amilanov-amd
403 lines
16 KiB
C++
403 lines
16 KiB
C++
/*
|
|
Copyright (c) 2021 Advanced Micro Devices, Inc. All rights reserved.
|
|
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.
|
|
*/
|
|
|
|
/*
|
|
This file verifies the following scenarios of hipMemcpy2DFromArrayAsync API
|
|
1. Negative Scenarios
|
|
2. Extent Validation Scenarios
|
|
3. hipMemcpy2DFromArrayAsync Basic Scenario
|
|
4. Pinned Memory scenarios on same and peer GPU
|
|
5. Device Context change scenario where memory is allocated in
|
|
one GPU and stream is created in peer GPU.
|
|
*/
|
|
|
|
#include <hip_test_common.hh>
|
|
#include <hip_test_checkers.hh>
|
|
|
|
|
|
static constexpr auto NUM_W{10};
|
|
static constexpr auto NUM_H{10};
|
|
|
|
/*
|
|
* This testcase copies the data from host to device of
|
|
hipMemcpy2DFromArrayAsync API
|
|
* INPUT: Copying Host variable hData(Initialized with value Phi(1.618))
|
|
* --> A_d device variable
|
|
* OUTPUT: For validating the result,Copying A_d device variable
|
|
* --> A_h host variable
|
|
* and verifying A_h with Phi
|
|
*/
|
|
TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_Basic") {
|
|
CHECK_IMAGE_SUPPORT
|
|
|
|
HIP_CHECK(hipSetDevice(0));
|
|
hipArray_t A_d{nullptr};
|
|
size_t width{sizeof(float) * NUM_W};
|
|
float *A_h{nullptr}, *hData{nullptr};
|
|
hipStream_t stream;
|
|
|
|
// Initialization of variables
|
|
HipTest::initArrays<float>(nullptr, nullptr, nullptr, &A_h, &hData, nullptr, width * NUM_H,
|
|
false);
|
|
hipChannelFormatDesc desc = hipCreateChannelDesc<float>();
|
|
HIP_CHECK(hipMallocArray(&A_d, &desc, NUM_W, NUM_H, hipArrayDefault));
|
|
HipTest::setDefaultData<float>(width * NUM_H, A_h, hData, nullptr);
|
|
HIP_CHECK(hipStreamCreate(&stream));
|
|
|
|
HIP_CHECK(hipMemcpy2DToArray(A_d, 0, 0, hData, width, width, NUM_H, hipMemcpyHostToDevice));
|
|
SECTION("Calling hipMemcpy2DFromArrayAsync() with user declared stream obj") {
|
|
HIP_CHECK(hipMemcpy2DFromArrayAsync(A_h, width, A_d, 0, 0, width, NUM_H, hipMemcpyDeviceToHost,
|
|
stream));
|
|
HIP_CHECK(hipStreamSynchronize(stream));
|
|
}
|
|
SECTION("Calling hipMemcpy2DFromArrayAsync() with hipStreamPerThread") {
|
|
HIP_CHECK(hipMemcpy2DFromArrayAsync(A_h, width, A_d, 0, 0, width, NUM_H, hipMemcpyDeviceToHost,
|
|
hipStreamPerThread));
|
|
HIP_CHECK(hipStreamSynchronize(hipStreamPerThread));
|
|
}
|
|
REQUIRE(HipTest::checkArray(A_h, hData, NUM_W, NUM_H) == true);
|
|
|
|
// Cleaning the memory
|
|
HIP_CHECK(hipFreeArray(A_d));
|
|
HIP_CHECK(hipStreamDestroy(stream));
|
|
HipTest::freeArrays<float>(nullptr, nullptr, nullptr, A_h, hData, nullptr, false);
|
|
}
|
|
|
|
/*
|
|
* This testcase verifies the extent validation scenarios
|
|
* of hipMemcpy2DFromArrayAsync API
|
|
*/
|
|
TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_ExtentValidation") {
|
|
CHECK_IMAGE_SUPPORT
|
|
|
|
HIP_CHECK(hipSetDevice(0));
|
|
hipArray_t A_d{nullptr};
|
|
size_t width{sizeof(float) * NUM_W};
|
|
float *A_h{nullptr}, *hData{nullptr}, *valData{nullptr};
|
|
hipStream_t stream;
|
|
|
|
// Initialization of variables
|
|
HipTest::initArrays<float>(nullptr, nullptr, nullptr, &A_h, &hData, nullptr, width * NUM_H,
|
|
false);
|
|
HipTest::initArrays<float>(nullptr, nullptr, nullptr, nullptr, &valData, nullptr, width * NUM_H,
|
|
false);
|
|
hipChannelFormatDesc desc = hipCreateChannelDesc<float>();
|
|
HIP_CHECK(hipMallocArray(&A_d, &desc, NUM_W, NUM_H, hipArrayDefault));
|
|
HIP_CHECK(hipStreamCreate(&stream));
|
|
|
|
SECTION("Destination width is 0") {
|
|
REQUIRE(hipMemcpy2DFromArrayAsync(A_h, 0, A_d, 0, 0, NUM_W * sizeof(float), NUM_H,
|
|
hipMemcpyDeviceToHost, stream) != hipSuccess);
|
|
}
|
|
// hipMemcpy2DFromArrayAsync API would return success for
|
|
// width and height as 0
|
|
// and does not perform any copy
|
|
// Validating the result with the initialized value
|
|
// 1.Initializing A_d with Pi value
|
|
// 2.copying A_d-->hData variable
|
|
// with height 0(copy will not be performed)
|
|
// 3 validating hData<-->A_h which will not be equal as copy is not done.
|
|
SECTION("Height is 0") {
|
|
HIP_CHECK(hipMemcpy2DToArray(A_d, 0, 0, A_h, width, width, NUM_H, hipMemcpyHostToDevice));
|
|
HIP_CHECK(hipMemcpy2DFromArrayAsync(hData, width, A_d, 0, 0, NUM_W * sizeof(float), 0,
|
|
hipMemcpyDeviceToHost, stream));
|
|
HIP_CHECK(hipStreamSynchronize(stream));
|
|
REQUIRE(HipTest::checkArray(hData, valData, NUM_W, NUM_H) == true);
|
|
}
|
|
// hipMemcpy2DFromArrayAsync API would return success for
|
|
// width and height as 0
|
|
// and does not perform any copy
|
|
// Validating the result with the initialized value
|
|
// 1.Initializing A_d with Pi value
|
|
// 2.copying A_d-->hData variable
|
|
// with width 0(copy will not be performed)
|
|
// 3 validating hData<-->A_h which will not be equal as copy is not done.
|
|
|
|
SECTION("Width is 0") {
|
|
HIP_CHECK(hipMemcpy2DToArray(A_d, 0, 0, A_h, width, width, NUM_H, hipMemcpyHostToDevice));
|
|
HIP_CHECK(hipMemcpy2DFromArrayAsync(hData, width, A_d, 0, 0, 0, NUM_H, hipMemcpyDeviceToHost,
|
|
stream));
|
|
HIP_CHECK(hipStreamSynchronize(stream));
|
|
REQUIRE(HipTest::checkArray(hData, valData, NUM_W, NUM_H) == true);
|
|
}
|
|
|
|
// Cleaning the memory
|
|
HIP_CHECK(hipFreeArray(A_d));
|
|
HIP_CHECK(hipStreamDestroy(stream));
|
|
HipTest::freeArrays<float>(nullptr, nullptr, nullptr, A_h, hData, nullptr, false);
|
|
HipTest::freeArrays<float>(nullptr, nullptr, nullptr, nullptr, valData, nullptr, false);
|
|
}
|
|
/*
|
|
* This Scenario Verifies hipMemcpy2DFromArrayAsync API by copying the
|
|
* data from pinned host memory to device on same GPU
|
|
* INPUT: Copying Host variable PinnMem(Initialized with value "10" )
|
|
* --> A_d device variable
|
|
* OUTPUT: For validating the result,Copying A_d device variable
|
|
* --> A_h host variable
|
|
* and verifying A_h with PinnedMem[0](i.e., 10)
|
|
*/
|
|
TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_PinnedHostMemSameGpu") {
|
|
CHECK_IMAGE_SUPPORT
|
|
|
|
HIP_CHECK(hipSetDevice(0));
|
|
hipArray_t A_d{nullptr};
|
|
constexpr auto def_val{10};
|
|
size_t width{sizeof(float) * NUM_W};
|
|
float *A_h{nullptr}, *PinnMem{nullptr};
|
|
hipStream_t stream;
|
|
|
|
// Initialization of variables
|
|
HipTest::initArrays<float>(nullptr, nullptr, nullptr, &A_h, nullptr, nullptr, width * NUM_H,
|
|
false);
|
|
HIP_CHECK(hipHostMalloc(reinterpret_cast<void**>(&PinnMem), width * NUM_H));
|
|
hipChannelFormatDesc desc = hipCreateChannelDesc<float>();
|
|
HIP_CHECK(hipMallocArray(&A_d, &desc, NUM_W, NUM_H, hipArrayDefault));
|
|
HipTest::setDefaultData<float>(width * NUM_H, A_h, nullptr, nullptr);
|
|
for (int i = 0; i < NUM_W * NUM_H; i++) {
|
|
PinnMem[i] = def_val + i;
|
|
}
|
|
HIP_CHECK(hipStreamCreate(&stream));
|
|
|
|
HIP_CHECK(hipMemcpy2DToArray(A_d, 0, 0, PinnMem, width, width, NUM_H, hipMemcpyHostToDevice));
|
|
HIP_CHECK(hipMemcpy2DFromArrayAsync(A_h, width, A_d, 0, 0, width, NUM_H, hipMemcpyDeviceToHost,
|
|
stream));
|
|
HIP_CHECK(hipStreamSynchronize(stream));
|
|
REQUIRE(HipTest::checkArray(A_h, PinnMem, NUM_W, NUM_H) == true);
|
|
|
|
// Cleaning the memory
|
|
HIP_CHECK(hipFreeArray(A_d));
|
|
HIP_CHECK(hipHostFree(PinnMem));
|
|
HIP_CHECK(hipStreamDestroy(stream));
|
|
HipTest::freeArrays<float>(nullptr, nullptr, nullptr, A_h, nullptr, nullptr, false);
|
|
}
|
|
/*
|
|
* This Scenario Verifies hipMemcpy2DFromArrayAsync API by copying the
|
|
* data from pinned host memory to device from Peer GPU.
|
|
* Device Memory is allocated in GPU 0 and the API is trigerred from GPU1
|
|
* INPUT: Initialize data, A_h --> A_d device variable
|
|
* whose memory is allocated in GPU 0
|
|
then A_d-->E_h in GPU1
|
|
* OUTPUT: validating the result by comparing A_h and E_h
|
|
*/
|
|
TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_multiDevicePinnedHostMem",
|
|
"[multigpu]") {
|
|
CHECK_IMAGE_SUPPORT
|
|
|
|
int numDevices = 0;
|
|
constexpr auto def_val{10};
|
|
HIP_CHECK(hipGetDeviceCount(&numDevices));
|
|
if (numDevices > 1) {
|
|
int canAccessPeer = 0;
|
|
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
|
|
if (canAccessPeer) {
|
|
HIP_CHECK(hipSetDevice(0));
|
|
hipArray_t A_d{nullptr};
|
|
size_t width{sizeof(float) * NUM_W};
|
|
float *A_h{nullptr}, *E_h{nullptr};
|
|
hipStream_t stream;
|
|
HIP_CHECK(hipStreamCreate(&stream));
|
|
|
|
// Initialization of variables
|
|
HipTest::initArrays<float>(nullptr, nullptr, nullptr, &A_h, nullptr, nullptr, width * NUM_H,
|
|
false);
|
|
hipChannelFormatDesc desc = hipCreateChannelDesc<float>();
|
|
HIP_CHECK(hipMallocArray(&A_d, &desc, NUM_W, NUM_H, hipArrayDefault));
|
|
HipTest::setDefaultData<float>(width * NUM_H, A_h, nullptr, nullptr);
|
|
HIP_CHECK(hipHostMalloc(reinterpret_cast<void**>(&E_h), width * NUM_H));
|
|
for (int i = 0; i < NUM_W * NUM_H; i++) {
|
|
E_h[i] = def_val + i;
|
|
}
|
|
|
|
HIP_CHECK(hipMemcpy2DToArray(A_d, 0, 0, A_h, width, width, NUM_H, hipMemcpyHostToDevice));
|
|
HIP_CHECK(hipSetDevice(1));
|
|
HIP_CHECK(hipMemcpy2DFromArrayAsync(E_h, width, A_d, 0, 0, width, NUM_H,
|
|
hipMemcpyDeviceToHost, stream));
|
|
HIP_CHECK(hipStreamSynchronize(stream));
|
|
REQUIRE(HipTest::checkArray(A_h, E_h, NUM_W, NUM_H) == true);
|
|
|
|
// Cleaning the memory
|
|
HIP_CHECK(hipFreeArray(A_d));
|
|
HIP_CHECK(hipHostFree(E_h));
|
|
HIP_CHECK(hipStreamDestroy(stream));
|
|
HipTest::freeArrays<float>(nullptr, nullptr, nullptr, A_h, nullptr, nullptr, false);
|
|
} else {
|
|
SUCCEED("Device Does not have P2P capability");
|
|
}
|
|
} else {
|
|
SUCCEED("Number of devices are < 2");
|
|
}
|
|
}
|
|
|
|
/*
|
|
* This scenario verifies the hipMemcpy2DFromArrayAsync API in case of device
|
|
* context change.
|
|
* Memory is allocated in GPU-0 and the API is triggered from GPU-1
|
|
* INPUT: Copying Host variable hData(Initial value Phi)
|
|
* --> A_d device variable
|
|
* whose memory is allocated in GPU 0
|
|
* OUTPUT: For validating the result,Copying A_d device variable
|
|
* --> A_h host variable
|
|
* and verifying A_h with Phi
|
|
* */
|
|
TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_multiDeviceContextChange",
|
|
"[multigpu]") {
|
|
CHECK_IMAGE_SUPPORT
|
|
|
|
int numDevices = 0;
|
|
HIP_CHECK(hipGetDeviceCount(&numDevices));
|
|
if (numDevices > 1) {
|
|
int canAccessPeer = 0;
|
|
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
|
|
if (canAccessPeer) {
|
|
HIP_CHECK(hipSetDevice(0));
|
|
hipArray_t A_d{nullptr};
|
|
size_t width{sizeof(float) * NUM_W};
|
|
float *A_h{nullptr}, *hData{nullptr};
|
|
hipStream_t stream;
|
|
|
|
// Initialization of variables
|
|
HipTest::initArrays<float>(nullptr, nullptr, nullptr, &A_h, &hData, nullptr, width * NUM_H,
|
|
false);
|
|
hipChannelFormatDesc desc = hipCreateChannelDesc<float>();
|
|
HIP_CHECK(hipMallocArray(&A_d, &desc, NUM_W, NUM_H, hipArrayDefault));
|
|
HipTest::setDefaultData<float>(width * NUM_H, A_h, hData, nullptr);
|
|
|
|
HIP_CHECK(hipSetDevice(1));
|
|
HIP_CHECK(hipStreamCreate(&stream));
|
|
HIP_CHECK(hipMemcpy2DToArray(A_d, 0, 0, hData, width, width, NUM_H, hipMemcpyHostToDevice));
|
|
|
|
HIP_CHECK(hipMemcpy2DFromArrayAsync(A_h, width, A_d, 0, 0, width, NUM_H,
|
|
hipMemcpyDeviceToHost, stream));
|
|
HIP_CHECK(hipStreamSynchronize(stream));
|
|
REQUIRE(HipTest::checkArray(A_h, hData, NUM_W, NUM_H) == true);
|
|
|
|
// Cleaning the memory
|
|
HIP_CHECK(hipFreeArray(A_d));
|
|
HIP_CHECK(hipStreamDestroy(stream));
|
|
HipTest::freeArrays<float>(nullptr, nullptr, nullptr, A_h, hData, nullptr, false);
|
|
} else {
|
|
SUCCEED("Device Does not have P2P capability");
|
|
}
|
|
} else {
|
|
SUCCEED("Number of devices are < 2");
|
|
}
|
|
}
|
|
/* This testcase verifies the negative scenarios
|
|
* of hipMemcpy2DFromArrayAsync API
|
|
*/
|
|
TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_Negative") {
|
|
CHECK_IMAGE_SUPPORT
|
|
|
|
HIP_CHECK(hipSetDevice(0));
|
|
hipArray_t A_d{nullptr};
|
|
size_t width{sizeof(float) * NUM_W};
|
|
float *A_h{nullptr}, *hData{nullptr};
|
|
hipStream_t stream;
|
|
HIP_CHECK(hipStreamCreate(&stream));
|
|
|
|
// Initialization of variables
|
|
HipTest::initArrays<float>(nullptr, nullptr, nullptr, &A_h, &hData, nullptr, width * NUM_H,
|
|
false);
|
|
HipTest::setDefaultData<float>(width * NUM_H, A_h, hData, nullptr);
|
|
hipChannelFormatDesc desc = hipCreateChannelDesc<float>();
|
|
HIP_CHECK(hipMallocArray(&A_d, &desc, NUM_W, NUM_H, hipArrayDefault));
|
|
|
|
SECTION("Nullptr to destination") {
|
|
REQUIRE(hipMemcpy2DFromArrayAsync(nullptr, width, A_d, 0, 0, width, NUM_H,
|
|
hipMemcpyDeviceToHost, stream) != hipSuccess);
|
|
}
|
|
|
|
SECTION("Nullptr to source") {
|
|
REQUIRE(hipMemcpy2DFromArrayAsync(A_h, width, nullptr, 0, 0, width, NUM_H,
|
|
hipMemcpyDeviceToHost, stream) != hipSuccess);
|
|
}
|
|
|
|
SECTION("Passing offset more than 0") {
|
|
REQUIRE(hipMemcpy2DFromArrayAsync(A_h, width, A_d, 1, 1, width, NUM_H, hipMemcpyDeviceToHost,
|
|
stream) != hipSuccess);
|
|
}
|
|
|
|
SECTION("Passing array more than allocated") {
|
|
REQUIRE(hipMemcpy2DFromArrayAsync(A_h, width, A_d, 0, 0, width + 2, NUM_H + 2,
|
|
hipMemcpyDeviceToHost, stream) != hipSuccess);
|
|
}
|
|
|
|
// Cleaning of Memory
|
|
HIP_CHECK(hipFreeArray(A_d));
|
|
HIP_CHECK(hipStreamDestroy(stream));
|
|
HipTest::freeArrays<float>(nullptr, nullptr, nullptr, A_h, hData, nullptr, false);
|
|
}
|
|
|
|
/**
|
|
* Test Description
|
|
* ------------------------
|
|
* - This testcase copies the data from host to device and launches
|
|
* hipMemcpy2DFromArrayAsync within the graph to trigger
|
|
* capturehipMemcpy2DFromArrayAsync internal api and verifies data in host.
|
|
* Test source
|
|
* ------------------------
|
|
* - unit/memory/hipMemcpy2DFromArrayAsync_old.cc
|
|
* Test requirements
|
|
* ------------------------
|
|
* - HIP_VERSION >= 6.0
|
|
*/
|
|
TEST_CASE("Unit_hipMemcpy2DFromArrayAsync_Capture") {
|
|
CHECK_IMAGE_SUPPORT
|
|
|
|
constexpr int kTestSizes[] = {3, 4, 100};
|
|
int num_rows = GENERATE_REF(from_range(std::begin(kTestSizes), std::end(kTestSizes)));
|
|
int num_cols = GENERATE_REF(from_range(std::begin(kTestSizes), std::end(kTestSizes)));
|
|
|
|
auto host_src = std::make_unique<int[]>(num_rows * num_cols);
|
|
auto host_dst = std::make_unique<int[]>(num_rows * num_cols);
|
|
|
|
for (int row = 0; row < num_rows; ++row) {
|
|
for (int col = 0; col < num_cols; ++col) {
|
|
host_src[row * num_cols + col] = row * num_cols + col;
|
|
}
|
|
}
|
|
|
|
hipArray_t device_array = nullptr;
|
|
hipChannelFormatDesc channel_desc = hipCreateChannelDesc<int>();
|
|
HIP_CHECK(hipMallocArray(&device_array, &channel_desc, num_cols, num_rows, hipArrayDefault));
|
|
HIP_CHECK(hipMemcpy2DToArray(device_array, 0, 0, host_src.get(), num_cols * sizeof(int),
|
|
num_cols * sizeof(int), num_rows, hipMemcpyHostToDevice));
|
|
|
|
hipStream_t stream = nullptr;
|
|
HIP_CHECK(hipStreamCreate(&stream));
|
|
|
|
GENERATE_CAPTURE();
|
|
BEGIN_CAPTURE(stream);
|
|
HIP_CHECK(hipMemcpy2DFromArrayAsync(host_dst.get(), sizeof(int) * num_cols, device_array, 0, 0,
|
|
sizeof(int) * num_cols, num_rows, hipMemcpyDeviceToHost,
|
|
stream));
|
|
END_CAPTURE(stream);
|
|
|
|
HIP_CHECK(hipStreamSynchronize(stream));
|
|
|
|
for (int row = 0; row < num_rows; ++row) {
|
|
for (int col = 0; col < num_cols; ++col) {
|
|
REQUIRE(host_dst[row * num_cols + col] == (row * num_cols + col));
|
|
}
|
|
}
|
|
|
|
HIP_CHECK(hipStreamDestroy(stream));
|
|
HIP_CHECK(hipFreeArray(device_array));
|
|
}
|