Compute Library
 23.05
fp32.cpp
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2020-2021 Arm Limited.
3  *
4  * SPDX-License-Identifier: MIT
5  *
6  * Permission is hereby granted, free of charge, to any person obtaining a copy
7  * of this software and associated documentation files (the "Software"), to
8  * deal in the Software without restriction, including without limitation the
9  * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
10  * sell copies of the Software, and to permit persons to whom the Software is
11  * furnished to do so, subject to the following conditions:
12  *
13  * The above copyright notice and this permission notice shall be included in all
14  * copies or substantial portions of the Software.
15  *
16  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
19  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
21  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
22  * SOFTWARE.
23  */
25 #include "src/core/NEON/NEMath.h"
26 
27 #include <arm_neon.h>
28 #include <cmath>
29 #include <cstddef>
30 
31 namespace arm_compute
32 {
33 namespace cpu
34 {
35 constexpr int step = 4;
36 
37 void fp32_neon_floor(const void *src, void *dst, int len)
38 {
41  ARM_COMPUTE_ASSERT(len >= 0);
42 
43  auto psrc = static_cast<const float *>(src);
44  auto pdst = static_cast<float *>(dst);
45 
46  for(; len >= step; len -= step)
47  {
48  vst1q_f32(pdst, vfloorq_f32(vld1q_f32(psrc)));
49  psrc += step;
50  pdst += step;
51  }
52 
53  for(; len > 0; --len)
54  {
55  *pdst = std::floor(*psrc);
56  ++pdst;
57  ++psrc;
58  }
59 }
60 } // namespace cpu
61 } // namespace arm_compute
#define ARM_COMPUTE_ASSERT(cond)
Definition: Validate.h:37
SimpleTensor< float > src
Definition: DFT.cpp:155
Copyright (c) 2017-2023 Arm Limited.
float32x4_t vfloorq_f32(float32x4_t val)
Calculate floor of a vector.
constexpr int step
Definition: fp32.cpp:35
void fp32_neon_floor(const void *src, void *dst, int len)
Definition: fp32.cpp:37
#define ARM_COMPUTE_ASSERT_NOT_NULLPTR(ptr)
Definition: Validate.h:38