/external/XNNPACK/src/qc8-igemm/gen/ |
D | 6x16c4-minmax-fp32-neondot.c | 259 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16c4__neondot() local 291 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16c4__neondot() 313 vacc4xCDEF = vcvtnq_s32_f32(vfpacc4xCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16c4__neondot()
|
D | 8x16c4-minmax-fp32-neondot.c | 311 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_8x16c4__neondot() local 357 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_8x16c4__neondot() 381 vacc4xCDEF = vcvtnq_s32_f32(vfpacc4xCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_8x16c4__neondot()
|
D | 6x16-minmax-fp32-neon-mlal-lane-prfm.c | 640 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neon_mlal_lane_prfm() local 672 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neon_mlal_lane_prfm() 695 vacc4xCDEF = vreinterpretq_s32_f32(vaddq_f32(vfpacc4xCDEF, vmagic_bias)); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neon_mlal_lane_prfm()
|
D | 6x16-minmax-fp32-neonv8-mlal-lane.c | 639 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane() local 671 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane() 693 vacc4xCDEF = vcvtnq_s32_f32(vfpacc4xCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane()
|
D | 6x16-minmax-fp32-neonv8-mlal-lane-prfm.c | 641 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane_prfm() local 673 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane_prfm() 695 vacc4xCDEF = vcvtnq_s32_f32(vfpacc4xCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane_prfm()
|
D | 6x16-minmax-fp32-neon-mlal-lane.c | 638 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neon_mlal_lane() local 670 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neon_mlal_lane() 693 vacc4xCDEF = vreinterpretq_s32_f32(vaddq_f32(vfpacc4xCDEF, vmagic_bias)); in xnn_qc8_igemm_minmax_fp32_ukernel_6x16__neon_mlal_lane()
|
/external/XNNPACK/src/qc8-gemm/gen/ |
D | 6x16c4-minmax-fp32-neondot.c | 239 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16c4__neondot() local 271 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16c4__neondot() 293 vacc4xCDEF = vcvtnq_s32_f32(vfpacc4xCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16c4__neondot()
|
D | 8x16c4-minmax-fp32-neondot.c | 287 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_8x16c4__neondot() local 333 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_8x16c4__neondot() 357 vacc4xCDEF = vcvtnq_s32_f32(vfpacc4xCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_8x16c4__neondot()
|
D | 6x16-minmax-fp32-neon-mlal-lane-prfm.c | 616 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neon_mlal_lane_prfm() local 648 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neon_mlal_lane_prfm() 671 vacc4xCDEF = vreinterpretq_s32_f32(vaddq_f32(vfpacc4xCDEF, vmagic_bias)); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neon_mlal_lane_prfm()
|
D | 6x16-minmax-fp32-neonv8-mlal-lane-prfm.c | 617 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane_prfm() local 649 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane_prfm() 671 vacc4xCDEF = vcvtnq_s32_f32(vfpacc4xCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane_prfm()
|
D | 6x16-minmax-fp32-neonv8-mlal-lane.c | 615 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane() local 647 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane() 669 vacc4xCDEF = vcvtnq_s32_f32(vfpacc4xCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neonv8_mlal_lane()
|
D | 6x16-minmax-fp32-neon-mlal-lane.c | 614 float32x4_t vfpacc4xCDEF = vcvtq_f32_s32(vacc4xCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neon_mlal_lane() local 646 vfpacc4xCDEF = vmulq_f32(vfpacc4xCDEF, vscaleCDEF); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neon_mlal_lane() 669 vacc4xCDEF = vreinterpretq_s32_f32(vaddq_f32(vfpacc4xCDEF, vmagic_bias)); in xnn_qc8_gemm_minmax_fp32_ukernel_6x16__neon_mlal_lane()
|