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
|
/*******************************************************************************
*
* MIT License
*
* Copyright (c) 2023 Advanced Micro Devices, Inc.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in all
* copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*
*******************************************************************************/
#include <gtest/gtest.h>
#include <gtest/gtest_common.hpp>
#include <miopen/miopen.h>
#include "tensor_holder.hpp"
#include "get_handle.hpp"
#include "cba.hpp"
#if MIOPEN_BACKEND_HIP
namespace {
bool IsTestSupportedForDevice()
{
using e_mask = enabled<Gpu::gfx94X, Gpu::gfx103X, Gpu::gfx110X>;
// gfx120X is not enabled due to WORKAROUND_SWDEV_479810
using d_mask = disabled<Gpu::None>;
return ::IsTestSupportedForDevMask<d_mask, e_mask>();
}
template <typename T>
class FusionSetArgTest : public ConvBiasActivInferTest<T>
{
public:
void SetUp() override
{
cba<T>::SetUp();
weights2 = tensor<T>{cba<T>::tensor_layout, cba<T>::conv_config.GetWeights()};
weights2.generate(tensor_elem_gen_integer{3});
cba<T>::weights = weights2;
auto&& handle = get_handle();
cba<T>::wei_dev = handle.Write(weights2.data);
handle.Finish();
}
void TearDown() override { cba<T>::TearDown(); }
template <typename Tp>
using cba = ConvBiasActivInferTest<Tp>;
tensor<T> weights2;
miopen::Allocator::ManageDataPtr wei_dev2;
};
bool SkipTest() { return get_handle_xnack(); }
} // namespace
using GPU_FusionSetArg_FP16 = FusionSetArgTest<float>;
TEST_P(GPU_FusionSetArg_FP16, TestSetArgApiCall)
{
// Original fusion_plan/args execution happens in cba_infer.cpp
// Original is checked independently and not sequentially, prior to FusionTestSetArgTest.
if(SkipTest())
{
test_skipped = true;
GTEST_SKIP() << "Fusion does not support xnack";
}
if(!IsTestSupportedForDevice())
{
test_skipped = true;
GTEST_SKIP() << "CBA fusion_test is not supported for this device";
}
using cba_float = cba<float>;
auto&& handle = get_handle();
auto convOp = std::make_shared<miopen::ConvForwardOpDescriptor>(cba_float::conv_desc,
cba_float::weights.desc);
miopenOperatorArgs_t fusion_args = static_cast<miopenOperatorArgs_t>(&(cba_float::params));
miopenFusionPlanDescriptor_t fusion_plan =
static_cast<miopenFusionPlanDescriptor_t>(&(cba_float::fusePlanDesc));
miopenFusionOpDescriptor_t conv_op = static_cast<miopenFusionOpDescriptor_t>(convOp.get());
EXPECT_EQ(miopenCompileFusionPlan(&handle, fusion_plan), 0);
EXPECT_EQ(miopenSetOpArgsConvForward(fusion_args,
conv_op,
&(cba_float::alpha),
&(cba_float::beta),
cba_float::wei_dev.get()),
0);
EXPECT_EQ(miopenExecuteFusionPlan(&handle,
fusion_plan,
&(cba_float::input.desc),
cba_float::in_dev.get(),
&(cba_float::output.desc),
cba_float::out_dev.get(),
fusion_args),
0);
handle.Finish();
using ConvParam = miopen::fusion::ConvolutionOpInvokeParam;
ConvParam* conv_param = dynamic_cast<ConvParam*>(miopen::deref(fusion_args).params[0].get());
ASSERT_EQ(conv_param->weights, wei_dev.get());
}
INSTANTIATE_TEST_SUITE_P(Full,
GPU_FusionSetArg_FP16,
testing::Combine(testing::Values(miopenActivationRELU),
testing::ValuesIn(GetNetwork1<ConvTestCaseBase>()),
testing::Values(miopenTensorNCHW)));
#endif
|
