/external/XNNPACK/src/qc8-igemm/gen/ |
D | 4x16c4-minmax-fp32-neondot.c | 202 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16c4__neondot() local 219 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16c4__neondot() 240 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16c4__neondot()
|
D | 6x16c4-minmax-fp32-neondot.c | 254 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16c4__neondot() local 283 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16c4__neondot() 308 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16c4__neondot()
|
D | 8x16c4-minmax-fp32-neondot.c | 306 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_8x16c4__neondot() local 347 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_8x16c4__neondot() 376 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_8x16c4__neondot()
|
D | 4x16-minmax-fp32-neon-mlal-lane.c | 481 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() local 498 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() 520 vacc3x89AB = vreinterpretq_s32_f32(vaddq_f32(vfpacc3x89AB, vmagic_bias)); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane()
|
D | 4x16-minmax-fp32-neon-mlal-lane-prfm.c | 483 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane_prfm() local 500 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane_prfm() 522 vacc3x89AB = vreinterpretq_s32_f32(vaddq_f32(vfpacc3x89AB, vmagic_bias)); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane_prfm()
|
D | 4x16-minmax-fp32-neonv8-mlal-lane.c | 482 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() local 499 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() 520 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane()
|
D | 4x16-minmax-fp32-neonv8-mlal-lane-prfm.c | 484 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane_prfm() local 501 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane_prfm() 522 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane_prfm()
|
D | 6x16-minmax-fp32-neon-mlal-lane-prfm.c | 635 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neon_mlal_lane_prfm() local 664 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neon_mlal_lane_prfm() 690 vacc3x89AB = vreinterpretq_s32_f32(vaddq_f32(vfpacc3x89AB, vmagic_bias)); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neon_mlal_lane_prfm()
|
/external/XNNPACK/src/qc8-gemm/gen/ |
D | 4x16c4-minmax-fp32-neondot.c | 186 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16c4__neondot() local 203 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16c4__neondot() 224 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16c4__neondot()
|
D | 6x16c4-minmax-fp32-neondot.c | 234 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16c4__neondot() local 263 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16c4__neondot() 288 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16c4__neondot()
|
D | 8x16c4-minmax-fp32-neondot.c | 282 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_8x16c4__neondot() local 323 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_8x16c4__neondot() 352 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_8x16c4__neondot()
|
D | 4x16-minmax-fp32-neon-mlal-lane.c | 461 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() local 478 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() 500 vacc3x89AB = vreinterpretq_s32_f32(vaddq_f32(vfpacc3x89AB, vmagic_bias)); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane()
|
D | 4x16-minmax-fp32-neon-mlal-lane-prfm.c | 463 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane_prfm() local 480 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane_prfm() 502 vacc3x89AB = vreinterpretq_s32_f32(vaddq_f32(vfpacc3x89AB, vmagic_bias)); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane_prfm()
|
D | 4x16-minmax-fp32-neonv8-mlal-lane-prfm.c | 464 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane_prfm() local 481 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane_prfm() 502 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane_prfm()
|
D | 4x16-minmax-fp32-neonv8-mlal-lane.c | 462 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() local 479 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() 500 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qc8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane()
|
/external/XNNPACK/src/qu8-igemm/gen/ |
D | 4x16c4-minmax-fp32-neondot.c | 244 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qu8_igemm_minmax_fp32_ukernel_4x16c4__neondot() local 262 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale); in xnn_qu8_igemm_minmax_fp32_ukernel_4x16c4__neondot() 279 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qu8_igemm_minmax_fp32_ukernel_4x16c4__neondot()
|
D | 4x16-minmax-fp32-neon-mlal-lane.c | 482 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qu8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() local 500 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale); in xnn_qu8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() 518 vacc3x89AB = vreinterpretq_s32_f32(vaddq_f32(vfpacc3x89AB, vmagic_bias)); in xnn_qu8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane()
|
D | 4x16-minmax-fp32-neonv8-mlal-lane.c | 483 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qu8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() local 501 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale); in xnn_qu8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() 518 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qu8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane()
|
/external/XNNPACK/src/qu8-gemm/gen/ |
D | 4x16c4-minmax-fp32-neondot.c | 226 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qu8_gemm_minmax_fp32_ukernel_4x16c4__neondot() local 244 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale); in xnn_qu8_gemm_minmax_fp32_ukernel_4x16c4__neondot() 261 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qu8_gemm_minmax_fp32_ukernel_4x16c4__neondot()
|
D | 4x16-minmax-fp32-neon-mlal-lane.c | 462 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qu8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() local 480 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale); in xnn_qu8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() 498 vacc3x89AB = vreinterpretq_s32_f32(vaddq_f32(vfpacc3x89AB, vmagic_bias)); in xnn_qu8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane()
|
D | 4x16-minmax-fp32-neonv8-mlal-lane.c | 463 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qu8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() local 481 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale); in xnn_qu8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() 498 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qu8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane()
|
/external/XNNPACK/src/qs8-igemm/gen/ |
D | 4x16-minmax-fp32-neon-mlal-lane.c | 481 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qs8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() local 499 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale); in xnn_qs8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() 517 vacc3x89AB = vreinterpretq_s32_f32(vaddq_f32(vfpacc3x89AB, vmagic_bias)); in xnn_qs8_igemm_minmax_fp32_ukernel_4x16__neon_mlal_lane()
|
D | 4x16-minmax-fp32-neonv8-mlal-lane.c | 482 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qs8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() local 500 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale); in xnn_qs8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() 517 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qs8_igemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane()
|
/external/XNNPACK/src/qs8-gemm/gen/ |
D | 4x16-minmax-fp32-neonv8-mlal-lane.c | 462 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qs8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() local 480 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale); in xnn_qs8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane() 497 vacc3x89AB = vcvtnq_s32_f32(vfpacc3x89AB); in xnn_qs8_gemm_minmax_fp32_ukernel_4x16__neonv8_mlal_lane()
|
D | 4x16-minmax-fp32-neon-mlal-lane.c | 461 float32x4_t vfpacc3x89AB = vcvtq_f32_s32(vacc3x89AB); in xnn_qs8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() local 479 vfpacc3x89AB = vmulq_f32(vfpacc3x89AB, vscale); in xnn_qs8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane() 497 vacc3x89AB = vreinterpretq_s32_f32(vaddq_f32(vfpacc3x89AB, vmagic_bias)); in xnn_qs8_gemm_minmax_fp32_ukernel_4x16__neon_mlal_lane()
|