/external/llvm-project/clang/test/CodeGen/arm-mve-intrinsics/ |
D | vqdmlad.c | 12 int8x16_t test_vqdmladhq_s8(int8x16_t inactive, int8x16_t a, int8x16_t b) { in test_vqdmladhq_s8() 51 int8x16_t test_vqdmladhxq_s8(int8x16_t inactive, int8x16_t a, int8x16_t b) { in test_vqdmladhxq_s8() 90 int8x16_t test_vqdmlsdhq_s8(int8x16_t inactive, int8x16_t a, int8x16_t b) { in test_vqdmlsdhq_s8() 129 int8x16_t test_vqdmlsdhxq_s8(int8x16_t inactive, int8x16_t a, int8x16_t b) { in test_vqdmlsdhxq_s8() 168 int8x16_t test_vqrdmladhq_s8(int8x16_t inactive, int8x16_t a, int8x16_t b) { in test_vqrdmladhq_s8() 207 int8x16_t test_vqrdmladhxq_s8(int8x16_t inactive, int8x16_t a, int8x16_t b) { in test_vqrdmladhxq_s8() 246 int8x16_t test_vqrdmlsdhq_s8(int8x16_t inactive, int8x16_t a, int8x16_t b) { in test_vqrdmlsdhq_s8() 285 int8x16_t test_vqrdmlsdhxq_s8(int8x16_t inactive, int8x16_t a, int8x16_t b) { in test_vqrdmlsdhxq_s8() 326 int8x16_t test_vqdmladhq_m_s8(int8x16_t inactive, int8x16_t a, int8x16_t b, mve_pred16_t p) { in test_vqdmladhq_m_s8() 371 int8x16_t test_vqdmladhxq_m_s8(int8x16_t inactive, int8x16_t a, int8x16_t b, mve_pred16_t p) { in test_vqdmladhxq_m_s8() [all …]
|
D | vhcaddq.c | 12 int8x16_t test_vhcaddq_rot90_s8(int8x16_t a, int8x16_t b) in test_vhcaddq_rot90_s8() 54 int8x16_t test_vhcaddq_rot270_s8(int8x16_t a, int8x16_t b) in test_vhcaddq_rot270_s8() 98 int8x16_t test_vhcaddq_rot90_x_s8(int8x16_t a, int8x16_t b, mve_pred16_t p) in test_vhcaddq_rot90_x_s8() 146 int8x16_t test_vhcaddq_rot270_x_s8(int8x16_t a, int8x16_t b, mve_pred16_t p) in test_vhcaddq_rot270_x_s8() 194 int8x16_t test_vhcaddq_rot90_m_s8(int8x16_t inactive, int8x16_t a, int8x16_t b, mve_pred16_t p) in test_vhcaddq_rot90_m_s8() 242 int8x16_t test_vhcaddq_rot270_m_s8(int8x16_t inactive, int8x16_t a, int8x16_t b, mve_pred16_t p) in test_vhcaddq_rot270_m_s8()
|
D | absneg.c | 42 int8x16_t test_vabsq_s8(int8x16_t a) in test_vabsq_s8() 88 int8x16_t test_vmvnq_s8(int8x16_t a) in test_vmvnq_s8() 174 int8x16_t test_vmvnq_m_s8(int8x16_t inactive, int8x16_t a, mve_pred16_t p) in test_vmvnq_m_s8() 270 int8x16_t test_vmvnq_x_s8(int8x16_t a, mve_pred16_t p) in test_vmvnq_x_s8() 392 int8x16_t test_vnegq_s8(int8x16_t a) in test_vnegq_s8() 438 int8x16_t test_vqabsq_s8(int8x16_t a) in test_vqabsq_s8() 490 int8x16_t test_vqnegq_s8(int8x16_t a) in test_vqnegq_s8() 570 int8x16_t test_vnegq_m_s8(int8x16_t inactive, int8x16_t a, mve_pred16_t p) in test_vnegq_m_s8() 650 int8x16_t test_vnegq_x_s8(int8x16_t a, mve_pred16_t p) in test_vnegq_x_s8() 730 int8x16_t test_vabsq_m_s8(int8x16_t inactive, int8x16_t a, mve_pred16_t p) in test_vabsq_m_s8() [all …]
|
D | vector-shift-var.c | 12 int8x16_t test_vshlq_s8(int8x16_t a, int8x16_t b) in test_vshlq_s8() 54 uint8x16_t test_vshlq_u8(uint8x16_t a, int8x16_t b) in test_vshlq_u8() 96 int8x16_t test_vshlq_r_s8(int8x16_t a, int32_t b) in test_vshlq_r_s8() 180 int8x16_t test_vqshlq_s8(int8x16_t a, int8x16_t b) in test_vqshlq_s8() 222 uint8x16_t test_vqshlq_u8(uint8x16_t a, int8x16_t b) in test_vqshlq_u8() 264 int8x16_t test_vqshlq_r_s8(int8x16_t a, int32_t b) in test_vqshlq_r_s8() 348 int8x16_t test_vrshlq_s8(int8x16_t a, int8x16_t b) in test_vrshlq_s8() 390 uint8x16_t test_vrshlq_u8(uint8x16_t a, int8x16_t b) in test_vrshlq_u8() 432 int8x16_t test_vrshlq_n_s8(int8x16_t a, int32_t b) in test_vrshlq_n_s8() 516 int8x16_t test_vqrshlq_s8(int8x16_t a, int8x16_t b) in test_vqrshlq_s8() [all …]
|
D | vqrdmulhq.c | 12 int8x16_t test_vqrdmulhq_s8(int8x16_t a, int8x16_t b) in test_vqrdmulhq_s8() 56 int8x16_t test_vqrdmulhq_m_s8(int8x16_t inactive, int8x16_t a, int8x16_t b, mve_pred16_t p) in test_vqrdmulhq_m_s8() 104 int8x16_t test_vqrdmulhq_n_s8(int8x16_t a, int8_t b) in test_vqrdmulhq_n_s8() 154 int8x16_t test_vqrdmulhq_m_n_s8(int8x16_t inactive, int8x16_t a, int8_t b, mve_pred16_t p) in test_vqrdmulhq_m_n_s8()
|
D | vqdmulhq.c | 12 int8x16_t test_vqdmulhq_s8(int8x16_t a, int8x16_t b) in test_vqdmulhq_s8() 56 int8x16_t test_vqdmulhq_m_s8(int8x16_t inactive, int8x16_t a, int8x16_t b, mve_pred16_t p) in test_vqdmulhq_m_s8() 104 int8x16_t test_vqdmulhq_n_s8(int8x16_t a, int8_t b) in test_vqdmulhq_n_s8() 154 int8x16_t test_vqdmulhq_m_n_s8(int8x16_t inactive, int8x16_t a, int8_t b, mve_pred16_t p) in test_vqdmulhq_m_n_s8()
|
D | vrev.c | 12 int8x16_t test_vrev16q_s8(int8x16_t a) in test_vrev16q_s8() 40 int8x16_t test_vrev32q_s8(int8x16_t a) in test_vrev32q_s8() 110 int8x16_t test_vrev64q_s8(int8x16_t a) in test_vrev64q_s8() 224 int8x16_t test_vrev16q_m_s8(int8x16_t inactive, int8x16_t a, mve_pred16_t p) in test_vrev16q_m_s8() 272 int8x16_t test_vrev32q_m_s8(int8x16_t inactive, int8x16_t a, mve_pred16_t p) in test_vrev32q_m_s8() 368 int8x16_t test_vrev64q_m_s8(int8x16_t inactive, int8x16_t a, mve_pred16_t p) in test_vrev64q_m_s8() 464 int8x16_t test_vrev16q_x_s8(int8x16_t a, mve_pred16_t p) in test_vrev16q_x_s8() 512 int8x16_t test_vrev32q_x_s8(int8x16_t a, mve_pred16_t p) in test_vrev32q_x_s8() 608 int8x16_t test_vrev64q_x_s8(int8x16_t a, mve_pred16_t p) in test_vrev64q_x_s8()
|
D | vmldav.c | 12 int32_t test_vmladavaq_s8(int32_t a, int8x16_t b, int8x16_t c) { in test_vmladavaq_s8() 90 int32_t test_vmladavaxq_s8(int32_t a, int8x16_t b, int8x16_t c) { in test_vmladavaxq_s8() 129 int32_t test_vmlsdavaq_s8(int32_t a, int8x16_t b, int8x16_t c) { in test_vmlsdavaq_s8() 168 int32_t test_vmlsdavaxq_s8(int32_t a, int8x16_t b, int8x16_t c) { in test_vmlsdavaxq_s8() 209 int32_t test_vmladavaq_p_s8(int32_t a, int8x16_t b, int8x16_t c, mve_pred16_t p) { in test_vmladavaq_p_s8() 299 int32_t test_vmladavaxq_p_s8(int32_t a, int8x16_t b, int8x16_t c, mve_pred16_t p) { in test_vmladavaxq_p_s8() 344 int32_t test_vmlsdavaq_p_s8(int32_t a, int8x16_t b, int8x16_t c, mve_pred16_t p) { in test_vmlsdavaq_p_s8() 389 int32_t test_vmlsdavaxq_p_s8(int32_t a, int8x16_t b, int8x16_t c, mve_pred16_t p) { in test_vmlsdavaxq_p_s8() 432 int32_t test_vmladavq_s8(int8x16_t a, int8x16_t b) { in test_vmladavq_s8() 510 int32_t test_vmladavxq_s8(int8x16_t a, int8x16_t b) { in test_vmladavxq_s8() [all …]
|
D | ternary.c | 129 int8x16_t test_vmlaq_n_s8(int8x16_t a, int8x16_t b, int8_t c) { in test_vmlaq_n_s8() 225 int8x16_t test_vmlasq_n_s8(int8x16_t a, int8x16_t b, int8_t c) { in test_vmlasq_n_s8() 319 int8x16_t test_vqdmlahq_n_s8(int8x16_t a, int8x16_t b, int8_t c) { in test_vqdmlahq_n_s8() 360 int8x16_t test_vqdmlashq_n_s8(int8x16_t m1, int8x16_t m2, int8_t add) { in test_vqdmlashq_n_s8() 401 int8x16_t test_vqrdmlahq_n_s8(int8x16_t a, int8x16_t b, int8_t c) { in test_vqrdmlahq_n_s8() 442 int8x16_t test_vqrdmlashq_n_s8(int8x16_t a, int8x16_t b, int8_t c) { in test_vqrdmlashq_n_s8() 615 int8x16_t test_vmlaq_m_n_s8(int8x16_t a, int8x16_t b, int8_t c, mve_pred16_t p) { in test_vmlaq_m_n_s8() 709 int8x16_t test_vmlasq_m_n_s8(int8x16_t a, int8x16_t b, int8_t c, mve_pred16_t p) { in test_vmlasq_m_n_s8() 803 int8x16_t test_vqdmlahq_m_n_s8(int8x16_t a, int8x16_t b, int8_t c, mve_pred16_t p) { in test_vqdmlahq_m_n_s8() 850 int8x16_t test_vqdmlashq_m_n_s8(int8x16_t m1, int8x16_t m2, int8_t add, mve_pred16_t p) { in test_vqdmlashq_m_n_s8() [all …]
|
D | vector-shift-imm.c | 12 int8x16_t test_vshlq_n_s8(int8x16_t a) in test_vshlq_n_s8() 54 int8x16_t test_vshlq_n_s8_trivial(int8x16_t a) in test_vshlq_n_s8_trivial() 180 int8x16_t test_vshrq_n_s8(int8x16_t a) in test_vshrq_n_s8() 222 int8x16_t test_vshrq_n_s8_trivial(int8x16_t a) in test_vshrq_n_s8_trivial() 347 int8x16_t test_vshlq_m_n_s8(int8x16_t inactive, int8x16_t a, mve_pred16_t p) in test_vshlq_m_n_s8() 443 int8x16_t test_vshrq_m_n_s8(int8x16_t inactive, int8x16_t a, mve_pred16_t p) in test_vshrq_m_n_s8() 539 int8x16_t test_vshlq_x_n_s8(int8x16_t a, mve_pred16_t p) in test_vshlq_x_n_s8() 635 int8x16_t test_vshrq_x_n_s8(int8x16_t a, mve_pred16_t p) in test_vshrq_x_n_s8() 729 int8x16_t test_vqshlq_n_s8(int8x16_t a) in test_vqshlq_n_s8() 813 uint8x16_t test_vqshluq_n_s8(int8x16_t a) in test_vqshluq_n_s8() [all …]
|
D | vector-shift-imm-dyadic.c | 12 int8x16_t test_vshrnbq_n_s16(int8x16_t a, int16x8_t b) in test_vshrnbq_n_s16() 68 int8x16_t test_vshrntq_n_s16(int8x16_t a, int16x8_t b) in test_vshrntq_n_s16() 126 int8x16_t test_vshrnbq_m_n_s16(int8x16_t a, int16x8_t b, mve_pred16_t p) in test_vshrnbq_m_n_s16() 190 int8x16_t test_vshrntq_m_n_s16(int8x16_t a, int16x8_t b, mve_pred16_t p) in test_vshrntq_m_n_s16() 252 int8x16_t test_vrshrnbq_n_s16(int8x16_t a, int16x8_t b) in test_vrshrnbq_n_s16() 308 int8x16_t test_vrshrntq_n_s16(int8x16_t a, int16x8_t b) in test_vrshrntq_n_s16() 366 int8x16_t test_vrshrnbq_m_n_s16(int8x16_t a, int16x8_t b, mve_pred16_t p) in test_vrshrnbq_m_n_s16() 430 int8x16_t test_vrshrntq_m_n_s16(int8x16_t a, int16x8_t b, mve_pred16_t p) in test_vrshrntq_m_n_s16() 492 int8x16_t test_vqshrnbq_n_s16(int8x16_t a, int16x8_t b) in test_vqshrnbq_n_s16() 548 int8x16_t test_vqshrntq_n_s16(int8x16_t a, int16x8_t b) in test_vqshrntq_n_s16() [all …]
|
D | vclz.c | 12 int8x16_t test_vclzq_s8(int8x16_t a) in test_vclzq_s8() 96 int8x16_t test_vclsq_s8(int8x16_t a) in test_vclsq_s8() 140 int8x16_t test_vclsq_m_s8(int8x16_t inactive, int8x16_t a, mve_pred16_t p) in test_vclsq_m_s8() 188 int8x16_t test_vclzq_m_s8(int8x16_t inactive, int8x16_t a, mve_pred16_t p) in test_vclzq_m_s8() 284 int8x16_t test_vclsq_x_s8(int8x16_t a, mve_pred16_t p) in test_vclsq_x_s8() 332 int8x16_t test_vclzq_x_s8(int8x16_t a, mve_pred16_t p) in test_vclzq_x_s8()
|
D | vcaddq.c | 54 int8x16_t test_vcaddq_rot90_s8(int8x16_t a, int8x16_t b) in test_vcaddq_rot90_s8() 166 int8x16_t test_vcaddq_rot270_s8(int8x16_t a, int8x16_t b) in test_vcaddq_rot270_s8() 287 int8x16_t test_vcaddq_rot90_m_s8(int8x16_t inactive, int8x16_t a, int8x16_t b, mve_pred16_t p) in test_vcaddq_rot90_m_s8() 415 int8x16_t test_vcaddq_rot270_m_s8(int8x16_t inactive, int8x16_t a, int8x16_t b, mve_pred16_t p) in test_vcaddq_rot270_m_s8() 543 int8x16_t test_vcaddq_rot90_x_s8(int8x16_t a, int8x16_t b, mve_pred16_t p) in test_vcaddq_rot90_x_s8() 671 int8x16_t test_vcaddq_rot270_x_s8(int8x16_t a, int8x16_t b, mve_pred16_t p) in test_vcaddq_rot270_x_s8()
|
D | vabdq.c | 12 int8x16_t test_vabdq_s8(int8x16_t a, int8x16_t b) in test_vabdq_s8() 72 int8x16_t test_vabdq_m_s8(int8x16_t inactive, int8x16_t a, int8x16_t b, mve_pred16_t p) in test_vabdq_m_s8()
|
D | reinterpret.c | 90 float16x8_t test_vreinterpretq_f16_s8(int8x16_t a) in test_vreinterpretq_f16_s8() 251 float32x4_t test_vreinterpretq_f32_s8(int8x16_t a) in test_vreinterpretq_f32_s8() 412 int16x8_t test_vreinterpretq_s16_s8(int8x16_t a) in test_vreinterpretq_s16_s8() 572 int32x4_t test_vreinterpretq_s32_s8(int8x16_t a) in test_vreinterpretq_s32_s8() 737 int64x2_t test_vreinterpretq_s64_s8(int8x16_t a) in test_vreinterpretq_s64_s8() 826 int8x16_t test_vreinterpretq_s8_f16(float16x8_t a) in test_vreinterpretq_s8_f16() 845 int8x16_t test_vreinterpretq_s8_f32(float32x4_t a) in test_vreinterpretq_s8_f32() 864 int8x16_t test_vreinterpretq_s8_s16(int16x8_t a) in test_vreinterpretq_s8_s16() 883 int8x16_t test_vreinterpretq_s8_s32(int32x4_t a) in test_vreinterpretq_s8_s32() 902 int8x16_t test_vreinterpretq_s8_s64(int64x2_t a) in test_vreinterpretq_s8_s64() [all …]
|
D | vmovn.c | 25 int8x16_t test_vmovnbq_s16(int8x16_t a, int16x8_t b) in test_vmovnbq_s16() 123 int8x16_t test_vmovntq_s16(int8x16_t a, int16x8_t b) in test_vmovntq_s16() 215 int8x16_t test_vmovnbq_m_s16(int8x16_t a, int16x8_t b, mve_pred16_t p) in test_vmovnbq_m_s16() 307 int8x16_t test_vmovntq_m_s16(int8x16_t a, int16x8_t b, mve_pred16_t p) in test_vmovntq_m_s16()
|
/external/clang/test/CodeGen/ |
D | aarch64-neon-tbl.c | 19 int8x8_t test_vqtbl1_s8(int8x16_t a, int8x8_t b) { in test_vqtbl1_s8() 175 int8x16_t test_vqtbl1q_s8(int8x16_t a, int8x16_t b) { in test_vqtbl1q_s8() 196 int8x16_t test_vqtbl2q_s8(int8x16x2_t a, int8x16_t b) { in test_vqtbl2q_s8() 220 int8x16_t test_vqtbl3q_s8(int8x16x3_t a, int8x16_t b) { in test_vqtbl3q_s8() 247 int8x16_t test_vqtbl4q_s8(int8x16x4_t a, int8x16_t b) { in test_vqtbl4q_s8() 351 int8x8_t test_vqtbx1_s8(int8x8_t a, int8x16_t b, int8x8_t c) { in test_vqtbx1_s8() 430 int8x16_t test_vqtbx1q_s8(int8x16_t a, int8x16_t b, int8x16_t c) { in test_vqtbx1q_s8() 451 int8x16_t test_vqtbx2q_s8(int8x16_t a, int8x16x2_t b, int8x16_t c) { in test_vqtbx2q_s8() 475 int8x16_t test_vqtbx3q_s8(int8x16_t a, int8x16x3_t b, int8x16_t c) { in test_vqtbx3q_s8() 502 int8x16_t test_vqtbx4q_s8(int8x16_t a, int8x16x4_t b, int8x16_t c) { in test_vqtbx4q_s8()
|
/external/llvm-project/clang/test/CodeGen/ |
D | aarch64-neon-tbl.c | 19 int8x8_t test_vqtbl1_s8(int8x16_t a, uint8x8_t b) { in test_vqtbl1_s8() 175 int8x16_t test_vqtbl1q_s8(int8x16_t a, int8x16_t b) { in test_vqtbl1q_s8() 196 int8x16_t test_vqtbl2q_s8(int8x16x2_t a, int8x16_t b) { in test_vqtbl2q_s8() 220 int8x16_t test_vqtbl3q_s8(int8x16x3_t a, int8x16_t b) { in test_vqtbl3q_s8() 247 int8x16_t test_vqtbl4q_s8(int8x16x4_t a, int8x16_t b) { in test_vqtbl4q_s8() 351 int8x8_t test_vqtbx1_s8(int8x8_t a, int8x16_t b, uint8x8_t c) { in test_vqtbx1_s8() 430 int8x16_t test_vqtbx1q_s8(int8x16_t a, int8x16_t b, uint8x16_t c) { in test_vqtbx1q_s8() 451 int8x16_t test_vqtbx2q_s8(int8x16_t a, int8x16x2_t b, int8x16_t c) { in test_vqtbx2q_s8() 475 int8x16_t test_vqtbx3q_s8(int8x16_t a, int8x16x3_t b, int8x16_t c) { in test_vqtbx3q_s8() 502 int8x16_t test_vqtbx4q_s8(int8x16_t a, int8x16x4_t b, int8x16_t c) { in test_vqtbx4q_s8()
|
D | aarch64-v8.6a-neon-intrinsics.c | 13 int32x4_t test_vmmlaq_s32(int32x4_t r, int8x16_t a, int8x16_t b) { in test_vmmlaq_s32() 27 int32x4_t test_vusmmlaq_s32(int32x4_t r, uint8x16_t a, int8x16_t b) { in test_vusmmlaq_s32() 73 int32x2_t test_vusdot_laneq_s32(int32x2_t r, uint8x8_t a, int8x16_t b) { in test_vusdot_laneq_s32() 93 int32x4_t test_vusdotq_s32(int32x4_t r, uint8x16_t a, int8x16_t b) { in test_vusdotq_s32() 119 int32x4_t test_vsudotq_lane_s32(int32x4_t r, int8x16_t a, uint8x8_t b) { in test_vsudotq_lane_s32() 132 int32x4_t test_vusdotq_laneq_s32(int32x4_t r, uint8x16_t a, int8x16_t b) { in test_vusdotq_laneq_s32() 145 int32x4_t test_vsudotq_laneq_s32(int32x4_t r, int8x16_t a, uint8x16_t b) { in test_vsudotq_laneq_s32()
|
D | aarch64-neon-misc.c | 66 uint8x16_t test_vceqzq_s8(int8x16_t a) { in test_vceqzq_s8() 276 uint8x16_t test_vcgezq_s8(int8x16_t a) { in test_vcgezq_s8() 382 uint8x16_t test_vclezq_s8(int8x16_t a) { in test_vclezq_s8() 488 uint8x16_t test_vcgtzq_s8(int8x16_t a) { in test_vcgtzq_s8() 594 uint8x16_t test_vcltzq_s8(int8x16_t a) { in test_vcltzq_s8() 685 int8x16_t test_vrev16q_s8(int8x16_t a) { in test_vrev16q_s8() 748 int8x16_t test_vrev32q_s8(int8x16_t a) { in test_vrev32q_s8() 853 int8x16_t test_vrev64q_s8(int8x16_t a) { in test_vrev64q_s8() 962 int16x8_t test_vpaddlq_s8(int8x16_t a) { in test_vpaddlq_s8() 1068 int16x8_t test_vpadalq_s8(int16x8_t a, int8x16_t b) { in test_vpadalq_s8() [all …]
|
/external/XNNPACK/src/qs8-igemm/gen/ |
D | 1x16c4-minmax-neondot.c | 68 const int8x16_t vb0123x0123 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot() 69 const int8x16_t vb0123x4567 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot() 70 const int8x16_t vb0123x89AB = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot() 71 const int8x16_t vb0123xCDEF = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot() 72 const int8x16_t vb4567x0123 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot() 73 const int8x16_t vb4567x4567 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot() 74 const int8x16_t vb4567x89AB = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot() 75 const int8x16_t vb4567xCDEF = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot() 95 const int8x16_t vb0123x0123 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot() 96 const int8x16_t vb0123x4567 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_igemm_minmax_ukernel_1x16c4__neondot() [all …]
|
/external/XNNPACK/src/qs8-gemm/gen/ |
D | 1x16c4-minmax-neondot.c | 59 const int8x16_t vb0123x0123 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_1x16c4__neondot() 60 const int8x16_t vb0123x4567 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_1x16c4__neondot() 61 const int8x16_t vb0123x89AB = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_1x16c4__neondot() 62 const int8x16_t vb0123xCDEF = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_1x16c4__neondot() 63 const int8x16_t vb4567x0123 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_1x16c4__neondot() 64 const int8x16_t vb4567x4567 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_1x16c4__neondot() 65 const int8x16_t vb4567x89AB = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_1x16c4__neondot() 66 const int8x16_t vb4567xCDEF = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_1x16c4__neondot() 86 const int8x16_t vb0123x0123 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_1x16c4__neondot() 87 const int8x16_t vb0123x4567 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_1x16c4__neondot() [all …]
|
D | 4x16c4-minmax-neondot.c | 92 const int8x16_t vb0123x0123 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_4x16c4__neondot() 93 const int8x16_t vb0123x4567 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_4x16c4__neondot() 94 const int8x16_t vb0123x89AB = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_4x16c4__neondot() 95 const int8x16_t vb0123xCDEF = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_4x16c4__neondot() 96 const int8x16_t vb4567x0123 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_4x16c4__neondot() 97 const int8x16_t vb4567x4567 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_4x16c4__neondot() 98 const int8x16_t vb4567x89AB = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_4x16c4__neondot() 99 const int8x16_t vb4567xCDEF = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_4x16c4__neondot() 146 const int8x16_t vb0123x0123 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_4x16c4__neondot() 147 const int8x16_t vb0123x4567 = vld1q_s8(w); w = (const void*) ((const int8_t*) w + 16); in xnn_qs8_gemm_minmax_ukernel_4x16c4__neondot() [all …]
|
/external/tensorflow/tensorflow/lite/kernels/internal/optimized/ |
D | depthwiseconv_3x3_filter_common.h | 37 inline int8x16_t vqtbl4q_s8(int8x16x4_t a, int8x16_t b) { in vqtbl4q_s8() 41 const int8x16_t high_bits = vshrq_n_s8(b, 4); in vqtbl4q_s8() 42 int8x16_t deleted_bit_3 = b; in vqtbl4q_s8() 52 const int8x16_t output_for_lower = in vqtbl4q_s8() 61 const int8x16_t output_for_higher = in vqtbl4q_s8() 66 int8x16_t output = vbslq_s8(mask, output_for_higher, output_for_lower); in vqtbl4q_s8() 79 inline void vzipq_s8_in_place(int8x16_t* a, int8x16_t* b) { in vzipq_s8_in_place() 86 inline void vzipq_s8x2_in_place(int8x16_t* a, int8x16_t* b) { in vzipq_s8x2_in_place() 95 inline void vtrn1_s8x2_in_place(int8x16_t* a, int8x16_t* b) { in vtrn1_s8x2_in_place() 103 inline int8x16_t vzip1q_s8(int8x16_t a, int8x16_t b) { in vzip1q_s8() [all …]
|
/external/webp/src/dsp/ |
D | dec_neon.c | 440 static int8x16_t FlipSign_NEON(const uint8x16_t v) { in FlipSign_NEON() 445 static uint8x16_t FlipSignBack_NEON(const int8x16_t v) { in FlipSignBack_NEON() 446 const int8x16_t sign_bit = vdupq_n_s8(0x80); in FlipSignBack_NEON() 450 static int8x16_t GetBaseDelta_NEON(const int8x16_t p1, const int8x16_t p0, in GetBaseDelta_NEON() 451 const int8x16_t q0, const int8x16_t q1) { in GetBaseDelta_NEON() 452 const int8x16_t q0_p0 = vqsubq_s8(q0, p0); // (q0-p0) in GetBaseDelta_NEON() 453 const int8x16_t p1_q1 = vqsubq_s8(p1, q1); // (p1-q1) in GetBaseDelta_NEON() 454 const int8x16_t s1 = vqaddq_s8(p1_q1, q0_p0); // (p1-q1) + 1 * (q0 - p0) in GetBaseDelta_NEON() 455 const int8x16_t s2 = vqaddq_s8(q0_p0, s1); // (p1-q1) + 2 * (q0 - p0) in GetBaseDelta_NEON() 456 const int8x16_t s3 = vqaddq_s8(q0_p0, s2); // (p1-q1) + 3 * (q0 - p0) in GetBaseDelta_NEON() [all …]
|