26 16 |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 575 | store32 arg(1) v27 |
Mike Klein | aab45b5 | 2019-07-02 15:39:23 -0500 | [diff] [blame] | 576 | |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 577 | 9 registers, 29 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 578 | 0 r0 = splat FF (3.5733111e-43) |
| 579 | 1 r1 = splat 100 (3.5873241e-43) |
Mike Klein | 754bad3 | 2019-06-05 10:47:46 -0500 | [diff] [blame] | 580 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 581 | 2 r2 = load32 arg(0) |
| 582 | 3 r3 = bit_and r2 r0 |
| 583 | 4 r4 = load32 arg(1) |
| 584 | 5 r5 = bit_and r4 r0 |
| 585 | 6 r6 = shr_i32 r2 24 |
| 586 | 7 r7 = sub_i32 r1 r6 |
| 587 | 8 r5 = mul_i16x2 r5 r7 |
| 588 | 9 r5 = shr_i32 r5 8 |
| 589 | 10 r5 = add_i32 r3 r5 |
| 590 | 11 r3 = bytes r2 2 |
| 591 | 12 r8 = bytes r4 2 |
| 592 | 13 r8 = mul_i16x2 r8 r7 |
| 593 | 14 r8 = shr_i32 r8 8 |
| 594 | 15 r8 = add_i32 r3 r8 |
| 595 | 16 r8 = pack r5 r8 8 |
| 596 | 17 r2 = bytes r2 3 |
| 597 | 18 r5 = bytes r4 3 |
| 598 | 19 r5 = mul_i16x2 r5 r7 |
| 599 | 20 r5 = shr_i32 r5 8 |
| 600 | 21 r5 = add_i32 r2 r5 |
| 601 | 22 r4 = shr_i32 r4 24 |
| 602 | 23 r7 = mul_i16x2 r4 r7 |
| 603 | 24 r7 = shr_i32 r7 8 |
| 604 | 25 r7 = add_i32 r6 r7 |
| 605 | 26 r7 = pack r5 r7 8 |
| 606 | 27 r7 = pack r8 r7 16 |
| 607 | 28 store32 arg(1) r7 |
Mike Klein | 821f5e8 | 2019-06-13 10:56:51 -0500 | [diff] [blame] | 608 | |
| 609 | I32 (SWAR) 8888 over 8888 |
Mike Klein | aab45b5 | 2019-07-02 15:39:23 -0500 | [diff] [blame] | 610 | 15 values: |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 611 | v0 = load32 arg(0) |
| 612 | v1 = bytes v0 404 |
Mike Klein | 5e533c9 | 2019-07-22 13:44:54 -0500 | [diff] [blame] | 613 | ↑ v2 = splat 1000100 (2.3510604e-38) |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 614 | v3 = sub_i16x2 v2 v1 |
| 615 | v4 = load32 arg(1) |
Mike Klein | 5e533c9 | 2019-07-22 13:44:54 -0500 | [diff] [blame] | 616 | ↑ v5 = splat FF00FF (2.3418409e-38) |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 617 | v6 = bit_and v4 v5 |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 618 | v7 = mul_i16x2 v6 v3 |
| 619 | v8 = shr_i16x2 v7 8 |
| 620 | v9 = shr_i16x2 v4 8 |
| 621 | v10 = mul_i16x2 v9 v3 |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 622 | v11 = bit_clear v10 v5 |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 623 | v12 = bit_or v8 v11 |
Mike Klein | c2fb3b4 | 2019-07-17 12:09:09 -0500 | [diff] [blame] | 624 | v13 = add_i32 v0 v12 |
| 625 | store32 arg(1) v13 |
Mike Klein | aab45b5 | 2019-07-02 15:39:23 -0500 | [diff] [blame] | 626 | |
Mike Klein | 2b7b2a2 | 2019-06-23 20:35:28 -0400 | [diff] [blame] | 627 | 6 registers, 15 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 628 | 0 r0 = splat 1000100 (2.3510604e-38) |
| 629 | 1 r1 = splat FF00FF (2.3418409e-38) |
Mike Klein | 821f5e8 | 2019-06-13 10:56:51 -0500 | [diff] [blame] | 630 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 631 | 2 r2 = load32 arg(0) |
| 632 | 3 r3 = bytes r2 404 |
| 633 | 4 r3 = sub_i16x2 r0 r3 |
| 634 | 5 r4 = load32 arg(1) |
| 635 | 6 r5 = bit_and r4 r1 |
| 636 | 7 r5 = mul_i16x2 r5 r3 |
| 637 | 8 r5 = shr_i16x2 r5 8 |
| 638 | 9 r4 = shr_i16x2 r4 8 |
| 639 | 10 r3 = mul_i16x2 r4 r3 |
| 640 | 11 r3 = bit_clear r3 r1 |
| 641 | 12 r3 = bit_or r5 r3 |
| 642 | 13 r3 = add_i32 r2 r3 |
| 643 | 14 store32 arg(1) r3 |
Mike Klein | 7b7077c | 2019-06-03 17:10:59 -0500 | [diff] [blame] | 644 | |
Mike Klein | f996311 | 2019-08-08 15:13:25 -0400 | [diff] [blame] | 645 | 6 values: |
Mike Klein | 0f61c12 | 2019-10-16 10:46:01 -0500 | [diff] [blame] | 646 | ↟ v0 = splat 1 (1.4012985e-45) |
| 647 | ↟ v1 = splat 2 (2.8025969e-45) |
Mike Klein | f996311 | 2019-08-08 15:13:25 -0400 | [diff] [blame] | 648 | ↑ v2 = add_i32 v0 v1 |
| 649 | v3 = load32 arg(0) |
| 650 | v4 = mul_i32 v3 v2 |
| 651 | store32 arg(0) v4 |
| 652 | |
| 653 | 2 registers, 6 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 654 | 0 r0 = splat 1 (1.4012985e-45) |
| 655 | 1 r1 = splat 2 (2.8025969e-45) |
| 656 | 2 r1 = add_i32 r0 r1 |
Mike Klein | f996311 | 2019-08-08 15:13:25 -0400 | [diff] [blame] | 657 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 658 | 3 r0 = load32 arg(0) |
| 659 | 4 r0 = mul_i32 r0 r1 |
| 660 | 5 store32 arg(0) r0 |
Mike Klein | f996311 | 2019-08-08 15:13:25 -0400 | [diff] [blame] | 661 | |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 662 | 19 values: |
| 663 | ↑ v0 = splat FF (3.5733111e-43) |
| 664 | v1 = load32 arg(0) |
| 665 | v2 = extract v1 0 v0 |
| 666 | v3 = load32 arg(1) |
| 667 | v4 = extract v3 0 v0 |
| 668 | v5 = add_i32 v2 v4 |
| 669 | v6 = extract v1 8 v0 |
| 670 | v7 = extract v3 8 v0 |
| 671 | v8 = add_i32 v6 v7 |
| 672 | v9 = pack v5 v8 8 |
| 673 | v10 = extract v1 16 v0 |
| 674 | v11 = extract v3 16 v0 |
| 675 | v12 = add_i32 v10 v11 |
| 676 | v13 = extract v1 24 v0 |
| 677 | v14 = extract v3 24 v0 |
| 678 | v15 = add_i32 v13 v14 |
| 679 | v16 = pack v12 v15 8 |
| 680 | v17 = pack v9 v16 16 |
| 681 | store32 arg(1) v17 |
| 682 | |
| 683 | 6 registers, 19 instructions: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 684 | 0 r0 = splat FF (3.5733111e-43) |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 685 | loop: |
Mike Klein | b5c4355 | 2020-01-07 11:39:30 -0600 | [diff] [blame] | 686 | 1 r1 = load32 arg(0) |
| 687 | 2 r2 = extract r1 0 r0 |
| 688 | 3 r3 = load32 arg(1) |
| 689 | 4 r4 = extract r3 0 r0 |
| 690 | 5 r4 = add_i32 r2 r4 |
| 691 | 6 r2 = extract r1 8 r0 |
| 692 | 7 r5 = extract r3 8 r0 |
| 693 | 8 r5 = add_i32 r2 r5 |
| 694 | 9 r5 = pack r4 r5 8 |
| 695 | 10 r4 = extract r1 16 r0 |
| 696 | 11 r2 = extract r3 16 r0 |
| 697 | 12 r2 = add_i32 r4 r2 |
| 698 | 13 r1 = extract r1 24 r0 |
| 699 | 14 r3 = extract r3 24 r0 |
| 700 | 15 r3 = add_i32 r1 r3 |
| 701 | 16 r3 = pack r2 r3 8 |
| 702 | 17 r3 = pack r5 r3 16 |
| 703 | 18 store32 arg(1) r3 |
Mike Klein | d48488b | 2019-10-22 12:27:58 -0500 | [diff] [blame] | 704 | |