/external/XNNPACK/src/f32-gemm/gen/ |
D | 4x2-minmax-wasm.c | 68 float vacc30 = vacc00; in xnn_f32_gemm_minmax_ukernel_4x2__wasm() local 88 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemm_minmax_ukernel_4x2__wasm() 100 vacc30 = __builtin_wasm_max_f32(vacc30, vmin); in xnn_f32_gemm_minmax_ukernel_4x2__wasm() 109 vacc30 = __builtin_wasm_min_f32(vacc30, vmax); in xnn_f32_gemm_minmax_ukernel_4x2__wasm() 113 c3[0] = vacc30; in xnn_f32_gemm_minmax_ukernel_4x2__wasm() 134 c3[0] = vacc30; in xnn_f32_gemm_minmax_ukernel_4x2__wasm()
|
D | 4x2-minmax-scalar.c | 68 float vacc30 = vacc00; in xnn_f32_gemm_minmax_ukernel_4x2__scalar() local 88 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemm_minmax_ukernel_4x2__scalar() 100 vacc30 = math_max_f32(vacc30, vmin); in xnn_f32_gemm_minmax_ukernel_4x2__scalar() 109 vacc30 = math_min_f32(vacc30, vmax); in xnn_f32_gemm_minmax_ukernel_4x2__scalar() 113 c3[0] = vacc30; in xnn_f32_gemm_minmax_ukernel_4x2__scalar() 134 c3[0] = vacc30; in xnn_f32_gemm_minmax_ukernel_4x2__scalar()
|
D | 4x4-minmax-wasm.c | 74 float vacc30 = vacc00; in xnn_f32_gemm_minmax_ukernel_4x4__wasm() local 104 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemm_minmax_ukernel_4x4__wasm() 124 vacc30 = __builtin_wasm_max_f32(vacc30, vmin); in xnn_f32_gemm_minmax_ukernel_4x4__wasm() 141 vacc30 = __builtin_wasm_min_f32(vacc30, vmax); in xnn_f32_gemm_minmax_ukernel_4x4__wasm() 147 c3[0] = vacc30; in xnn_f32_gemm_minmax_ukernel_4x4__wasm() 176 c3[0] = vacc30; in xnn_f32_gemm_minmax_ukernel_4x4__wasm() 178 vacc30 = vacc32; in xnn_f32_gemm_minmax_ukernel_4x4__wasm() 194 c3[0] = vacc30; in xnn_f32_gemm_minmax_ukernel_4x4__wasm()
|
D | 4x4-minmax-scalar.c | 74 float vacc30 = vacc00; in xnn_f32_gemm_minmax_ukernel_4x4__scalar() local 104 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemm_minmax_ukernel_4x4__scalar() 124 vacc30 = math_max_f32(vacc30, vmin); in xnn_f32_gemm_minmax_ukernel_4x4__scalar() 141 vacc30 = math_min_f32(vacc30, vmax); in xnn_f32_gemm_minmax_ukernel_4x4__scalar() 147 c3[0] = vacc30; in xnn_f32_gemm_minmax_ukernel_4x4__scalar() 176 c3[0] = vacc30; in xnn_f32_gemm_minmax_ukernel_4x4__scalar() 178 vacc30 = vacc32; in xnn_f32_gemm_minmax_ukernel_4x4__scalar() 194 c3[0] = vacc30; in xnn_f32_gemm_minmax_ukernel_4x4__scalar()
|
D | 4x4-relu-wasm.c | 72 float vacc30 = vacc00; in xnn_f32_gemm_relu_ukernel_4x4__wasm() local 102 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemm_relu_ukernel_4x4__wasm() 122 vacc30 = __builtin_wasm_max_f32(vacc30, 0.0f); in xnn_f32_gemm_relu_ukernel_4x4__wasm() 128 c3[0] = vacc30; in xnn_f32_gemm_relu_ukernel_4x4__wasm() 157 c3[0] = vacc30; in xnn_f32_gemm_relu_ukernel_4x4__wasm() 159 vacc30 = vacc32; in xnn_f32_gemm_relu_ukernel_4x4__wasm() 175 c3[0] = vacc30; in xnn_f32_gemm_relu_ukernel_4x4__wasm()
|
D | 4x4-relu-scalar.c | 72 float vacc30 = vacc00; in xnn_f32_gemm_relu_ukernel_4x4__scalar() local 102 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemm_relu_ukernel_4x4__scalar() 122 vacc30 = math_max_f32(vacc30, 0.0f); in xnn_f32_gemm_relu_ukernel_4x4__scalar() 128 c3[0] = vacc30; in xnn_f32_gemm_relu_ukernel_4x4__scalar() 157 c3[0] = vacc30; in xnn_f32_gemm_relu_ukernel_4x4__scalar() 159 vacc30 = vacc32; in xnn_f32_gemm_relu_ukernel_4x4__scalar() 175 c3[0] = vacc30; in xnn_f32_gemm_relu_ukernel_4x4__scalar()
|
D | 4x2-relu-scalar.c | 66 float vacc30 = vacc00; in xnn_f32_gemm_relu_ukernel_4x2__scalar() local 86 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemm_relu_ukernel_4x2__scalar() 98 vacc30 = math_max_f32(vacc30, 0.0f); in xnn_f32_gemm_relu_ukernel_4x2__scalar() 102 c3[0] = vacc30; in xnn_f32_gemm_relu_ukernel_4x2__scalar() 123 c3[0] = vacc30; in xnn_f32_gemm_relu_ukernel_4x2__scalar()
|
D | 4x2-relu-wasm.c | 66 float vacc30 = vacc00; in xnn_f32_gemm_relu_ukernel_4x2__wasm() local 86 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemm_relu_ukernel_4x2__wasm() 98 vacc30 = __builtin_wasm_max_f32(vacc30, 0.0f); in xnn_f32_gemm_relu_ukernel_4x2__wasm() 102 c3[0] = vacc30; in xnn_f32_gemm_relu_ukernel_4x2__wasm() 123 c3[0] = vacc30; in xnn_f32_gemm_relu_ukernel_4x2__wasm()
|
D | 4x4-scalar.c | 72 float vacc30 = vacc00; in xnn_f32_gemm_ukernel_4x4__scalar() local 102 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemm_ukernel_4x4__scalar() 112 c3[0] = vacc30; in xnn_f32_gemm_ukernel_4x4__scalar() 141 c3[0] = vacc30; in xnn_f32_gemm_ukernel_4x4__scalar() 143 vacc30 = vacc32; in xnn_f32_gemm_ukernel_4x4__scalar() 159 c3[0] = vacc30; in xnn_f32_gemm_ukernel_4x4__scalar()
|
D | 4x2-scalar.c | 66 float vacc30 = vacc00; in xnn_f32_gemm_ukernel_4x2__scalar() local 86 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemm_ukernel_4x2__scalar() 94 c3[0] = vacc30; in xnn_f32_gemm_ukernel_4x2__scalar() 115 c3[0] = vacc30; in xnn_f32_gemm_ukernel_4x2__scalar()
|
/external/XNNPACK/src/f32-igemm/gen/ |
D | 4x2-minmax-scalar.c | 65 float vacc30 = vacc00; in xnn_f32_igemm_minmax_ukernel_4x2__scalar() local 110 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_igemm_minmax_ukernel_4x2__scalar() 124 vacc30 = math_max_f32(vacc30, vmin); in xnn_f32_igemm_minmax_ukernel_4x2__scalar() 133 vacc30 = math_min_f32(vacc30, vmax); in xnn_f32_igemm_minmax_ukernel_4x2__scalar() 137 c3[0] = vacc30; in xnn_f32_igemm_minmax_ukernel_4x2__scalar() 154 c3[0] = vacc30; in xnn_f32_igemm_minmax_ukernel_4x2__scalar()
|
D | 4x2-minmax-wasm.c | 65 float vacc30 = vacc00; in xnn_f32_igemm_minmax_ukernel_4x2__wasm() local 110 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_igemm_minmax_ukernel_4x2__wasm() 124 vacc30 = __builtin_wasm_max_f32(vacc30, vmin); in xnn_f32_igemm_minmax_ukernel_4x2__wasm() 133 vacc30 = __builtin_wasm_min_f32(vacc30, vmax); in xnn_f32_igemm_minmax_ukernel_4x2__wasm() 137 c3[0] = vacc30; in xnn_f32_igemm_minmax_ukernel_4x2__wasm() 154 c3[0] = vacc30; in xnn_f32_igemm_minmax_ukernel_4x2__wasm()
|
D | 4x4-minmax-wasm.c | 71 float vacc30 = vacc00; in xnn_f32_igemm_minmax_ukernel_4x4__wasm() local 126 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_igemm_minmax_ukernel_4x4__wasm() 148 vacc30 = __builtin_wasm_max_f32(vacc30, vmin); in xnn_f32_igemm_minmax_ukernel_4x4__wasm() 165 vacc30 = __builtin_wasm_min_f32(vacc30, vmax); in xnn_f32_igemm_minmax_ukernel_4x4__wasm() 171 c3[0] = vacc30; in xnn_f32_igemm_minmax_ukernel_4x4__wasm() 196 c3[0] = vacc30; in xnn_f32_igemm_minmax_ukernel_4x4__wasm() 198 vacc30 = vacc32; in xnn_f32_igemm_minmax_ukernel_4x4__wasm() 214 c3[0] = vacc30; in xnn_f32_igemm_minmax_ukernel_4x4__wasm()
|
D | 4x4-minmax-scalar.c | 71 float vacc30 = vacc00; in xnn_f32_igemm_minmax_ukernel_4x4__scalar() local 126 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_igemm_minmax_ukernel_4x4__scalar() 148 vacc30 = math_max_f32(vacc30, vmin); in xnn_f32_igemm_minmax_ukernel_4x4__scalar() 165 vacc30 = math_min_f32(vacc30, vmax); in xnn_f32_igemm_minmax_ukernel_4x4__scalar() 171 c3[0] = vacc30; in xnn_f32_igemm_minmax_ukernel_4x4__scalar() 196 c3[0] = vacc30; in xnn_f32_igemm_minmax_ukernel_4x4__scalar() 198 vacc30 = vacc32; in xnn_f32_igemm_minmax_ukernel_4x4__scalar() 214 c3[0] = vacc30; in xnn_f32_igemm_minmax_ukernel_4x4__scalar()
|
D | 4x4-relu-wasm.c | 69 float vacc30 = vacc00; in xnn_f32_igemm_relu_ukernel_4x4__wasm() local 124 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_igemm_relu_ukernel_4x4__wasm() 146 vacc30 = __builtin_wasm_max_f32(vacc30, 0.0f); in xnn_f32_igemm_relu_ukernel_4x4__wasm() 152 c3[0] = vacc30; in xnn_f32_igemm_relu_ukernel_4x4__wasm() 177 c3[0] = vacc30; in xnn_f32_igemm_relu_ukernel_4x4__wasm() 179 vacc30 = vacc32; in xnn_f32_igemm_relu_ukernel_4x4__wasm() 195 c3[0] = vacc30; in xnn_f32_igemm_relu_ukernel_4x4__wasm()
|
D | 4x4-relu-scalar.c | 69 float vacc30 = vacc00; in xnn_f32_igemm_relu_ukernel_4x4__scalar() local 124 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_igemm_relu_ukernel_4x4__scalar() 146 vacc30 = math_max_f32(vacc30, 0.0f); in xnn_f32_igemm_relu_ukernel_4x4__scalar() 152 c3[0] = vacc30; in xnn_f32_igemm_relu_ukernel_4x4__scalar() 177 c3[0] = vacc30; in xnn_f32_igemm_relu_ukernel_4x4__scalar() 179 vacc30 = vacc32; in xnn_f32_igemm_relu_ukernel_4x4__scalar() 195 c3[0] = vacc30; in xnn_f32_igemm_relu_ukernel_4x4__scalar()
|
D | 4x2-relu-scalar.c | 63 float vacc30 = vacc00; in xnn_f32_igemm_relu_ukernel_4x2__scalar() local 108 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_igemm_relu_ukernel_4x2__scalar() 122 vacc30 = math_max_f32(vacc30, 0.0f); in xnn_f32_igemm_relu_ukernel_4x2__scalar() 126 c3[0] = vacc30; in xnn_f32_igemm_relu_ukernel_4x2__scalar() 143 c3[0] = vacc30; in xnn_f32_igemm_relu_ukernel_4x2__scalar()
|
D | 4x2-relu-wasm.c | 63 float vacc30 = vacc00; in xnn_f32_igemm_relu_ukernel_4x2__wasm() local 108 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_igemm_relu_ukernel_4x2__wasm() 122 vacc30 = __builtin_wasm_max_f32(vacc30, 0.0f); in xnn_f32_igemm_relu_ukernel_4x2__wasm() 126 c3[0] = vacc30; in xnn_f32_igemm_relu_ukernel_4x2__wasm() 143 c3[0] = vacc30; in xnn_f32_igemm_relu_ukernel_4x2__wasm()
|
D | 4x4-scalar.c | 69 float vacc30 = vacc00; in xnn_f32_igemm_ukernel_4x4__scalar() local 124 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_igemm_ukernel_4x4__scalar() 136 c3[0] = vacc30; in xnn_f32_igemm_ukernel_4x4__scalar() 161 c3[0] = vacc30; in xnn_f32_igemm_ukernel_4x4__scalar() 163 vacc30 = vacc32; in xnn_f32_igemm_ukernel_4x4__scalar() 179 c3[0] = vacc30; in xnn_f32_igemm_ukernel_4x4__scalar()
|
D | 4x2-scalar.c | 63 float vacc30 = vacc00; in xnn_f32_igemm_ukernel_4x2__scalar() local 108 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_igemm_ukernel_4x2__scalar() 118 c3[0] = vacc30; in xnn_f32_igemm_ukernel_4x2__scalar() 135 c3[0] = vacc30; in xnn_f32_igemm_ukernel_4x2__scalar()
|
/external/XNNPACK/src/f32-gemm/gen-inc/ |
D | 4x4inc-minmax-wasm.c | 75 float vacc30 = acc[12]; in xnn_f32_gemminc_minmax_ukernel_4x4__wasm() local 106 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemminc_minmax_ukernel_4x4__wasm() 126 vacc30 = __builtin_wasm_max_f32(vacc30, vmin); in xnn_f32_gemminc_minmax_ukernel_4x4__wasm() 143 vacc30 = __builtin_wasm_min_f32(vacc30, vmax); in xnn_f32_gemminc_minmax_ukernel_4x4__wasm() 149 c3[0] = vacc30; in xnn_f32_gemminc_minmax_ukernel_4x4__wasm() 178 c3[0] = vacc30; in xnn_f32_gemminc_minmax_ukernel_4x4__wasm() 180 vacc30 = vacc32; in xnn_f32_gemminc_minmax_ukernel_4x4__wasm() 196 c3[0] = vacc30; in xnn_f32_gemminc_minmax_ukernel_4x4__wasm()
|
D | 4x4inc-minmax-scalar.c | 75 float vacc30 = acc[12]; in xnn_f32_gemminc_minmax_ukernel_4x4__scalar() local 106 vacc30 = math_muladd_f32(va3, vb0, vacc30); in xnn_f32_gemminc_minmax_ukernel_4x4__scalar() 126 vacc30 = math_max_f32(vacc30, vmin); in xnn_f32_gemminc_minmax_ukernel_4x4__scalar() 143 vacc30 = math_min_f32(vacc30, vmax); in xnn_f32_gemminc_minmax_ukernel_4x4__scalar() 149 c3[0] = vacc30; in xnn_f32_gemminc_minmax_ukernel_4x4__scalar() 178 c3[0] = vacc30; in xnn_f32_gemminc_minmax_ukernel_4x4__scalar() 180 vacc30 = vacc32; in xnn_f32_gemminc_minmax_ukernel_4x4__scalar() 196 c3[0] = vacc30; in xnn_f32_gemminc_minmax_ukernel_4x4__scalar()
|