Files
oneDNN/tests/gtests/test_global_scratchpad.cpp

107 lines
3.7 KiB
C++

/*******************************************************************************
* Copyright 2020-2022 Intel Corporation
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*******************************************************************************/
#include "dnnl_test_common.hpp"
#include "gtest/gtest.h"
#include "oneapi/dnnl/dnnl.hpp"
namespace dnnl {
using dt = memory::data_type;
using tag = memory::format_tag;
// This test checks that globally defined primitive_t object will be
// successfully destroyed after finishing the program despite the order of
// internal objects destruction.
// The cause was thread-local non-trivially-constructed object in
// global_scratchpad_t object which got destroyed before global_scratchpad_t
// causing a crash.
class global_scratchpad_t : public ::testing::Test {};
struct conv_ctx_t {
conv_ctx_t() : eng_(engine::kind::cpu, 0) {}
struct conv_t {
conv_t() = default;
memory::desc src_md;
memory::desc wei_md;
memory::desc dst_md;
convolution_forward::primitive_desc pd;
memory src_mem;
memory wei_mem;
memory dst_mem;
primitive prim;
};
void Setup(const memory::dims &src_dims, const memory::dims &wei_dims,
const memory::dims &dst_dims, const memory::dims &strides_dims,
const memory::dims &dilations_dims,
const memory::dims &padding_left,
const memory::dims &padding_right) {
c_.src_md = memory::desc(src_dims, dt::f32, tag::any);
c_.wei_md = memory::desc(wei_dims, dt::f32, tag::any);
c_.dst_md = memory::desc(dst_dims, dt::f32, tag::any);
c_.pd = convolution_forward::primitive_desc(eng_, prop_kind::forward,
algorithm::convolution_direct, c_.src_md, c_.wei_md, c_.dst_md,
strides_dims, dilations_dims, padding_left, padding_right);
c_.src_mem = test::make_memory(c_.pd.src_desc(), eng_);
c_.wei_mem = test::make_memory(c_.pd.weights_desc(), eng_);
c_.dst_mem = test::make_memory(c_.pd.dst_desc(), eng_);
c_.prim = convolution_forward(c_.pd);
}
engine eng_;
struct conv_t c_;
};
conv_ctx_t global_conv_ctx1;
conv_ctx_t global_conv_ctx2;
HANDLE_EXCEPTIONS_FOR_TEST(global_scratchpad_t, TestGlobalScratchpad) {
#if defined(DNNL_WITH_SYCL) && defined(TEST_DNNL_DPCPP_BUFFER)
// It seems static USM data doesn't get along with OpenCL runtime.
// TODO: investigate.
if (get_test_engine_kind() == engine::kind::gpu) return;
#endif
memory::dims src1 = {1, 1, 3, 4};
memory::dims wei1 = {1, 1, 3, 3};
memory::dims dst1 = {1, 1, 8, 5};
memory::dims str1 = {1, 1};
memory::dims dil1 = {0, 0};
memory::dims pad_l1 = {3, 1};
memory::dims pad_r1 = {4, 2};
global_conv_ctx1.Setup(src1, wei1, dst1, str1, dil1, pad_l1, pad_r1);
memory::dims src2 = {256, 3, 227, 227};
memory::dims wei2 = {96, 3, 11, 11};
memory::dims dst2 = {256, 96, 55, 55};
memory::dims str2 = {4, 4};
memory::dims dil2 = {0, 0};
memory::dims pad_l2 = {0, 0};
memory::dims pad_r2 = {0, 0};
global_conv_ctx2.Setup(src2, wei2, dst2, str2, dil2, pad_l2, pad_r2);
// if something goes wrong, test should return 139 on Linux.
};
} // namespace dnnl