| /external/XNNPACK/src/f32-gemm/gen-inc/ |
| D | 2x4-scalar.c | 55 float vacc12 = acc[6]; in xnn_f32_gemminc_ukernel_2x4__scalar() local
76 vacc12 += va1 * vb2; in xnn_f32_gemminc_ukernel_2x4__scalar()
89 vacc12 = math_max_f32(vacc12, vmin); in xnn_f32_gemminc_ukernel_2x4__scalar()
99 vacc12 = math_min_f32(vacc12, vmax); in xnn_f32_gemminc_ukernel_2x4__scalar()
105 c1[2] = vacc12; in xnn_f32_gemminc_ukernel_2x4__scalar()
122 vacc10 = vacc12; in xnn_f32_gemminc_ukernel_2x4__scalar()
|
| D | 2x4-wasm.c | 55 float vacc12 = acc[6]; in xnn_f32_gemminc_ukernel_2x4__wasm() local
76 vacc12 += va1 * vb2; in xnn_f32_gemminc_ukernel_2x4__wasm()
89 vacc12 = __builtin_wasm_max_f32(vacc12, vmin); in xnn_f32_gemminc_ukernel_2x4__wasm()
99 vacc12 = __builtin_wasm_min_f32(vacc12, vmax); in xnn_f32_gemminc_ukernel_2x4__wasm()
105 c1[2] = vacc12; in xnn_f32_gemminc_ukernel_2x4__wasm()
122 vacc10 = vacc12; in xnn_f32_gemminc_ukernel_2x4__wasm()
|
| D | 4x4-wasm.c | 67 float vacc12 = acc[6]; in xnn_f32_gemminc_ukernel_4x4__wasm() local
98 vacc12 += va1 * vb2; in xnn_f32_gemminc_ukernel_4x4__wasm()
119 vacc12 = __builtin_wasm_max_f32(vacc12, vmin); in xnn_f32_gemminc_ukernel_4x4__wasm()
137 vacc12 = __builtin_wasm_min_f32(vacc12, vmax); in xnn_f32_gemminc_ukernel_4x4__wasm()
161 c1[2] = vacc12; in xnn_f32_gemminc_ukernel_4x4__wasm()
188 vacc10 = vacc12; in xnn_f32_gemminc_ukernel_4x4__wasm()
|
| D | 4x4-scalar.c | 67 float vacc12 = acc[6]; in xnn_f32_gemminc_ukernel_4x4__scalar() local
98 vacc12 += va1 * vb2; in xnn_f32_gemminc_ukernel_4x4__scalar()
119 vacc12 = math_max_f32(vacc12, vmin); in xnn_f32_gemminc_ukernel_4x4__scalar()
137 vacc12 = math_min_f32(vacc12, vmax); in xnn_f32_gemminc_ukernel_4x4__scalar()
161 c1[2] = vacc12; in xnn_f32_gemminc_ukernel_4x4__scalar()
188 vacc10 = vacc12; in xnn_f32_gemminc_ukernel_4x4__scalar()
|
| /external/XNNPACK/src/f32-gemm/gen/ |
| D | 2x4-wasm.c | 54 float vacc12 = vacc02; in xnn_f32_gemm_ukernel_2x4__wasm() local
74 vacc12 += va1 * vb2; in xnn_f32_gemm_ukernel_2x4__wasm()
87 vacc12 = __builtin_wasm_max_f32(vacc12, vmin); in xnn_f32_gemm_ukernel_2x4__wasm()
97 vacc12 = __builtin_wasm_min_f32(vacc12, vmax); in xnn_f32_gemm_ukernel_2x4__wasm()
103 c1[2] = vacc12; in xnn_f32_gemm_ukernel_2x4__wasm()
120 vacc10 = vacc12; in xnn_f32_gemm_ukernel_2x4__wasm()
|
| D | 2x4-scalar.c | 54 float vacc12 = vacc02; in xnn_f32_gemm_ukernel_2x4__scalar() local
74 vacc12 += va1 * vb2; in xnn_f32_gemm_ukernel_2x4__scalar()
87 vacc12 = math_max_f32(vacc12, vmin); in xnn_f32_gemm_ukernel_2x4__scalar()
97 vacc12 = math_min_f32(vacc12, vmax); in xnn_f32_gemm_ukernel_2x4__scalar()
103 c1[2] = vacc12; in xnn_f32_gemm_ukernel_2x4__scalar()
120 vacc10 = vacc12; in xnn_f32_gemm_ukernel_2x4__scalar()
|
| D | 4x4-scalar.c | 66 float vacc12 = vacc02; in xnn_f32_gemm_ukernel_4x4__scalar() local
96 vacc12 += va1 * vb2; in xnn_f32_gemm_ukernel_4x4__scalar()
117 vacc12 = math_max_f32(vacc12, vmin); in xnn_f32_gemm_ukernel_4x4__scalar()
135 vacc12 = math_min_f32(vacc12, vmax); in xnn_f32_gemm_ukernel_4x4__scalar()
159 c1[2] = vacc12; in xnn_f32_gemm_ukernel_4x4__scalar()
186 vacc10 = vacc12; in xnn_f32_gemm_ukernel_4x4__scalar()
|
| D | 4x4-wasm.c | 66 float vacc12 = vacc02; in xnn_f32_gemm_ukernel_4x4__wasm() local
96 vacc12 += va1 * vb2; in xnn_f32_gemm_ukernel_4x4__wasm()
117 vacc12 = __builtin_wasm_max_f32(vacc12, vmin); in xnn_f32_gemm_ukernel_4x4__wasm()
135 vacc12 = __builtin_wasm_min_f32(vacc12, vmax); in xnn_f32_gemm_ukernel_4x4__wasm()
159 c1[2] = vacc12; in xnn_f32_gemm_ukernel_4x4__wasm()
186 vacc10 = vacc12; in xnn_f32_gemm_ukernel_4x4__wasm()
|
| /external/XNNPACK/src/f32-igemm/gen/ |
| D | 2x4-wasm.c | 55 float vacc12 = vacc02; in xnn_f32_igemm_ukernel_2x4__wasm() local
90 vacc12 += va1 * vb2; in xnn_f32_igemm_ukernel_2x4__wasm()
105 vacc12 = __builtin_wasm_max_f32(vacc12, vmin); in xnn_f32_igemm_ukernel_2x4__wasm()
115 vacc12 = __builtin_wasm_min_f32(vacc12, vmax); in xnn_f32_igemm_ukernel_2x4__wasm()
121 c1[2] = vacc12; in xnn_f32_igemm_ukernel_2x4__wasm()
136 vacc10 = vacc12; in xnn_f32_igemm_ukernel_2x4__wasm()
|
| D | 2x4-scalar.c | 55 float vacc12 = vacc02; in xnn_f32_igemm_ukernel_2x4__scalar() local
90 vacc12 += va1 * vb2; in xnn_f32_igemm_ukernel_2x4__scalar()
105 vacc12 = math_max_f32(vacc12, vmin); in xnn_f32_igemm_ukernel_2x4__scalar()
115 vacc12 = math_min_f32(vacc12, vmax); in xnn_f32_igemm_ukernel_2x4__scalar()
121 c1[2] = vacc12; in xnn_f32_igemm_ukernel_2x4__scalar()
136 vacc10 = vacc12; in xnn_f32_igemm_ukernel_2x4__scalar()
|
| D | 4x4-scalar.c | 63 float vacc12 = vacc02; in xnn_f32_igemm_ukernel_4x4__scalar() local
118 vacc12 += va1 * vb2; in xnn_f32_igemm_ukernel_4x4__scalar()
141 vacc12 = math_max_f32(vacc12, vmin); in xnn_f32_igemm_ukernel_4x4__scalar()
159 vacc12 = math_min_f32(vacc12, vmax); in xnn_f32_igemm_ukernel_4x4__scalar()
183 c1[2] = vacc12; in xnn_f32_igemm_ukernel_4x4__scalar()
206 vacc10 = vacc12; in xnn_f32_igemm_ukernel_4x4__scalar()
|
| D | 4x4-wasm.c | 63 float vacc12 = vacc02; in xnn_f32_igemm_ukernel_4x4__wasm() local
118 vacc12 += va1 * vb2; in xnn_f32_igemm_ukernel_4x4__wasm()
141 vacc12 = __builtin_wasm_max_f32(vacc12, vmin); in xnn_f32_igemm_ukernel_4x4__wasm()
159 vacc12 = __builtin_wasm_min_f32(vacc12, vmax); in xnn_f32_igemm_ukernel_4x4__wasm()
183 c1[2] = vacc12; in xnn_f32_igemm_ukernel_4x4__wasm()
206 vacc10 = vacc12; in xnn_f32_igemm_ukernel_4x4__wasm()
|
| /external/XNNPACK/src/q8-gemm/ |
| D | 2x4c8-sse2.c | 71 __m128i vacc12 = vacc02; in xnn_q8_gemm_ukernel_2x4c8__sse2() local
100 vacc12 = _mm_add_epi32(vacc12, _mm_madd_epi16(vxa1, vxb2)); in xnn_q8_gemm_ukernel_2x4c8__sse2()
105 __m128i vacc1x0123 = sse_reduce4_i32(vacc10, vacc11, vacc12, vacc13); in xnn_q8_gemm_ukernel_2x4c8__sse2()
|