File size: 2,116 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
// 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 % 16 == 0
$assert BATCH_TILE >= 16
$SIMD_TILE = BATCH_TILE // 16
#include <assert.h>

#include <immintrin.h>

#include <xnnpack/common.h>
#include <xnnpack/intrinsics-polyfill.h>
#include <xnnpack/vcvt.h>


void xnn_f32_f16_vcvt_ukernel__avx512skx_x${BATCH_TILE}(
    size_t batch,
    const float* input,
    void* output,
    const union xnn_f32_f16_cvt_params params[restrict XNN_MIN_ELEMENTS(1)])
{
  assert(batch != 0);
  assert(batch % sizeof(float) == 0);
  assert(input != NULL);
  assert(output != NULL);

  uint16_t* o = (uint16_t*) output;
  $if BATCH_TILE > 16:
    for (; batch >= ${BATCH_TILE} * sizeof(float); batch -= ${BATCH_TILE} * sizeof(float)) {
      const __m512 vf0 = _mm512_loadu_ps(input);
      $for N in range(1, SIMD_TILE):
        const __m512 vf${N} = _mm512_loadu_ps(input + ${N * 16});
      input += ${BATCH_TILE};

      _mm256_storeu_si256((__m256i*) o, _mm512_cvtps_ph(vf0, _MM_FROUND_NO_EXC | _MM_FROUND_TO_NEAREST_INT));
      $for N in range(1, SIMD_TILE):
        _mm256_storeu_si256((__m256i*) (o + ${N * 16}), _mm512_cvtps_ph(vf${N}, _MM_FROUND_NO_EXC | _MM_FROUND_TO_NEAREST_INT));
      o += ${BATCH_TILE};
    }
  for (; batch >= 16 * sizeof(float); batch -= 16 * sizeof(float)) {
    const __m512 vf = _mm512_loadu_ps(input);
    input += 16;

    _mm256_storeu_si256((__m256i*) o, _mm512_cvtps_ph(vf, _MM_FROUND_NO_EXC | _MM_FROUND_TO_NEAREST_INT));
    o += 16;
  }
  if XNN_UNLIKELY(batch != 0) {
    assert(batch >= 1 * sizeof(float));
    assert(batch <= 15 * sizeof(float));

    // Prepare mask for valid elements (depends on batch).
    batch >>= XNN_LOG2_SIZEOF_FLOAT;
    const __mmask16 vmask = _cvtu32_mask16((uint16_t) ((uint32_t) (UINT32_C(1) << batch) - UINT32_C(1)));

    const __m512 vf = _mm512_maskz_loadu_ps(vmask, input);
    const __m256i vh = _mm512_cvtps_ph(vf, _MM_FROUND_NO_EXC | _MM_FROUND_TO_NEAREST_INT);
    _mm256_mask_storeu_epi16(o, vmask, vh);
  }
}