/external/XNNPACK/src/f32-dwconv/gen/ |
D | up2x25-wasm-acc2.c | 182 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_ukernel_up2x25__wasm_acc2() local 200 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__wasm_acc2() 218 vacc1p1 = math_muladd_f32(vi5x1, vk5x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__wasm_acc2() 236 vacc1p1 = math_muladd_f32(vi7x1, vk7x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__wasm_acc2() 254 vacc1p1 = math_muladd_f32(vi9x1, vk9x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__wasm_acc2() 272 vacc1p1 = math_muladd_f32(vi11x1, vk11x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__wasm_acc2() 290 vacc1p1 = math_muladd_f32(vi13x1, vk13x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__wasm_acc2() 308 vacc1p1 = math_muladd_f32(vi15x1, vk15x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__wasm_acc2() 326 vacc1p1 = math_muladd_f32(vi17x1, vk17x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__wasm_acc2() 344 vacc1p1 = math_muladd_f32(vi19x1, vk19x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__wasm_acc2() [all …]
|
D | up2x25-scalar-acc2.c | 182 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_ukernel_up2x25__scalar_acc2() local 200 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__scalar_acc2() 218 vacc1p1 = math_muladd_f32(vi5x1, vk5x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__scalar_acc2() 236 vacc1p1 = math_muladd_f32(vi7x1, vk7x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__scalar_acc2() 254 vacc1p1 = math_muladd_f32(vi9x1, vk9x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__scalar_acc2() 272 vacc1p1 = math_muladd_f32(vi11x1, vk11x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__scalar_acc2() 290 vacc1p1 = math_muladd_f32(vi13x1, vk13x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__scalar_acc2() 308 vacc1p1 = math_muladd_f32(vi15x1, vk15x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__scalar_acc2() 326 vacc1p1 = math_muladd_f32(vi17x1, vk17x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__scalar_acc2() 344 vacc1p1 = math_muladd_f32(vi19x1, vk19x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x25__scalar_acc2() [all …]
|
D | up2x25-minmax-scalar-acc2.c | 184 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_minmax_ukernel_up2x25__scalar_acc2() local 202 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__scalar_acc2() 220 vacc1p1 = math_muladd_f32(vi5x1, vk5x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__scalar_acc2() 238 vacc1p1 = math_muladd_f32(vi7x1, vk7x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__scalar_acc2() 256 vacc1p1 = math_muladd_f32(vi9x1, vk9x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__scalar_acc2() 274 vacc1p1 = math_muladd_f32(vi11x1, vk11x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__scalar_acc2() 292 vacc1p1 = math_muladd_f32(vi13x1, vk13x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__scalar_acc2() 310 vacc1p1 = math_muladd_f32(vi15x1, vk15x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__scalar_acc2() 328 vacc1p1 = math_muladd_f32(vi17x1, vk17x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__scalar_acc2() 346 vacc1p1 = math_muladd_f32(vi19x1, vk19x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__scalar_acc2() [all …]
|
D | up2x25-minmax-wasm-acc2.c | 184 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_minmax_ukernel_up2x25__wasm_acc2() local 202 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__wasm_acc2() 220 vacc1p1 = math_muladd_f32(vi5x1, vk5x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__wasm_acc2() 238 vacc1p1 = math_muladd_f32(vi7x1, vk7x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__wasm_acc2() 256 vacc1p1 = math_muladd_f32(vi9x1, vk9x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__wasm_acc2() 274 vacc1p1 = math_muladd_f32(vi11x1, vk11x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__wasm_acc2() 292 vacc1p1 = math_muladd_f32(vi13x1, vk13x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__wasm_acc2() 310 vacc1p1 = math_muladd_f32(vi15x1, vk15x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__wasm_acc2() 328 vacc1p1 = math_muladd_f32(vi17x1, vk17x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__wasm_acc2() 346 vacc1p1 = math_muladd_f32(vi19x1, vk19x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x25__wasm_acc2() [all …]
|
D | up2x9-scalar-acc2.c | 102 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_ukernel_up2x9__scalar_acc2() local 120 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x9__scalar_acc2() 138 vacc1p1 = math_muladd_f32(vi5x1, vk5x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x9__scalar_acc2() 156 vacc1p1 = math_muladd_f32(vi7x1, vk7x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x9__scalar_acc2() 171 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_ukernel_up2x9__scalar_acc2()
|
D | up2x9-wasm-acc2.c | 102 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_ukernel_up2x9__wasm_acc2() local 120 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x9__wasm_acc2() 138 vacc1p1 = math_muladd_f32(vi5x1, vk5x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x9__wasm_acc2() 156 vacc1p1 = math_muladd_f32(vi7x1, vk7x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x9__wasm_acc2() 171 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_ukernel_up2x9__wasm_acc2()
|
D | up2x9-minmax-scalar-acc2.c | 104 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_minmax_ukernel_up2x9__scalar_acc2() local 122 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x9__scalar_acc2() 140 vacc1p1 = math_muladd_f32(vi5x1, vk5x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x9__scalar_acc2() 158 vacc1p1 = math_muladd_f32(vi7x1, vk7x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x9__scalar_acc2() 173 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_minmax_ukernel_up2x9__scalar_acc2()
|
D | up2x9-minmax-wasm-acc2.c | 104 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_minmax_ukernel_up2x9__wasm_acc2() local 122 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x9__wasm_acc2() 140 vacc1p1 = math_muladd_f32(vi5x1, vk5x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x9__wasm_acc2() 158 vacc1p1 = math_muladd_f32(vi7x1, vk7x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x9__wasm_acc2() 173 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_minmax_ukernel_up2x9__wasm_acc2()
|
D | up2x4-wasm-acc2.c | 77 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_ukernel_up2x4__wasm_acc2() local 95 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x4__wasm_acc2() 101 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_ukernel_up2x4__wasm_acc2()
|
D | up2x4-scalar-acc2.c | 77 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_ukernel_up2x4__scalar_acc2() local 95 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_ukernel_up2x4__scalar_acc2() 101 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_ukernel_up2x4__scalar_acc2()
|
D | up2x4-minmax-scalar-acc2.c | 79 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_minmax_ukernel_up2x4__scalar_acc2() local 97 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x4__scalar_acc2() 103 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_minmax_ukernel_up2x4__scalar_acc2()
|
D | up2x4-minmax-wasm-acc2.c | 79 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_minmax_ukernel_up2x4__wasm_acc2() local 97 vacc1p1 = math_muladd_f32(vi3x1, vk3x1, vacc1p1); in xnn_f32_dwconv_minmax_ukernel_up2x4__wasm_acc2() 103 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_minmax_ukernel_up2x4__wasm_acc2()
|
D | up2x3-wasm-acc2.c | 72 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_ukernel_up2x3__wasm_acc2() local 87 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_ukernel_up2x3__wasm_acc2()
|
D | up2x3-scalar-acc2.c | 72 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_ukernel_up2x3__scalar_acc2() local 87 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_ukernel_up2x3__scalar_acc2()
|
D | up2x3-minmax-wasm-acc2.c | 74 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_minmax_ukernel_up2x3__wasm_acc2() local 89 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_minmax_ukernel_up2x3__wasm_acc2()
|
D | up2x3-minmax-scalar-acc2.c | 74 float vacc1p1 = vi1x1 * vk1x1; in xnn_f32_dwconv_minmax_ukernel_up2x3__scalar_acc2() local 89 vacc1p0 = vacc1p0 + vacc1p1; in xnn_f32_dwconv_minmax_ukernel_up2x3__scalar_acc2()
|