xref: /aosp_15_r20/external/ComputeLibrary/src/cpu/operators/CpuMul.cpp (revision c217d954acce2dbc11938adb493fc0abd69584f3)
1 /*
2  * Copyright (c) 2016-2022 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  */
24 #include "src/cpu/operators/CpuMul.h"
25 
26 #include "arm_compute/core/TensorInfo.h"
27 #include "arm_compute/core/Validate.h"
28 #include "arm_compute/runtime/NEON/NEScheduler.h"
29 #include "src/common/utils/Log.h"
30 #include "src/cpu/kernels/CpuMulKernel.h"
31 
32 namespace arm_compute
33 {
34 namespace cpu
35 {
validate(const ITensorInfo * src1,const ITensorInfo * src2,const ITensorInfo * dst,float scale,ConvertPolicy overflow_policy,RoundingPolicy rounding_policy,const ActivationLayerInfo & act_info)36 Status CpuMul::validate(const ITensorInfo *src1, const ITensorInfo *src2, const ITensorInfo *dst, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy,
37                         const ActivationLayerInfo &act_info)
38 {
39     ARM_COMPUTE_RETURN_ERROR_ON(act_info.enabled());
40     return kernels::CpuMulKernel::validate(src1, src2, dst, scale, overflow_policy, rounding_policy);
41 }
42 
configure(ITensorInfo * src1,ITensorInfo * src2,ITensorInfo * dst,float scale,ConvertPolicy overflow_policy,RoundingPolicy rounding_policy,const ActivationLayerInfo & act_info)43 void CpuMul::configure(ITensorInfo *src1, ITensorInfo *src2, ITensorInfo *dst, float scale, ConvertPolicy overflow_policy, RoundingPolicy rounding_policy,
44                        const ActivationLayerInfo &act_info)
45 {
46     ARM_COMPUTE_UNUSED(act_info);
47     ARM_COMPUTE_LOG_PARAMS(src1, src2, dst, scale, overflow_policy, rounding_policy, act_info);
48 
49     auto k = std::make_unique<kernels::CpuMulKernel>();
50     k->configure(src1, src2, dst, scale, overflow_policy, rounding_policy);
51     _kernel = std::move(k);
52 }
53 
run(ITensorPack & tensors)54 void CpuMul::run(ITensorPack &tensors)
55 {
56     ARM_COMPUTE_ERROR_ON_MSG(tensors.empty(), "No inputs provided");
57     auto split_dimension = static_cast<kernels::CpuMulKernel *>(_kernel.get())->get_split_dimension_hint();
58     NEScheduler::get().schedule_op(_kernel.get(), split_dimension, _kernel->window(), tensors);
59 }
60 
validate(const ITensorInfo * src1,const ITensorInfo * src2,const ITensorInfo * dst,const ActivationLayerInfo & act_info)61 Status CpuComplexMul::validate(const ITensorInfo *src1, const ITensorInfo *src2, const ITensorInfo *dst, const ActivationLayerInfo &act_info)
62 {
63     ARM_COMPUTE_RETURN_ERROR_ON(act_info.enabled());
64     return kernels::CpuComplexMulKernel::validate(src1, src2, dst);
65 }
66 
configure(ITensorInfo * src1,ITensorInfo * src2,ITensorInfo * dst,const ActivationLayerInfo & act_info)67 void CpuComplexMul::configure(ITensorInfo *src1, ITensorInfo *src2, ITensorInfo *dst, const ActivationLayerInfo &act_info)
68 {
69     ARM_COMPUTE_UNUSED(act_info);
70     ARM_COMPUTE_LOG_PARAMS(src1, src2, dst, act_info);
71 
72     auto k = std::make_unique<kernels::CpuComplexMulKernel>();
73     k->configure(src1, src2, dst);
74     _kernel = std::move(k);
75 }
76 
run(ITensorPack & tensors)77 void CpuComplexMul::run(ITensorPack &tensors)
78 {
79     ARM_COMPUTE_ERROR_ON_MSG(tensors.empty(), "No inputs provided");
80     NEScheduler::get().schedule_op(_kernel.get(), Window::DimY, _kernel->window(), tensors);
81 }
82 } // namespace cpu
83 } // namespace arm_compute