Hugging Face's logo

Androidonnxfork
/
test

Model card Files Community
5
test
File size: 7,710 Bytes 
8b7c501
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
 
// Copyright 2021 Google LLC
//
// This source code is licensed under the BSD-style license found in the
// LICENSE file in the root directory of this source tree.

$assert BATCH_TILE % 8 == 0
$assert BATCH_TILE >= 8
$SIMD_TILE = BATCH_TILE // 8
$ABC = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ"
#include <assert.h>

#include <wasm_simd128.h>

#include <xnnpack/common.h>
#include <xnnpack/vcvt.h>


$WASM_V32X4_LANESELECT = "wasm_i32x4_relaxed_laneselect" if RELAXED else "wasm_v128_bitselect"
$ISA = "wasmrelaxedsimd" if RELAXED else "wasmsimd"
void xnn_f16_f32_vcvt_ukernel__${ISA}_int16_x${BATCH_TILE}(
    size_t batch,
    const void* input,
    float* output,
    const union xnn_f16_f32_cvt_params params[restrict XNN_MIN_ELEMENTS(1)]) XNN_OOB_READS
{
  assert(batch != 0);
  assert(batch % sizeof(uint16_t) == 0);
  assert(input != NULL);
  assert(output != NULL);

  const v128_t vsign_mask = wasm_v128_load64_splat(params->wasmsimd_int16.sign_mask);
  const v128_t vexp_offset = wasm_v128_load64_splat(params->wasmsimd_int16.exp_offset);
  const v128_t vexp_scale = wasm_v128_load64_splat(params->wasmsimd_int16.exp_scale);
  const v128_t vmagic_mask = wasm_v128_load64_splat(params->wasmsimd_int16.magic_mask);
  const v128_t vmagic_bias = wasm_v128_load64_splat(params->wasmsimd_int16.magic_bias);
  const v128_t vdenorm_cutoff = wasm_v128_load64_splat(params->wasmsimd_int16.denorm_cutoff);

  const uint16_t* i = (const uint16_t*) input;
  $if BATCH_TILE > 8:
    for (; batch >= ${BATCH_TILE} * sizeof(uint16_t); batch -= ${BATCH_TILE} * sizeof(uint16_t)) {
      const v128_t vh0 = wasm_v128_load(i);
      $for N in range(1, SIMD_TILE):
        const v128_t vh${N} = wasm_v128_load(i + ${N * 8});
      i += ${BATCH_TILE};

      $for N in range(SIMD_TILE):
        const v128_t vsign${N} = wasm_v128_and(vh${N}, vsign_mask);

      $for N in range(SIMD_TILE):
        const v128_t vnonsign${N} = wasm_v128_xor(vh${N}, vsign${N});

      $for N in range(SIMD_TILE):
        const v128_t vprenorm${N*2} = wasm_i16x8_shl(vnonsign${N}, 13);
        const v128_t vprenorm${N*2+1} = wasm_i16x8_add(wasm_u16x8_shr(vnonsign${N}, 3), vexp_offset);

      $for N in range(SIMD_TILE):
        const v128_t vnorm${N*2} = wasm_f32x4_mul(wasm_v16x8_shuffle(vprenorm${N*2}, vprenorm${N*2+1}, 0,  8, 1,  9, 2, 10, 3, 11), vexp_scale);
        const v128_t vnorm${N*2+1} = wasm_f32x4_mul(wasm_v16x8_shuffle(vprenorm${N*2}, vprenorm${N*2+1}, 4, 12, 5, 13, 6, 14, 7, 15), vexp_scale);

      $for N in range(SIMD_TILE):
        const v128_t vdenorm${N*2} = wasm_f32x4_sub(wasm_v16x8_shuffle(vnonsign${N}, vmagic_mask, 0,  8, 1,  9, 2, 10, 3, 11), vmagic_bias);
        const v128_t vdenorm${N*2+1} = wasm_f32x4_sub(wasm_v16x8_shuffle(vnonsign${N}, vmagic_mask, 4, 12, 5, 13, 6, 14, 7, 15), vmagic_bias);

      $for N in range(SIMD_TILE):
        const v128_t vmask${N} = wasm_i16x8_gt(vnonsign${N}, vdenorm_cutoff);
      const v128_t vzero = wasm_i16x8_const_splat(0);

      $for N in range(SIMD_TILE):
        const v128_t vxmask${N*2} = wasm_i32x4_extend_low_i16x8(vmask${N});
        const v128_t vxmask${N*2+1} = wasm_i32x4_extend_high_i16x8(vmask${N});

      $for N in range(SIMD_TILE):
        const v128_t vabsf${N*2} = ${WASM_V32X4_LANESELECT}(vnorm${N*2}, vdenorm${N*2}, vxmask${N*2});
        const v128_t vsignf${N*2} = wasm_v16x8_shuffle(vzero, vsign${N}, 0,  8, 1,  9, 2, 10, 3, 11);
        const v128_t vabsf${N*2+1} = ${WASM_V32X4_LANESELECT}(vnorm${N*2+1}, vdenorm${N*2+1}, vxmask${N*2+1});
        const v128_t vsignf${N*2+1} = wasm_v16x8_shuffle(vzero, vsign${N}, 4, 12, 5, 13, 6, 14, 7, 15);

      $for N in range(SIMD_TILE):
        const v128_t vf${N*2} = wasm_v128_or(vsignf${N*2}, vabsf${N*2});
        const v128_t vf${N*2+1} = wasm_v128_or(vsignf${N*2+1}, vabsf${N*2+1});

      wasm_v128_store(output, vf0);
      $for N in range(1, 2*SIMD_TILE):
        wasm_v128_store(output + ${N*4}, vf${N});
      output += ${BATCH_TILE};
    }
  for (; batch >= 8 * sizeof(uint16_t); batch -= 8 * sizeof(uint16_t)) {
    const v128_t vh = wasm_v128_load(i);
    i += 8;

    const v128_t vsign = wasm_v128_and(vh, vsign_mask);

    const v128_t vnonsign = wasm_v128_xor(vh, vsign);

    const v128_t vprenorm_lo = wasm_i16x8_shl(vnonsign, 13);
    const v128_t vprenorm_hi = wasm_i16x8_add(wasm_u16x8_shr(vnonsign, 3), vexp_offset);

    const v128_t vnorm_lo = wasm_f32x4_mul(wasm_v16x8_shuffle(vprenorm_lo, vprenorm_hi, 0,  8, 1,  9, 2, 10, 3, 11), vexp_scale);
    const v128_t vnorm_hi = wasm_f32x4_mul(wasm_v16x8_shuffle(vprenorm_lo, vprenorm_hi, 4, 12, 5, 13, 6, 14, 7, 15), vexp_scale);

    const v128_t vdenorm_lo = wasm_f32x4_sub(wasm_v16x8_shuffle(vnonsign, vmagic_mask, 0,  8, 1,  9, 2, 10, 3, 11), vmagic_bias);
    const v128_t vdenorm_hi = wasm_f32x4_sub(wasm_v16x8_shuffle(vnonsign, vmagic_mask, 4, 12, 5, 13, 6, 14, 7, 15), vmagic_bias);

    const v128_t vmask = wasm_i16x8_gt(vnonsign, vdenorm_cutoff);
    const v128_t vzero = wasm_i16x8_const_splat(0);

    const v128_t vxmask_lo = wasm_i32x4_extend_low_i16x8(vmask);
    const v128_t vxmask_hi = wasm_i32x4_extend_high_i16x8(vmask);

    const v128_t vabsf_lo = ${WASM_V32X4_LANESELECT}(vnorm_lo, vdenorm_lo, vxmask_lo);
    const v128_t vsignf_lo = wasm_v16x8_shuffle(vzero, vsign, 0,  8, 1,  9, 2, 10, 3, 11);
    const v128_t vabsf_hi = ${WASM_V32X4_LANESELECT}(vnorm_hi, vdenorm_hi, vxmask_hi);
    const v128_t vsignf_hi = wasm_v16x8_shuffle(vzero, vsign, 4, 12, 5, 13, 6, 14, 7, 15);

    const v128_t vf_lo = wasm_v128_or(vsignf_lo, vabsf_lo);
    const v128_t vf_hi = wasm_v128_or(vsignf_hi, vabsf_hi);

    wasm_v128_store(output, vf_lo);
    wasm_v128_store(output + 4, vf_hi);
    output += 8;
  }
  if XNN_UNLIKELY(batch != 0) {
    assert(batch >= 1 * sizeof(uint16_t));
    assert(batch <= 7 * sizeof(uint16_t));
    const v128_t vh = wasm_v128_load(i);

    const v128_t vsign = wasm_v128_and(vh, vsign_mask);

    const v128_t vnonsign = wasm_v128_xor(vh, vsign);

    const v128_t vprenorm_lo = wasm_i16x8_shl(vnonsign, 13);
    const v128_t vprenorm_hi = wasm_i16x8_add(wasm_u16x8_shr(vnonsign, 3), vexp_offset);

    const v128_t vnorm_lo = wasm_f32x4_mul(wasm_v16x8_shuffle(vprenorm_lo, vprenorm_hi, 0,  8, 1,  9, 2, 10, 3, 11), vexp_scale);
    const v128_t vnorm_hi = wasm_f32x4_mul(wasm_v16x8_shuffle(vprenorm_lo, vprenorm_hi, 4, 12, 5, 13, 6, 14, 7, 15), vexp_scale);

    const v128_t vdenorm_lo = wasm_f32x4_sub(wasm_v16x8_shuffle(vnonsign, vmagic_mask, 0,  8, 1,  9, 2, 10, 3, 11), vmagic_bias);
    const v128_t vdenorm_hi = wasm_f32x4_sub(wasm_v16x8_shuffle(vnonsign, vmagic_mask, 4, 12, 5, 13, 6, 14, 7, 15), vmagic_bias);

    const v128_t vmask = wasm_i16x8_gt(vnonsign, vdenorm_cutoff);
    const v128_t vzero = wasm_i16x8_const_splat(0);

    const v128_t vxmask_lo = wasm_i32x4_extend_low_i16x8(vmask);
    const v128_t vxmask_hi = wasm_i32x4_extend_high_i16x8(vmask);

    const v128_t vabsf_lo = ${WASM_V32X4_LANESELECT}(vnorm_lo, vdenorm_lo, vxmask_lo);
    const v128_t vsignf_lo = wasm_v16x8_shuffle(vzero, vsign, 0,  8, 1,  9, 2, 10, 3, 11);
    const v128_t vabsf_hi = ${WASM_V32X4_LANESELECT}(vnorm_hi, vdenorm_hi, vxmask_hi);
    const v128_t vsignf_hi = wasm_v16x8_shuffle(vzero, vsign, 4, 12, 5, 13, 6, 14, 7, 15);

    v128_t vf = wasm_v128_or(vsignf_lo, vabsf_lo);

    if (batch & (4 * sizeof(uint16_t))) {
      wasm_v128_store(output, vf);
      output += 4;

      vf = wasm_v128_or(vsignf_hi, vabsf_hi);
    }
    if (batch & (2 * sizeof(uint16_t))) {
      wasm_v128_store64_lane(output, vf, 0);
      vf = wasm_v64x2_shuffle(vf, vf, 1, 1);
      output += 2;
    }
    if (batch & (1 * sizeof(uint16_t))) {
      wasm_v128_store32_lane(output, vf, 0);
    }
  }
}