• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 // Auto-generated file. Do not edit!
2 //   Template: src/f32-vrnd/vrndu-neon.c.in
3 //   Generator: tools/xngen
4 //
5 // Copyright 2020 Google LLC
6 //
7 // This source code is licensed under the BSD-style license found in the
8 // LICENSE file in the root directory of this source tree.
9 
10 #include <assert.h>
11 
12 #include <arm_neon.h>
13 
14 #include <xnnpack/common.h>
15 #include <xnnpack/math.h>
16 #include <xnnpack/vunary.h>
17 
18 
xnn_f32_vrndu_ukernel__neon_x8(size_t n,const float * x,float * y,const union xnn_f32_rnd_params params[restrict XNN_MIN_ELEMENTS (1)])19 void xnn_f32_vrndu_ukernel__neon_x8(
20     size_t n,
21     const float* x,
22     float* y,
23     const union xnn_f32_rnd_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_DISABLE_TSAN
24 {
25   assert(n != 0);
26   assert(n % sizeof(float) == 0);
27 
28   const float32x4_t vintegral_threshold = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000)));
29   const float32x4_t vone = vmovq_n_f32(1.0f);
30   for (; n >= 8 * sizeof(float); n -= 8 * sizeof(float)) {
31     const float32x4_t vx0123 = vld1q_f32(x); x += 4;
32     const float32x4_t vx4567 = vld1q_f32(x); x += 4;
33 
34     const int32x4_t vintx0123 = vcvtq_s32_f32(vx0123);
35     const int32x4_t vintx4567 = vcvtq_s32_f32(vx4567);
36 
37     uint32x4_t vrndmask0123 = vcaltq_f32(vx0123, vintegral_threshold);
38     uint32x4_t vrndmask4567 = vcaltq_f32(vx4567, vintegral_threshold);
39 
40     const float32x4_t vprerndx0123 = vcvtq_f32_s32(vintx0123);
41     const float32x4_t vprerndx4567 = vcvtq_f32_s32(vintx4567);
42 
43     vrndmask0123 = vbicq_u32(vrndmask0123, vmovq_n_u32(UINT32_C(0x80000000)));
44     vrndmask4567 = vbicq_u32(vrndmask4567, vmovq_n_u32(UINT32_C(0x80000000)));
45 
46     const float32x4_t vrndx0123 = vbslq_f32(vrndmask0123, vprerndx0123, vx0123);
47     const float32x4_t vrndx4567 = vbslq_f32(vrndmask4567, vprerndx4567, vx4567);
48 
49     uint32x4_t vadjmask0123 = vcgeq_f32(vrndx0123, vx0123);
50     uint32x4_t vadjmask4567 = vcgeq_f32(vrndx4567, vx4567);
51 
52     const float32x4_t vadjrndx0123 = vaddq_f32(vrndx0123, vone);
53     const float32x4_t vadjrndx4567 = vaddq_f32(vrndx4567, vone);
54 
55     vadjmask0123 = vorrq_u32(vadjmask0123, vmovq_n_u32(UINT32_C(0x80000000)));
56     vadjmask4567 = vorrq_u32(vadjmask4567, vmovq_n_u32(UINT32_C(0x80000000)));
57 
58     const float32x4_t vy0123 = vbslq_f32(vadjmask0123, vrndx0123, vadjrndx0123);
59     const float32x4_t vy4567 = vbslq_f32(vadjmask4567, vrndx4567, vadjrndx4567);
60 
61     vst1q_f32(y, vy0123); y += 4;
62     vst1q_f32(y, vy4567); y += 4;
63   }
64   for (; n >= 4 * sizeof(float); n -= 4 * sizeof(float)) {
65     const float32x4_t vx = vld1q_f32(x); x += 4;
66     const int32x4_t vintx = vcvtq_s32_f32(vx);
67     uint32x4_t vrndmask = vcaltq_f32(vx, vintegral_threshold);
68     const float32x4_t vprerndx = vcvtq_f32_s32(vintx);
69     vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000)));
70     const float32x4_t vrndx = vbslq_f32(vrndmask, vprerndx, vx);
71     uint32x4_t vadjmask = vcgeq_f32(vrndx, vx);
72     const float32x4_t vadjrndx = vaddq_f32(vrndx, vone);
73     vadjmask = vorrq_u32(vadjmask, vmovq_n_u32(UINT32_C(0x80000000)));
74     const float32x4_t vy = vbslq_f32(vadjmask, vrndx, vadjrndx);
75     vst1q_f32(y, vy); y += 4;
76   }
77   if XNN_UNLIKELY(n != 0) {
78     const float32x4_t vx = vld1q_f32(x);
79     const int32x4_t vintx = vcvtq_s32_f32(vx);
80     const float32x4_t vprerndx = vcvtq_f32_s32(vintx);
81     uint32x4_t vrndmask = vcaltq_f32(vx, vintegral_threshold);
82     vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000)));
83     const float32x4_t vrndx = vbslq_f32(vrndmask, vprerndx, vx);
84     uint32x4_t vadjmask = vcgeq_f32(vrndx, vx);
85     const float32x4_t vadjrndx = vaddq_f32(vrndx, vone);
86     vadjmask = vorrq_u32(vadjmask, vmovq_n_u32(UINT32_C(0x80000000)));
87     const float32x4_t vy = vbslq_f32(vadjmask, vrndx, vadjrndx);
88     float32x2_t vy_lo = vget_low_f32(vy);
89     if (n & (2 * sizeof(float))) {
90       vst1_f32(y, vy_lo); y += 2;
91       vy_lo = vget_high_f32(vy);
92     }
93     if (n & (1 * sizeof(float))) {
94       vst1_lane_f32(y, vy_lo, 0);
95     }
96   }
97 }
98