AI/rocm-systems
2
0
انشعاب 0
کد مسائل تقاضاهای واکشی Actions Packages پروژه‌ها انتشارها دانشنامه فعالیت

EXSWHTEC-97 - Implement tests for hipMemcpy2D APIs #53

فهرست منبع 
Change-Id: I674741e641b1ebd6adb5c60c05c2d8ade27510c6
This commit is contained in:
Mirza Halilčević
2023-12-28 19:33:15 +01:00
کامیت شده توسط Rakesh Roy
والد 72aa591328
کامیت c9e664b8e6
6فایلهای تغییر یافته به همراه1652 افزوده شده و 986 حذف شده
دانلود پرونده وصله دانلود فایل تغییرات diff Expand all files Collapse all files
catch/unit/memory/CMakeLists.txt
+2
مشاهده پرونده
@@ -40,7 +40,9 @@ set(TEST_SRC
hipMemcpyParam2D.cc
hipMemcpyParam2DAsync.cc
hipMemcpy2D.cc
hipMemcpy2D_old.cc
hipMemcpy2DAsync.cc
hipMemcpy2DAsync_old.cc
hipMemcpy2DFromArray.cc
hipMemcpy2DFromArray_old.cc
hipMemcpy2DFromArrayAsync.cc
catch/unit/memory/hipMemcpy2D.cc
+112 -457
مشاهده پرونده
@@ -1,496 +1,151 @@
/*
Copyright (c) 2021-2023 Advanced Micro Devices, Inc. All rights reserved.
Copyright (c) 2022 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 WARRANNTY OF ANY KIND, EXPRESS OR
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
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 INN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR INN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
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.
*/
/**
* @addtogroup hipMemcpy2D hipMemcpy2D
* @{
* @ingroup MemcpyTest
* `hipMemcpy2D(void* dst, size_t dpitch, const void* src,
* size_t spitch, size_t width, size_t height,
* hipMemcpyKind kind)` -
* Copies data between host and device.
*/
// Testcase Description:
// 1) Verifies the working of Memcpy2D API negative scenarios by
// Pass NULL to destination pointer
// Pass NULL to Source pointer
// Pass width greater than spitch/dpitch
// 2) Verifies hipMemcpy2D API by
// pass 0 to destionation pitch
// pass 0 to source pitch
// pass 0 to width
// pass 0 to height
// 3) Verifies working of Memcpy2D API on host memory and pinned host memory by
// performing D2H, D2D and H2D memory kind copies on same GPU
// 4) Verifies working of Memcpy2D API for the following scenarios
// H2D-D2D-D2H on host and device memory
// H2D-D2D-D2H on pinned host and device memory
// H2D-D2D-D2H functionalities where memory is allocated in GPU-0
// and API is triggered from GPU-1
#include "memcpy2d_tests_common.hh"
#include <hip_test_common.hh>
#include <hip_test_checkers.hh>
#include <hip/hip_runtime_api.h>
#include <resource_guards.hh>
#include <utils.hh>
static constexpr auto NUM_W{16};
static constexpr auto NUM_H{16};
static constexpr auto COLUMNS{8};
static constexpr auto ROWS{8};
TEST_CASE("Unit_hipMemcpy2D_Positive_Basic") {
constexpr bool async = false;
/**
* Test Description
* ------------------------
* - This testcases performs the following scenarios of hipMemcpy2D API on same GPU
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
SECTION("Device to Host") { Memcpy2DDeviceToHostShell<async>(hipMemcpy2D); }
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "B_d" using D2D copy
"B_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy2D_H2D-D2D-D2H", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
// 1 refers to pinned host memory
auto mem_type = GENERATE(0, 1);
HIP_CHECK(hipSetDevice(0));
TestType *A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr},
*B_d{nullptr};
size_t pitch_A, pitch_B;
size_t width{NUM_W * sizeof(TestType)};
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
SECTION("Device to Device") {
SECTION("Peer access disabled") { Memcpy2DDeviceToDeviceShell<async, false>(hipMemcpy2D); }
SECTION("Peer access enabled") { Memcpy2DDeviceToDeviceShell<async, true>(hipMemcpy2D); }
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&B_d),
&pitch_B, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
SECTION("Host to Device") { Memcpy2DHostToDeviceShell<async>(hipMemcpy2D); }
// Host to Device
HIP_CHECK(hipMemcpy2D(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyHostToDevice));
// Performs D2D on same GPU device
HIP_CHECK(hipMemcpy2D(B_d, pitch_B, A_d,
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice));
// hipMemcpy2D Device to Host
HIP_CHECK(hipMemcpy2D(B_h, COLUMNS*sizeof(TestType), B_d, pitch_B,
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost));
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
HIP_CHECK(hipFree(B_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
}
SECTION("Host to Host") { Memcpy2DHostToHostShell<async>(hipMemcpy2D); }
}
/**
* Test Description
* ------------------------
* - This testcase performs the following scenarios of hipMemcpy2D API on same GPU.
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
The src and dst input pointers to hipMemCpy2D add an offset to the pointers
returned by the allocation functions.
TEST_CASE("Unit_hipMemcpy2D_Positive_Synchronization_Behavior") {
HIP_CHECK(hipDeviceSynchronize());
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "B_d" using D2D copy
"B_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
SECTION("Host to Device") { Memcpy2DHtoDSyncBehavior(hipMemcpy2D, true); }
TEMPLATE_TEST_CASE("Unit_hipMemcpy2D_H2D-D2D-D2H_WithOffset", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
// 1 refers to pinned host memory
auto mem_type = GENERATE(0, 1);
HIP_CHECK(hipSetDevice(0));
TestType *A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr},
*B_d{nullptr};
size_t pitch_A, pitch_B;
size_t width{NUM_W * sizeof(TestType)};
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
SECTION("Device to Host") {
Memcpy2DDtoHPageableSyncBehavior(hipMemcpy2D, true);
Memcpy2DDtoHPinnedSyncBehavior(hipMemcpy2D, true);
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&B_d),
&pitch_B, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
// Host to Device
HIP_CHECK(hipMemcpy2D(A_d+COLUMNS*sizeof(TestType), pitch_A, A_h,
COLUMNS*sizeof(TestType), COLUMNS*sizeof(TestType),
ROWS, hipMemcpyHostToDevice));
// Performs D2D on same GPU device
HIP_CHECK(hipMemcpy2D(B_d+COLUMNS*sizeof(TestType), pitch_B,
A_d+COLUMNS*sizeof(TestType),
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice));
// hipMemcpy2D Device to Host
HIP_CHECK(hipMemcpy2D(B_h, COLUMNS*sizeof(TestType),
B_d+COLUMNS*sizeof(TestType), pitch_B,
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost));
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
HIP_CHECK(hipFree(B_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
SECTION("Device to Device") {
#if HT_NVIDIA
Memcpy2DDtoDSyncBehavior(hipMemcpy2D, false);
#else
Memcpy2DDtoDSyncBehavior(hipMemcpy2D, true);
#endif
}
#if HT_NVIDIA // Disabled on AMD due to defect - EXSWHTEC-232
SECTION("Host to Host") { Memcpy2DHtoHSyncBehavior(hipMemcpy2D, true); }
#endif
}
/**
* Test Description
* ------------------------
* - This testcases performs the following scenarios of hipMemcpy2D API on Peer GPU
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
3. Device context change where memory is allocated in GPU-0
and API is trigerred from GPU-1
TEST_CASE("Unit_hipMemcpy2D_Positive_Parameters") {
constexpr bool async = false;
Memcpy2DZeroWidthHeight<async>(hipMemcpy2D);
}
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "X_d" using D2D copy
"X_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEST_CASE("Unit_hipMemcpy2D_Negative_Parameters") {
constexpr size_t cols = 128;
constexpr size_t rows = 128;
TEMPLATE_TEST_CASE("Unit_hipMemcpy2D_multiDevice-D2D", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
auto mem_type = GENERATE(0, 1);
int numDevices = 0;
int canAccessPeer = 0;
TestType* A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(TestType)};
HIP_CHECK(hipGetDeviceCount(&numDevices));
if (numDevices > 1) {
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
if (canAccessPeer) {
HIP_CHECK(hipSetDevice(0));
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
char *X_d{nullptr};
size_t pitch_X;
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&X_d),
&pitch_X, width, NUM_H));
// Change device
HIP_CHECK(hipSetDevice(1));
// Host to Device
HIP_CHECK(hipMemcpy2D(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS, hipMemcpyHostToDevice));
// Device to Device
HIP_CHECK(hipMemcpy2D(X_d, pitch_X, A_d,
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice));
// Device to Host
HIP_CHECK(hipMemcpy2D(B_h, COLUMNS*sizeof(TestType), X_d,
pitch_X, COLUMNS*sizeof(TestType), ROWS, hipMemcpyDeviceToHost));
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
}
HIP_CHECK(hipFree(X_d));
} else {
SUCCEED("Machine does not seem to have P2P");
constexpr auto NegativeTests = [](void* dst, size_t dpitch, const void* src, size_t spitch,
size_t width, size_t height, hipMemcpyKind kind) {
SECTION("dst == nullptr") {
HIP_CHECK_ERROR(hipMemcpy2D(nullptr, dpitch, src, spitch, width, height, kind),
hipErrorInvalidValue);
}
} else {
SUCCEED("skipped the testcase as no of devices is less than 2");
}
}
/**
* Test Description
* ------------------------
* - This Testcase verifies the null size checks of hipMemcpy2D API
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEST_CASE("Unit_hipMemcpy2D_SizeCheck") {
CHECK_IMAGE_SUPPORT
HIP_CHECK(hipSetDevice(0));
int* A_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(int)};
// Allocating memory
HipTest::initArrays<int>(nullptr, nullptr, nullptr,
&A_h, nullptr, nullptr, NUM_W*NUM_H);
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<int>(NUM_W*NUM_H, A_h, nullptr, nullptr);
SECTION("hipMemcpy2D API where Source Pitch is zero") {
REQUIRE(hipMemcpy2D(A_h, 0, A_d,
pitch_A, NUM_W, NUM_H,
hipMemcpyDeviceToHost) != hipSuccess);
}
SECTION("hipMemcpy2D API where Destination Pitch is zero") {
REQUIRE(hipMemcpy2D(A_h, width, A_d,
0, NUM_W, NUM_H,
hipMemcpyDeviceToHost) != hipSuccess);
}
SECTION("hipMemcpy2D API where height is zero") {
REQUIRE(hipMemcpy2D(A_h, width, A_d,
pitch_A, NUM_W, 0,
hipMemcpyDeviceToHost) == hipSuccess);
}
SECTION("hipMemcpy2D API where width is zero") {
REQUIRE(hipMemcpy2D(A_h, width, A_d,
pitch_A, 0, NUM_H,
hipMemcpyDeviceToHost) == hipSuccess);
}
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
free(A_h);
}
/**
* Test Description
* ------------------------
* - This Testcase verifies all the negative scenarios of hipMemcpy2D API
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEST_CASE("Unit_hipMemcpy2D_Negative") {
CHECK_IMAGE_SUPPORT
HIP_CHECK(hipSetDevice(0));
int* A_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(int)};
// Allocating memory
HipTest::initArrays<int>(nullptr, nullptr, nullptr,
&A_h, nullptr, nullptr, NUM_W*NUM_H);
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<int>(NUM_W*NUM_H, A_h, nullptr, nullptr);
SECTION("hipMemcpy2D API by Passing nullptr to destination") {
REQUIRE(hipMemcpy2D(nullptr, width, A_d,
pitch_A, COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost) != hipSuccess);
}
SECTION("hipMemcpy2D API by Passing nullptr to destination") {
REQUIRE(hipMemcpy2D(nullptr, width, nullptr,
pitch_A, COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost) != hipSuccess);
}
SECTION("hipMemcpy2D API where width is greater than destination pitch") {
REQUIRE(hipMemcpy2D(A_h, 10, A_d, pitch_A,
COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost) != hipSuccess);
}
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
free(A_h);
}
static void hipMemcpy2D_Basic_Size_Test(size_t inc) {
constexpr int defaultProgramSize = 256 * 1024 * 1024;
constexpr int N = 2;
constexpr int value = 42;
int *in, *out, *dev;
size_t newSize = 0, inp = 0;
size_t size = sizeof(int) * N * inc;
size_t free, total;
HIP_CHECK(hipMemGetInfo(&free, &total));
if ( free < 2 * size )
newSize = ( free - defaultProgramSize ) / 2;
else
newSize = size;
INFO("Array size: " << size/1024.0/1024.0 << " MB or " << size << " Bytes.");
INFO("Free memory: " << free/1024.0/1024.0 << " MB or " << free << " Bytes");
INFO("NewSize:" << newSize/1024.0/1024.0 << "MB or " << newSize << " Bytes");
HIP_CHECK(hipHostMalloc(&in, newSize));
HIP_CHECK(hipHostMalloc(&out, newSize));
HIP_CHECK(hipMalloc(&dev, newSize));
inp = newSize / (sizeof(int) * N);
for (size_t i=0; i < N; i++) {
in[i * inp] = value;
}
size_t pitch = sizeof(int) * inp;
HIP_CHECK(hipMemcpy2D(dev, pitch, in, pitch, sizeof(int),
N, hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy2D(out, pitch, dev, pitch, sizeof(int),
N, hipMemcpyDeviceToHost));
for (size_t i=0; i < N; i++) {
REQUIRE(out[i * inp] == value);
}
HIP_CHECK(hipFree(dev));
HIP_CHECK(hipHostFree(in));
HIP_CHECK(hipHostFree(out));
}
/**
* Test Description
* ------------------------
* - This testcase performs multidevice size check on hipMemcpy2D API
1. Verify hipMemcpy2D with 1 << 20 size
2. Verify hipMemcpy2D with 1 << 21 size
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEST_CASE("Unit_hipMemcpy2D_multiDevice_Basic_Size_Test") {
CHECK_IMAGE_SUPPORT
size_t input = 1 << 20;
int numDevices = 0;
HIP_CHECK(hipGetDeviceCount(&numDevices));
for (int i=0; i < numDevices; i++) {
HIP_CHECK(hipSetDevice(i));
SECTION("Verify hipMemcpy2D with 1 << 20 size") {
hipMemcpy2D_Basic_Size_Test(input);
SECTION("src == nullptr") {
HIP_CHECK_ERROR(hipMemcpy2D(dst, dpitch, nullptr, spitch, width, height, kind),
hipErrorInvalidValue);
}
SECTION("Verify hipMemcpy2D with 1 << 21 size") {
input <<= 1;
hipMemcpy2D_Basic_Size_Test(input);
SECTION("dpitch < width") {
HIP_CHECK_ERROR(hipMemcpy2D(dst, width - 1, src, spitch, width, height, kind),
hipErrorInvalidPitchValue);
}
SECTION("spitch < width") {
HIP_CHECK_ERROR(hipMemcpy2D(dst, dpitch, src, width - 1, width, height, kind),
hipErrorInvalidPitchValue);
}
SECTION("dpitch > max pitch") {
int attr = 0;
HIP_CHECK(hipDeviceGetAttribute(&attr, hipDeviceAttributeMaxPitch, 0));
HIP_CHECK_ERROR(
hipMemcpy2D(dst, static_cast<size_t>(attr) + 1, src, spitch, width, height, kind),
hipErrorInvalidValue);
}
SECTION("spitch > max pitch") {
int attr = 0;
HIP_CHECK(hipDeviceGetAttribute(&attr, hipDeviceAttributeMaxPitch, 0));
HIP_CHECK_ERROR(
hipMemcpy2D(dst, dpitch, src, static_cast<size_t>(attr) + 1, width, height, kind),
hipErrorInvalidValue);
}
#if HT_NVIDIA // Disabled on AMD due to defect - EXSWHTEC-234
SECTION("Invalid MemcpyKind") {
HIP_CHECK_ERROR(
hipMemcpy2D(dst, dpitch, src, spitch, width, height, static_cast<hipMemcpyKind>(-1)),
hipErrorInvalidMemcpyDirection);
}
#endif
};
SECTION("Host to Device") {
LinearAllocGuard2D<int> device_alloc(cols, rows);
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, device_alloc.pitch() * rows);
NegativeTests(device_alloc.ptr(), device_alloc.pitch(), host_alloc.ptr(), device_alloc.pitch(),
device_alloc.width(), device_alloc.height(), hipMemcpyHostToDevice);
}
SECTION("Device to Host") {
LinearAllocGuard2D<int> device_alloc(cols, rows);
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, device_alloc.pitch() * rows);
NegativeTests(host_alloc.ptr(), device_alloc.pitch(), device_alloc.ptr(), device_alloc.pitch(),
device_alloc.width(), device_alloc.height(), hipMemcpyDeviceToHost);
}
SECTION("Host to Host") {
LinearAllocGuard<int> src_alloc(LinearAllocs::hipHostMalloc, cols * rows * sizeof(int));
LinearAllocGuard<int> dst_alloc(LinearAllocs::hipHostMalloc, cols * rows * sizeof(int));
NegativeTests(dst_alloc.ptr(), cols * sizeof(int), src_alloc.ptr(), cols * sizeof(int),
cols * sizeof(int), rows, hipMemcpyHostToHost);
}
SECTION("Device to Device") {
LinearAllocGuard2D<int> src_alloc(cols, rows);
LinearAllocGuard2D<int> dst_alloc(cols, rows);
NegativeTests(dst_alloc.ptr(), dst_alloc.pitch(), src_alloc.ptr(), src_alloc.pitch(),
dst_alloc.width(), dst_alloc.height(), hipMemcpyDeviceToDevice);
}
}
catch/unit/memory/hipMemcpy2DAsync.cc
+162 -529
مشاهده پرونده
@@ -1,555 +1,188 @@
/*
Copyright (c) 2021-2023 Advanced Micro Devices, Inc. All rights reserved.
Copyright (c) 2022 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 WARRANNTY OF ANY KIND, EXPRESS OR
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
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 INN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR INN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
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.
*/
/**
* @addtogroup hipMemcpy2DAsync hipMemcpy2DAsync
* @{
* @ingroup MemcpyTest
* `hipMemcpy2DAsync(void* dst, size_t dpitch, const void* src,
* size_t spitch, size_t width, size_t height,
* hipMemcpyKind kind, hipStream_t stream = 0 )` -
* Copies data between host and device.
*/
// Testcase Description:
// 1) Verifies the working of Memcpy2DAsync API negative scenarios by
// Pass NULL to destination pointer
// Pass NULL to Source pointer
// Pass width greater than spitch/dpitch
// 2) Verifies hipMemcpy2DAsync API by
// pass 0 to destionation pitch
// pass 0 to source pitch
// pass 0 to width
// pass 0 to height
// 3) Verifies working of Memcpy2DAsync API on host memory
// and pinned host memory by
// performing D2H, D2D and H2D memory kind copies on same GPU
// 4) Verifies working of Memcpy2DAsync API on host memory
// and pinned host memory by
// performing D2H, D2D and H2D memory kind copies on peer GPU
// 5) Verifies working of Memcpy2DAsync API where memory is allocated
// in GPU-0 and stream is created on GPU-1
#include "memcpy2d_tests_common.hh"
#include <hip_test_common.hh>
#include <hip_test_checkers.hh>
#include <hip/hip_runtime_api.h>
#include <resource_guards.hh>
#include <utils.hh>
static constexpr auto NUM_W{16};
static constexpr auto NUM_H{16};
static constexpr auto COLUMNS{6};
static constexpr auto ROWS{6};
TEST_CASE("Unit_hipMemcpy2DAsync_Positive_Basic") {
using namespace std::placeholders;
/**
* Test Description
* ------------------------
* - This performs the following scenarios of hipMemcpy2DAsync API on same GPU
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
constexpr bool async = true;
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "B_d" using D2D copy
"B_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 5.2
*/
const auto stream_type = GENERATE(Streams::nullstream, Streams::perThread, Streams::created);
const StreamGuard stream_guard(stream_type);
const hipStream_t stream = stream_guard.stream();
TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_Host&PinnedMem", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
// 1 refers to pinned host memory
auto mem_type = GENERATE(0, 1);
HIP_CHECK(hipSetDevice(0));
TestType *A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr},
*B_d{nullptr};
size_t pitch_A, pitch_B;
size_t width{NUM_W * sizeof(TestType)};
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&B_d),
&pitch_B, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
SECTION("Calling Async apis with stream object created by user") {
// Host to Device
HIP_CHECK(hipMemcpy2DAsync(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyHostToDevice, stream));
// Performs D2D on same GPU device
HIP_CHECK(hipMemcpy2DAsync(B_d, pitch_B, A_d,
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice, stream));
// hipMemcpy2DAsync Device to Host
HIP_CHECK(hipMemcpy2DAsync(B_h, COLUMNS*sizeof(TestType), B_d, pitch_B,
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost, stream));
HIP_CHECK(hipStreamSynchronize(stream));
}
SECTION("Calling Async apis with hipStreamPerThread") {
// Host to Device
HIP_CHECK(hipMemcpy2DAsync(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyHostToDevice, hipStreamPerThread));
// Performs D2D on same GPU device
HIP_CHECK(hipMemcpy2DAsync(B_d, pitch_B, A_d, pitch_A,
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToDevice, hipStreamPerThread));
// hipMemcpy2DAsync Device to Host
HIP_CHECK(hipMemcpy2DAsync(B_h, COLUMNS*sizeof(TestType), B_d, pitch_B,
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost, hipStreamPerThread));
HIP_CHECK(hipStreamSynchronize(hipStreamPerThread));
SECTION("Device to Host") {
Memcpy2DDeviceToHostShell<async>(
std::bind(hipMemcpy2DAsync, _1, _2, _3, _4, _5, _6, _7, stream), stream);
}
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
HIP_CHECK(hipFree(B_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
}
HIP_CHECK(hipStreamDestroy(stream));
}
/**
* Test Description
* ------------------------
* - This testcases performs the following scenarios of hipMemcpy2DAsync API on Peer GPU
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "X_d" using D2D copy
"X_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 5.2
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice-Host&PinnedMem", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
auto mem_type = GENERATE(0, 1);
int numDevices = 0;
int canAccessPeer = 0;
TestType* A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(TestType)};
HIP_CHECK(hipGetDeviceCount(&numDevices));
hipStream_t stream;
if (numDevices > 1) {
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
if (canAccessPeer) {
HIP_CHECK(hipSetDevice(0));
HIP_CHECK(hipStreamCreate(&stream));
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
// Host to Device
HIP_CHECK(hipMemcpy2DAsync(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS, hipMemcpyHostToDevice, stream));
// Change device
HIP_CHECK(hipSetDevice(1));
char *X_d{nullptr};
size_t pitch_X;
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&X_d),
&pitch_X, width, NUM_H));
// Device to Device
HIP_CHECK(hipMemcpy2DAsync(X_d, pitch_X, A_d,
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice, stream));
// Device to Host
HIP_CHECK(hipMemcpy2DAsync(B_h, COLUMNS*sizeof(TestType), X_d,
pitch_X, COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost, stream));
HIP_CHECK(hipStreamSynchronize(stream));
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
}
HIP_CHECK(hipFree(X_d));
HIP_CHECK(hipStreamDestroy(stream));
} else {
SUCCEED("Machine does not seem to have P2P");
SECTION("Device to Device") {
SECTION("Peer access disabled") {
Memcpy2DDeviceToDeviceShell<async, false>(
std::bind(hipMemcpy2DAsync, _1, _2, _3, _4, _5, _6, _7, stream), stream);
}
} else {
SUCCEED("skipped the testcase as no of devices is less than 2");
}
}
/**
* Test Description
* ------------------------
* - This testcases performs the following scenarios of hipMemcpy2DAsync API on Peer GPU
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
Memory is allocated in GPU-0 and Stream is created in GPU-1
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "X_d" using D2D copy
"X_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 5.2
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice-StreamOnDiffDevice", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
auto mem_type = GENERATE(0, 1);
int numDevices = 0;
int canAccessPeer = 0;
TestType* A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(TestType)};
HIP_CHECK(hipGetDeviceCount(&numDevices));
hipStream_t stream;
if (numDevices > 1) {
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
if (canAccessPeer) {
HIP_CHECK(hipSetDevice(0));
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
char *X_d{nullptr};
size_t pitch_X;
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&X_d),
&pitch_X, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
// Change device
HIP_CHECK(hipSetDevice(1));
HIP_CHECK(hipStreamCreate(&stream));
// Host to Device
HIP_CHECK(hipMemcpy2DAsync(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS, hipMemcpyHostToDevice, stream));
// Device to Device
HIP_CHECK(hipMemcpy2DAsync(X_d, pitch_X, A_d,
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice, stream));
// Device to Host
HIP_CHECK(hipMemcpy2DAsync(B_h, COLUMNS*sizeof(TestType), X_d,
pitch_X, COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost, stream));
HIP_CHECK(hipStreamSynchronize(stream));
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
}
HIP_CHECK(hipFree(X_d));
HIP_CHECK(hipStreamDestroy(stream));
} else {
SUCCEED("Machine does not seem to have P2P");
}
} else {
SUCCEED("skipped the testcase as no of devices is less than 2");
}
}
/**
* Test Description
* ------------------------
* - This testcase verifies the null checks of hipMemcpy2DAsync API
1. hipMemcpy2DAsync API where Source Pitch is zero
2. hipMemcpy2DAsync API where Destination Pitch is zero
3. hipMemcpy2DAsync API where height is zero
4. hipMemcpy2DAsync API where width is zero
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 5.2
*/
TEST_CASE("Unit_hipMemcpy2DAsync_SizeCheck") {
CHECK_IMAGE_SUPPORT
HIP_CHECK(hipSetDevice(0));
int* A_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(int)};
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
// Allocating memory
HipTest::initArrays<int>(nullptr, nullptr, nullptr,
&A_h, nullptr, nullptr, NUM_W*NUM_H);
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<int>(NUM_W*NUM_H, A_h, nullptr, nullptr);
SECTION("hipMemcpy2DAsync API where Source Pitch is zero") {
REQUIRE(hipMemcpy2DAsync(A_h, 0, A_d,
pitch_A, NUM_W, NUM_H,
hipMemcpyDeviceToHost, stream) != hipSuccess);
}
SECTION("hipMemcpy2DAsync API where Destination Pitch is zero") {
REQUIRE(hipMemcpy2DAsync(A_h, width, A_d,
0, NUM_W, NUM_H,
hipMemcpyDeviceToHost, stream) != hipSuccess);
}
SECTION("hipMemcpy2DAsync API where height is zero") {
REQUIRE(hipMemcpy2DAsync(A_h, width, A_d,
pitch_A, NUM_W, 0,
hipMemcpyDeviceToHost, stream) == hipSuccess);
}
SECTION("hipMemcpy2DAsync API where width is zero") {
REQUIRE(hipMemcpy2DAsync(A_h, width, A_d,
pitch_A, 0, NUM_H,
hipMemcpyDeviceToHost, stream) == hipSuccess);
}
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
free(A_h);
}
/**
* Test Description
* ------------------------
* - This testcase performs the negative scenarios of hipMemcpy2DAsync API
1. hipMemcpy2DAsync API by Passing nullptr to destination
2. hipMemcpy2DAsync API by Passing nullptr to source
3. hipMemcpy2DAsync API where width is > destination pitch
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 5.2
*/
TEST_CASE("Unit_hipMemcpy2DAsync_Negative") {
CHECK_IMAGE_SUPPORT
HIP_CHECK(hipSetDevice(0));
int* A_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(int)};
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
// Allocating memory
HipTest::initArrays<int>(nullptr, nullptr, nullptr,
&A_h, nullptr, nullptr, NUM_W*NUM_H);
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<int>(NUM_W*NUM_H, A_h, nullptr, nullptr);
SECTION("hipMemcpy2DAsync API by Passing nullptr to destination") {
REQUIRE(hipMemcpy2DAsync(nullptr, width, A_d,
pitch_A, COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost, stream) != hipSuccess);
}
SECTION("hipMemcpy2DAsync API by Passing nullptr to source") {
REQUIRE(hipMemcpy2DAsync(A_h, width, nullptr,
pitch_A, COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost, stream) != hipSuccess);
}
SECTION("hipMemcpy2DAsync API where width is > destination pitch") {
REQUIRE(hipMemcpy2DAsync(A_h, 10, A_d, pitch_A,
COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost, stream) != hipSuccess);
}
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
HIP_CHECK(hipStreamDestroy(stream));
free(A_h);
}
static void hipMemcpy2DAsync_Basic_Size_Test(size_t inc) {
constexpr int defaultProgramSize = 256 * 1024 * 1024;
constexpr int N = 2;
constexpr int value = 42;
int *in, *out, *dev;
size_t newSize = 0, inp = 0;
size_t size = sizeof(int) * N * inc;
size_t free, total;
HIP_CHECK(hipMemGetInfo(&free, &total));
if ( free < 2 * size )
newSize = ( free - defaultProgramSize ) / 2;
else
newSize = size;
INFO("Array size: " << size/1024.0/1024.0 << " MB or " << size << " Bytes.");
INFO("Free memory: " << free/1024.0/1024.0 << " MB or " << free << " Bytes");
INFO("NewSize:" << newSize/1024.0/1024.0 << "MB or " << newSize << " Bytes");
HIP_CHECK(hipHostMalloc(&in, newSize));
HIP_CHECK(hipHostMalloc(&out, newSize));
HIP_CHECK(hipMalloc(&dev, newSize));
inp = newSize / (sizeof(int) * N);
for (size_t i=0; i < N; i++) {
in[i * inp] = value;
}
size_t pitch = sizeof(int) * inp;
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
HIP_CHECK(hipMemcpy2DAsync(dev, pitch, in, pitch, sizeof(int),
N, hipMemcpyHostToDevice, stream));
HIP_CHECK(hipMemcpy2DAsync(out, pitch, dev, pitch, sizeof(int),
N, hipMemcpyDeviceToHost, stream));
HIP_CHECK(hipStreamSynchronize(stream));
for (size_t i=0; i < N; i++) {
REQUIRE(out[i * inp] == value);
}
HIP_CHECK(hipFree(dev));
HIP_CHECK(hipHostFree(in));
HIP_CHECK(hipHostFree(out));
HIP_CHECK(hipStreamDestroy(stream));
}
/**
* Test Description
* ------------------------
* - This testcase performs multidevice size check on hipMemcpy2DAsync API
1. Verify hipMemcpy2DAsync with 1 << 20 size
2. Verify hipMemcpy2DAsync with 1 << 21 size
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice_Basic_Size_Test") {
CHECK_IMAGE_SUPPORT
size_t input = 1 << 20;
int numDevices = 0;
HIP_CHECK(hipGetDeviceCount(&numDevices));
for (int i=0; i < numDevices; i++) {
HIP_CHECK(hipSetDevice(i));
SECTION("Verify hipMemcpy2DAsync with 1 << 20 size") {
hipMemcpy2DAsync_Basic_Size_Test(input);
}
SECTION("Verify hipMemcpy2DAsync with 1 << 21 size") {
input <<= 1;
hipMemcpy2DAsync_Basic_Size_Test(input);
SECTION("Peer access enabled") {
Memcpy2DDeviceToDeviceShell<async, true>(
std::bind(hipMemcpy2DAsync, _1, _2, _3, _4, _5, _6, _7, stream), stream);
}
}
SECTION("Host to Device") {
Memcpy2DHostToDeviceShell<async>(
std::bind(hipMemcpy2DAsync, _1, _2, _3, _4, _5, _6, _7, stream), stream);
}
SECTION("Host to Host") {
Memcpy2DHostToHostShell<async>(std::bind(hipMemcpy2DAsync, _1, _2, _3, _4, _5, _6, _7, stream),
stream);
}
}
TEST_CASE("Unit_hipMemcpy2DAsync_Positive_Synchronization_Behavior") {
using namespace std::placeholders;
HIP_CHECK(hipDeviceSynchronize());
SECTION("Host to Device") {
Memcpy2DHtoDSyncBehavior(std::bind(hipMemcpy2DAsync, _1, _2, _3, _4, _5, _6, _7, nullptr),
false);
}
#if HT_NVIDIA // Disabled on AMD due to defect - EXSWHTEC-233
SECTION("Device to Pageable Host") {
Memcpy2DDtoHPageableSyncBehavior(
std::bind(hipMemcpy2DAsync, _1, _2, _3, _4, _5, _6, _7, nullptr), true);
}
#endif
SECTION("Device to Pinned Host") {
Memcpy2DDtoHPinnedSyncBehavior(std::bind(hipMemcpy2DAsync, _1, _2, _3, _4, _5, _6, _7, nullptr),
false);
}
SECTION("Device to Device") {
Memcpy2DDtoDSyncBehavior(std::bind(hipMemcpy2DAsync, _1, _2, _3, _4, _5, _6, _7, nullptr),
false);
}
#if HT_NVIDIA // Disabled on AMD due to defect - EXSWHTEC-233
SECTION("Host to Host") {
Memcpy2DHtoHSyncBehavior(std::bind(hipMemcpy2DAsync, _1, _2, _3, _4, _5, _6, _7, nullptr),
true);
}
#endif
}
TEST_CASE("Unit_hipMemcpy2DAsync_Positive_Parameters") {
using namespace std::placeholders;
constexpr bool async = true;
Memcpy2DZeroWidthHeight<async>(std::bind(hipMemcpy2DAsync, _1, _2, _3, _4, _5, _6, _7, nullptr));
}
TEST_CASE("Unit_hipMemcpy2DAsync_Negative_Parameters") {
constexpr size_t cols = 128;
constexpr size_t rows = 128;
constexpr auto NegativeTests = [](void* dst, size_t dpitch, const void* src, size_t spitch,
size_t width, size_t height, hipMemcpyKind kind) {
SECTION("dst == nullptr") {
HIP_CHECK_ERROR(hipMemcpy2DAsync(nullptr, dpitch, src, spitch, width, height, kind, nullptr),
hipErrorInvalidValue);
}
SECTION("src == nullptr") {
HIP_CHECK_ERROR(hipMemcpy2DAsync(dst, dpitch, nullptr, spitch, width, height, kind, nullptr),
hipErrorInvalidValue);
}
SECTION("dpitch < width") {
HIP_CHECK_ERROR(hipMemcpy2DAsync(dst, width - 1, src, spitch, width, height, kind, nullptr),
hipErrorInvalidPitchValue);
}
SECTION("spitch < width") {
HIP_CHECK_ERROR(hipMemcpy2DAsync(dst, dpitch, src, width - 1, width, height, kind, nullptr),
hipErrorInvalidPitchValue);
}
SECTION("dpitch > max pitch") {
int attr = 0;
HIP_CHECK(hipDeviceGetAttribute(&attr, hipDeviceAttributeMaxPitch, 0));
HIP_CHECK_ERROR(hipMemcpy2DAsync(dst, static_cast<size_t>(attr) + 1, src, spitch, width,
height, kind, nullptr),
hipErrorInvalidValue);
}
SECTION("spitch > max pitch") {
int attr = 0;
HIP_CHECK(hipDeviceGetAttribute(&attr, hipDeviceAttributeMaxPitch, 0));
HIP_CHECK_ERROR(hipMemcpy2DAsync(dst, dpitch, src, static_cast<size_t>(attr) + 1, width,
height, kind, nullptr),
hipErrorInvalidValue);
}
#if HT_NVIDIA // Disabled on AMD due to defect - EXSWHTEC-234
SECTION("Invalid MemcpyKind") {
HIP_CHECK_ERROR(hipMemcpy2DAsync(dst, dpitch, src, spitch, width, height,
static_cast<hipMemcpyKind>(-1), nullptr),
hipErrorInvalidMemcpyDirection);
}
#endif
#if HT_NVIDIA // Disabled on AMD due to defect - EXSWHTEC-235
SECTION("Invalid stream") {
StreamGuard stream_guard(Streams::created);
HIP_CHECK(hipStreamDestroy(stream_guard.stream()));
HIP_CHECK_ERROR(
hipMemcpy2DAsync(dst, dpitch, src, spitch, width, height, kind, stream_guard.stream()),
hipErrorContextIsDestroyed);
}
#endif
};
SECTION("Host to device") {
LinearAllocGuard2D<int> device_alloc(cols, rows);
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, device_alloc.pitch() * rows);
NegativeTests(device_alloc.ptr(), device_alloc.pitch(), host_alloc.ptr(), device_alloc.pitch(),
device_alloc.width(), device_alloc.height(), hipMemcpyHostToDevice);
}
SECTION("Device to host") {
LinearAllocGuard2D<int> device_alloc(cols, rows);
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, device_alloc.pitch() * rows);
NegativeTests(host_alloc.ptr(), device_alloc.pitch(), device_alloc.ptr(), device_alloc.pitch(),
device_alloc.width(), device_alloc.height(), hipMemcpyDeviceToHost);
}
SECTION("Host to host") {
LinearAllocGuard<int> src_alloc(LinearAllocs::hipHostMalloc, cols * rows * sizeof(int));
LinearAllocGuard<int> dst_alloc(LinearAllocs::hipHostMalloc, cols * rows * sizeof(int));
NegativeTests(dst_alloc.ptr(), cols * sizeof(int), src_alloc.ptr(), cols * sizeof(int),
cols * sizeof(int), rows, hipMemcpyHostToHost);
}
SECTION("Device to device") {
LinearAllocGuard2D<int> src_alloc(cols, rows);
LinearAllocGuard2D<int> dst_alloc(cols, rows);
NegativeTests(dst_alloc.ptr(), dst_alloc.pitch(), src_alloc.ptr(), src_alloc.pitch(),
dst_alloc.width(), dst_alloc.height(), hipMemcpyDeviceToDevice);
}
}
catch/unit/memory/hipMemcpy2DAsync_old.cc
+555
مشاهده پرونده
@@ -0,0 +1,555 @@
/*
Copyright (c) 2021-2023 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 WARRANNTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNNESS 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 INN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR INN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.
*/
/**
* @addtogroup hipMemcpy2DAsync hipMemcpy2DAsync
* @{
* @ingroup MemcpyTest
* `hipMemcpy2DAsync(void* dst, size_t dpitch, const void* src,
* size_t spitch, size_t width, size_t height,
* hipMemcpyKind kind, hipStream_t stream = 0 )` -
* Copies data between host and device.
*/
// Testcase Description:
// 1) Verifies the working of Memcpy2DAsync API negative scenarios by
// Pass NULL to destination pointer
// Pass NULL to Source pointer
// Pass width greater than spitch/dpitch
// 2) Verifies hipMemcpy2DAsync API by
// pass 0 to destionation pitch
// pass 0 to source pitch
// pass 0 to width
// pass 0 to height
// 3) Verifies working of Memcpy2DAsync API on host memory
// and pinned host memory by
// performing D2H, D2D and H2D memory kind copies on same GPU
// 4) Verifies working of Memcpy2DAsync API on host memory
// and pinned host memory by
// performing D2H, D2D and H2D memory kind copies on peer GPU
// 5) Verifies working of Memcpy2DAsync API where memory is allocated
// in GPU-0 and stream is created on GPU-1
#include <hip_test_common.hh>
#include <hip_test_checkers.hh>
static constexpr auto NUM_W{16};
static constexpr auto NUM_H{16};
static constexpr auto COLUMNS{6};
static constexpr auto ROWS{6};
/**
* Test Description
* ------------------------
* - This performs the following scenarios of hipMemcpy2DAsync API on same GPU
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "B_d" using D2D copy
"B_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 5.2
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_Host&PinnedMem", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
// 1 refers to pinned host memory
auto mem_type = GENERATE(0, 1);
HIP_CHECK(hipSetDevice(0));
TestType *A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr},
*B_d{nullptr};
size_t pitch_A, pitch_B;
size_t width{NUM_W * sizeof(TestType)};
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&B_d),
&pitch_B, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
SECTION("Calling Async apis with stream object created by user") {
// Host to Device
HIP_CHECK(hipMemcpy2DAsync(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyHostToDevice, stream));
// Performs D2D on same GPU device
HIP_CHECK(hipMemcpy2DAsync(B_d, pitch_B, A_d,
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice, stream));
// hipMemcpy2DAsync Device to Host
HIP_CHECK(hipMemcpy2DAsync(B_h, COLUMNS*sizeof(TestType), B_d, pitch_B,
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost, stream));
HIP_CHECK(hipStreamSynchronize(stream));
}
SECTION("Calling Async apis with hipStreamPerThread") {
// Host to Device
HIP_CHECK(hipMemcpy2DAsync(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyHostToDevice, hipStreamPerThread));
// Performs D2D on same GPU device
HIP_CHECK(hipMemcpy2DAsync(B_d, pitch_B, A_d, pitch_A,
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToDevice, hipStreamPerThread));
// hipMemcpy2DAsync Device to Host
HIP_CHECK(hipMemcpy2DAsync(B_h, COLUMNS*sizeof(TestType), B_d, pitch_B,
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost, hipStreamPerThread));
HIP_CHECK(hipStreamSynchronize(hipStreamPerThread));
}
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
HIP_CHECK(hipFree(B_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
}
HIP_CHECK(hipStreamDestroy(stream));
}
/**
* Test Description
* ------------------------
* - This testcases performs the following scenarios of hipMemcpy2DAsync API on Peer GPU
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "X_d" using D2D copy
"X_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 5.2
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice-Host&PinnedMem", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
auto mem_type = GENERATE(0, 1);
int numDevices = 0;
int canAccessPeer = 0;
TestType* A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(TestType)};
HIP_CHECK(hipGetDeviceCount(&numDevices));
hipStream_t stream;
if (numDevices > 1) {
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
if (canAccessPeer) {
HIP_CHECK(hipSetDevice(0));
HIP_CHECK(hipStreamCreate(&stream));
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
// Host to Device
HIP_CHECK(hipMemcpy2DAsync(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS, hipMemcpyHostToDevice, stream));
// Change device
HIP_CHECK(hipSetDevice(1));
char *X_d{nullptr};
size_t pitch_X;
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&X_d),
&pitch_X, width, NUM_H));
// Device to Device
HIP_CHECK(hipMemcpy2DAsync(X_d, pitch_X, A_d,
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice, stream));
// Device to Host
HIP_CHECK(hipMemcpy2DAsync(B_h, COLUMNS*sizeof(TestType), X_d,
pitch_X, COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost, stream));
HIP_CHECK(hipStreamSynchronize(stream));
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
}
HIP_CHECK(hipFree(X_d));
HIP_CHECK(hipStreamDestroy(stream));
} else {
SUCCEED("Machine does not seem to have P2P");
}
} else {
SUCCEED("skipped the testcase as no of devices is less than 2");
}
}
/**
* Test Description
* ------------------------
* - This testcases performs the following scenarios of hipMemcpy2DAsync API on Peer GPU
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
Memory is allocated in GPU-0 and Stream is created in GPU-1
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "X_d" using D2D copy
"X_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 5.2
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice-StreamOnDiffDevice", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
auto mem_type = GENERATE(0, 1);
int numDevices = 0;
int canAccessPeer = 0;
TestType* A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(TestType)};
HIP_CHECK(hipGetDeviceCount(&numDevices));
hipStream_t stream;
if (numDevices > 1) {
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
if (canAccessPeer) {
HIP_CHECK(hipSetDevice(0));
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
char *X_d{nullptr};
size_t pitch_X;
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&X_d),
&pitch_X, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
// Change device
HIP_CHECK(hipSetDevice(1));
HIP_CHECK(hipStreamCreate(&stream));
// Host to Device
HIP_CHECK(hipMemcpy2DAsync(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS, hipMemcpyHostToDevice, stream));
// Device to Device
HIP_CHECK(hipMemcpy2DAsync(X_d, pitch_X, A_d,
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice, stream));
// Device to Host
HIP_CHECK(hipMemcpy2DAsync(B_h, COLUMNS*sizeof(TestType), X_d,
pitch_X, COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost, stream));
HIP_CHECK(hipStreamSynchronize(stream));
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
}
HIP_CHECK(hipFree(X_d));
HIP_CHECK(hipStreamDestroy(stream));
} else {
SUCCEED("Machine does not seem to have P2P");
}
} else {
SUCCEED("skipped the testcase as no of devices is less than 2");
}
}
/**
* Test Description
* ------------------------
* - This testcase verifies the null checks of hipMemcpy2DAsync API
1. hipMemcpy2DAsync API where Source Pitch is zero
2. hipMemcpy2DAsync API where Destination Pitch is zero
3. hipMemcpy2DAsync API where height is zero
4. hipMemcpy2DAsync API where width is zero
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 5.2
*/
TEST_CASE("Unit_hipMemcpy2DAsync_SizeCheck") {
CHECK_IMAGE_SUPPORT
HIP_CHECK(hipSetDevice(0));
int* A_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(int)};
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
// Allocating memory
HipTest::initArrays<int>(nullptr, nullptr, nullptr,
&A_h, nullptr, nullptr, NUM_W*NUM_H);
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<int>(NUM_W*NUM_H, A_h, nullptr, nullptr);
SECTION("hipMemcpy2DAsync API where Source Pitch is zero") {
REQUIRE(hipMemcpy2DAsync(A_h, 0, A_d,
pitch_A, NUM_W, NUM_H,
hipMemcpyDeviceToHost, stream) != hipSuccess);
}
SECTION("hipMemcpy2DAsync API where Destination Pitch is zero") {
REQUIRE(hipMemcpy2DAsync(A_h, width, A_d,
0, NUM_W, NUM_H,
hipMemcpyDeviceToHost, stream) != hipSuccess);
}
SECTION("hipMemcpy2DAsync API where height is zero") {
REQUIRE(hipMemcpy2DAsync(A_h, width, A_d,
pitch_A, NUM_W, 0,
hipMemcpyDeviceToHost, stream) == hipSuccess);
}
SECTION("hipMemcpy2DAsync API where width is zero") {
REQUIRE(hipMemcpy2DAsync(A_h, width, A_d,
pitch_A, 0, NUM_H,
hipMemcpyDeviceToHost, stream) == hipSuccess);
}
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
free(A_h);
}
/**
* Test Description
* ------------------------
* - This testcase performs the negative scenarios of hipMemcpy2DAsync API
1. hipMemcpy2DAsync API by Passing nullptr to destination
2. hipMemcpy2DAsync API by Passing nullptr to source
3. hipMemcpy2DAsync API where width is > destination pitch
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 5.2
*/
TEST_CASE("Unit_hipMemcpy2DAsync_Negative") {
CHECK_IMAGE_SUPPORT
HIP_CHECK(hipSetDevice(0));
int* A_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(int)};
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
// Allocating memory
HipTest::initArrays<int>(nullptr, nullptr, nullptr,
&A_h, nullptr, nullptr, NUM_W*NUM_H);
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<int>(NUM_W*NUM_H, A_h, nullptr, nullptr);
SECTION("hipMemcpy2DAsync API by Passing nullptr to destination") {
REQUIRE(hipMemcpy2DAsync(nullptr, width, A_d,
pitch_A, COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost, stream) != hipSuccess);
}
SECTION("hipMemcpy2DAsync API by Passing nullptr to source") {
REQUIRE(hipMemcpy2DAsync(A_h, width, nullptr,
pitch_A, COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost, stream) != hipSuccess);
}
SECTION("hipMemcpy2DAsync API where width is > destination pitch") {
REQUIRE(hipMemcpy2DAsync(A_h, 10, A_d, pitch_A,
COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost, stream) != hipSuccess);
}
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
HIP_CHECK(hipStreamDestroy(stream));
free(A_h);
}
static void hipMemcpy2DAsync_Basic_Size_Test(size_t inc) {
constexpr int defaultProgramSize = 256 * 1024 * 1024;
constexpr int N = 2;
constexpr int value = 42;
int *in, *out, *dev;
size_t newSize = 0, inp = 0;
size_t size = sizeof(int) * N * inc;
size_t free, total;
HIP_CHECK(hipMemGetInfo(&free, &total));
if ( free < 2 * size )
newSize = ( free - defaultProgramSize ) / 2;
else
newSize = size;
INFO("Array size: " << size/1024.0/1024.0 << " MB or " << size << " Bytes.");
INFO("Free memory: " << free/1024.0/1024.0 << " MB or " << free << " Bytes");
INFO("NewSize:" << newSize/1024.0/1024.0 << "MB or " << newSize << " Bytes");
HIP_CHECK(hipHostMalloc(&in, newSize));
HIP_CHECK(hipHostMalloc(&out, newSize));
HIP_CHECK(hipMalloc(&dev, newSize));
inp = newSize / (sizeof(int) * N);
for (size_t i=0; i < N; i++) {
in[i * inp] = value;
}
size_t pitch = sizeof(int) * inp;
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
HIP_CHECK(hipMemcpy2DAsync(dev, pitch, in, pitch, sizeof(int),
N, hipMemcpyHostToDevice, stream));
HIP_CHECK(hipMemcpy2DAsync(out, pitch, dev, pitch, sizeof(int),
N, hipMemcpyDeviceToHost, stream));
HIP_CHECK(hipStreamSynchronize(stream));
for (size_t i=0; i < N; i++) {
REQUIRE(out[i * inp] == value);
}
HIP_CHECK(hipFree(dev));
HIP_CHECK(hipHostFree(in));
HIP_CHECK(hipHostFree(out));
HIP_CHECK(hipStreamDestroy(stream));
}
/**
* Test Description
* ------------------------
* - This testcase performs multidevice size check on hipMemcpy2DAsync API
1. Verify hipMemcpy2DAsync with 1 << 20 size
2. Verify hipMemcpy2DAsync with 1 << 21 size
* Test source
* ------------------------
* - unit/memory/hipMemcpy2DAsync.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEST_CASE("Unit_hipMemcpy2DAsync_multiDevice_Basic_Size_Test") {
CHECK_IMAGE_SUPPORT
size_t input = 1 << 20;
int numDevices = 0;
HIP_CHECK(hipGetDeviceCount(&numDevices));
for (int i=0; i < numDevices; i++) {
HIP_CHECK(hipSetDevice(i));
SECTION("Verify hipMemcpy2DAsync with 1 << 20 size") {
hipMemcpy2DAsync_Basic_Size_Test(input);
}
SECTION("Verify hipMemcpy2DAsync with 1 << 21 size") {
input <<= 1;
hipMemcpy2DAsync_Basic_Size_Test(input);
}
}
}
catch/unit/memory/hipMemcpy2D_old.cc
+496
مشاهده پرونده
@@ -0,0 +1,496 @@
/*
Copyright (c) 2021-2023 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 WARRANNTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNNESS 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 INN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR INN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.
*/
/**
* @addtogroup hipMemcpy2D hipMemcpy2D
* @{
* @ingroup MemcpyTest
* `hipMemcpy2D(void* dst, size_t dpitch, const void* src,
* size_t spitch, size_t width, size_t height,
* hipMemcpyKind kind)` -
* Copies data between host and device.
*/
// Testcase Description:
// 1) Verifies the working of Memcpy2D API negative scenarios by
// Pass NULL to destination pointer
// Pass NULL to Source pointer
// Pass width greater than spitch/dpitch
// 2) Verifies hipMemcpy2D API by
// pass 0 to destionation pitch
// pass 0 to source pitch
// pass 0 to width
// pass 0 to height
// 3) Verifies working of Memcpy2D API on host memory and pinned host memory by
// performing D2H, D2D and H2D memory kind copies on same GPU
// 4) Verifies working of Memcpy2D API for the following scenarios
// H2D-D2D-D2H on host and device memory
// H2D-D2D-D2H on pinned host and device memory
// H2D-D2D-D2H functionalities where memory is allocated in GPU-0
// and API is triggered from GPU-1
#include <hip_test_common.hh>
#include <hip_test_checkers.hh>
static constexpr auto NUM_W{16};
static constexpr auto NUM_H{16};
static constexpr auto COLUMNS{8};
static constexpr auto ROWS{8};
/**
* Test Description
* ------------------------
* - This testcases performs the following scenarios of hipMemcpy2D API on same GPU
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "B_d" using D2D copy
"B_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy2D_H2D-D2D-D2H", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
// 1 refers to pinned host memory
auto mem_type = GENERATE(0, 1);
HIP_CHECK(hipSetDevice(0));
TestType *A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr},
*B_d{nullptr};
size_t pitch_A, pitch_B;
size_t width{NUM_W * sizeof(TestType)};
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&B_d),
&pitch_B, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
// Host to Device
HIP_CHECK(hipMemcpy2D(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyHostToDevice));
// Performs D2D on same GPU device
HIP_CHECK(hipMemcpy2D(B_d, pitch_B, A_d,
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice));
// hipMemcpy2D Device to Host
HIP_CHECK(hipMemcpy2D(B_h, COLUMNS*sizeof(TestType), B_d, pitch_B,
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost));
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
HIP_CHECK(hipFree(B_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
}
}
/**
* Test Description
* ------------------------
* - This testcase performs the following scenarios of hipMemcpy2D API on same GPU.
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
The src and dst input pointers to hipMemCpy2D add an offset to the pointers
returned by the allocation functions.
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "B_d" using D2D copy
"B_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy2D_H2D-D2D-D2H_WithOffset", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
// 1 refers to pinned host memory
auto mem_type = GENERATE(0, 1);
HIP_CHECK(hipSetDevice(0));
TestType *A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr},
*B_d{nullptr};
size_t pitch_A, pitch_B;
size_t width{NUM_W * sizeof(TestType)};
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&B_d),
&pitch_B, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
// Host to Device
HIP_CHECK(hipMemcpy2D(A_d+COLUMNS*sizeof(TestType), pitch_A, A_h,
COLUMNS*sizeof(TestType), COLUMNS*sizeof(TestType),
ROWS, hipMemcpyHostToDevice));
// Performs D2D on same GPU device
HIP_CHECK(hipMemcpy2D(B_d+COLUMNS*sizeof(TestType), pitch_B,
A_d+COLUMNS*sizeof(TestType),
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice));
// hipMemcpy2D Device to Host
HIP_CHECK(hipMemcpy2D(B_h, COLUMNS*sizeof(TestType),
B_d+COLUMNS*sizeof(TestType), pitch_B,
COLUMNS*sizeof(TestType), ROWS,
hipMemcpyDeviceToHost));
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
HIP_CHECK(hipFree(B_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
}
}
/**
* Test Description
* ------------------------
* - This testcases performs the following scenarios of hipMemcpy2D API on Peer GPU
1. H2D-D2D-D2H for Host Memory<-->Device Memory
2. H2D-D2D-D2H for Pinned Host Memory<-->Device Memory
3. Device context change where memory is allocated in GPU-0
and API is trigerred from GPU-1
Input : "A_h" initialized based on data type
"A_h" --> "A_d" using H2D copy
"A_d" --> "X_d" using D2D copy
"X_d" --> "B_h" using D2H copy
Output: Validating A_h with B_h both should be equal for
the number of COLUMNS and ROWS copied
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy2D_multiDevice-D2D", ""
, int, float, double) {
CHECK_IMAGE_SUPPORT
auto mem_type = GENERATE(0, 1);
int numDevices = 0;
int canAccessPeer = 0;
TestType* A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(TestType)};
HIP_CHECK(hipGetDeviceCount(&numDevices));
if (numDevices > 1) {
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
if (canAccessPeer) {
HIP_CHECK(hipSetDevice(0));
// Allocating memory
if (mem_type) {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, true);
} else {
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_h, &B_h, &C_h, NUM_W*NUM_H, false);
}
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<TestType>(NUM_W*NUM_H, A_h, B_h, C_h);
char *X_d{nullptr};
size_t pitch_X;
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&X_d),
&pitch_X, width, NUM_H));
// Change device
HIP_CHECK(hipSetDevice(1));
// Host to Device
HIP_CHECK(hipMemcpy2D(A_d, pitch_A, A_h, COLUMNS*sizeof(TestType),
COLUMNS*sizeof(TestType), ROWS, hipMemcpyHostToDevice));
// Device to Device
HIP_CHECK(hipMemcpy2D(X_d, pitch_X, A_d,
pitch_A, COLUMNS*sizeof(TestType),
ROWS, hipMemcpyDeviceToDevice));
// Device to Host
HIP_CHECK(hipMemcpy2D(B_h, COLUMNS*sizeof(TestType), X_d,
pitch_X, COLUMNS*sizeof(TestType), ROWS, hipMemcpyDeviceToHost));
// Validating the result
REQUIRE(HipTest::checkArray<TestType>(A_h, B_h, COLUMNS, ROWS) == true);
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
if (mem_type) {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, true);
} else {
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr,
A_h, B_h, C_h, false);
}
HIP_CHECK(hipFree(X_d));
} else {
SUCCEED("Machine does not seem to have P2P");
}
} else {
SUCCEED("skipped the testcase as no of devices is less than 2");
}
}
/**
* Test Description
* ------------------------
* - This Testcase verifies the null size checks of hipMemcpy2D API
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEST_CASE("Unit_hipMemcpy2D_SizeCheck") {
CHECK_IMAGE_SUPPORT
HIP_CHECK(hipSetDevice(0));
int* A_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(int)};
// Allocating memory
HipTest::initArrays<int>(nullptr, nullptr, nullptr,
&A_h, nullptr, nullptr, NUM_W*NUM_H);
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<int>(NUM_W*NUM_H, A_h, nullptr, nullptr);
SECTION("hipMemcpy2D API where Source Pitch is zero") {
REQUIRE(hipMemcpy2D(A_h, 0, A_d,
pitch_A, NUM_W, NUM_H,
hipMemcpyDeviceToHost) != hipSuccess);
}
SECTION("hipMemcpy2D API where Destination Pitch is zero") {
REQUIRE(hipMemcpy2D(A_h, width, A_d,
0, NUM_W, NUM_H,
hipMemcpyDeviceToHost) != hipSuccess);
}
SECTION("hipMemcpy2D API where height is zero") {
REQUIRE(hipMemcpy2D(A_h, width, A_d,
pitch_A, NUM_W, 0,
hipMemcpyDeviceToHost) == hipSuccess);
}
SECTION("hipMemcpy2D API where width is zero") {
REQUIRE(hipMemcpy2D(A_h, width, A_d,
pitch_A, 0, NUM_H,
hipMemcpyDeviceToHost) == hipSuccess);
}
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
free(A_h);
}
/**
* Test Description
* ------------------------
* - This Testcase verifies all the negative scenarios of hipMemcpy2D API
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEST_CASE("Unit_hipMemcpy2D_Negative") {
CHECK_IMAGE_SUPPORT
HIP_CHECK(hipSetDevice(0));
int* A_h{nullptr}, *A_d{nullptr};
size_t pitch_A;
size_t width{NUM_W * sizeof(int)};
// Allocating memory
HipTest::initArrays<int>(nullptr, nullptr, nullptr,
&A_h, nullptr, nullptr, NUM_W*NUM_H);
HIP_CHECK(hipMallocPitch(reinterpret_cast<void**>(&A_d),
&pitch_A, width, NUM_H));
// Initialize the data
HipTest::setDefaultData<int>(NUM_W*NUM_H, A_h, nullptr, nullptr);
SECTION("hipMemcpy2D API by Passing nullptr to destination") {
REQUIRE(hipMemcpy2D(nullptr, width, A_d,
pitch_A, COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost) != hipSuccess);
}
SECTION("hipMemcpy2D API by Passing nullptr to destination") {
REQUIRE(hipMemcpy2D(nullptr, width, nullptr,
pitch_A, COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost) != hipSuccess);
}
SECTION("hipMemcpy2D API where width is greater than destination pitch") {
REQUIRE(hipMemcpy2D(A_h, 10, A_d, pitch_A,
COLUMNS*sizeof(int), ROWS,
hipMemcpyDeviceToHost) != hipSuccess);
}
// DeAllocating the memory
HIP_CHECK(hipFree(A_d));
free(A_h);
}
static void hipMemcpy2D_Basic_Size_Test(size_t inc) {
constexpr int defaultProgramSize = 256 * 1024 * 1024;
constexpr int N = 2;
constexpr int value = 42;
int *in, *out, *dev;
size_t newSize = 0, inp = 0;
size_t size = sizeof(int) * N * inc;
size_t free, total;
HIP_CHECK(hipMemGetInfo(&free, &total));
if ( free < 2 * size )
newSize = ( free - defaultProgramSize ) / 2;
else
newSize = size;
INFO("Array size: " << size/1024.0/1024.0 << " MB or " << size << " Bytes.");
INFO("Free memory: " << free/1024.0/1024.0 << " MB or " << free << " Bytes");
INFO("NewSize:" << newSize/1024.0/1024.0 << "MB or " << newSize << " Bytes");
HIP_CHECK(hipHostMalloc(&in, newSize));
HIP_CHECK(hipHostMalloc(&out, newSize));
HIP_CHECK(hipMalloc(&dev, newSize));
inp = newSize / (sizeof(int) * N);
for (size_t i=0; i < N; i++) {
in[i * inp] = value;
}
size_t pitch = sizeof(int) * inp;
HIP_CHECK(hipMemcpy2D(dev, pitch, in, pitch, sizeof(int),
N, hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy2D(out, pitch, dev, pitch, sizeof(int),
N, hipMemcpyDeviceToHost));
for (size_t i=0; i < N; i++) {
REQUIRE(out[i * inp] == value);
}
HIP_CHECK(hipFree(dev));
HIP_CHECK(hipHostFree(in));
HIP_CHECK(hipHostFree(out));
}
/**
* Test Description
* ------------------------
* - This testcase performs multidevice size check on hipMemcpy2D API
1. Verify hipMemcpy2D with 1 << 20 size
2. Verify hipMemcpy2D with 1 << 21 size
* Test source
* ------------------------
* - unit/memory/hipMemcpy2D.cc
* Test requirements
* ------------------------
* - HIP_VERSION >= 6.0
*/
TEST_CASE("Unit_hipMemcpy2D_multiDevice_Basic_Size_Test") {
CHECK_IMAGE_SUPPORT
size_t input = 1 << 20;
int numDevices = 0;
HIP_CHECK(hipGetDeviceCount(&numDevices));
for (int i=0; i < numDevices; i++) {
HIP_CHECK(hipSetDevice(i));
SECTION("Verify hipMemcpy2D with 1 << 20 size") {
hipMemcpy2D_Basic_Size_Test(input);
}
SECTION("Verify hipMemcpy2D with 1 << 21 size") {
input <<= 1;
hipMemcpy2D_Basic_Size_Test(input);
}
}
}
catch/unit/memory/memcpy2d_tests_common.hh
+325
مشاهده پرونده
@@ -0,0 +1,325 @@
/*
Copyright (c) 2022 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.
*/
#pragma once
#include <hip_test_common.hh>
#include <hip/hip_runtime_api.h>
#include <utils.hh>
#include <resource_guards.hh>
template <bool should_synchronize, typename F>
void Memcpy2DDeviceToHostShell(F memcpy_func, const hipStream_t kernel_stream = nullptr) {
const auto kind = GENERATE(hipMemcpyDeviceToHost, hipMemcpyDefault);
constexpr size_t cols = 127;
constexpr size_t rows = 128;
LinearAllocGuard2D<int> device_alloc(cols, rows);
const size_t host_pitch = GENERATE_REF(device_alloc.width(), device_alloc.width() + 64);
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, host_pitch * rows);
const dim3 threads_per_block(32, 32);
const dim3 blocks(cols / threads_per_block.x + 1, rows / threads_per_block.y + 1);
Iota<<<blocks, threads_per_block>>>(device_alloc.ptr(), device_alloc.pitch(),
device_alloc.width_logical(), device_alloc.height(), 1);
HIP_CHECK(hipGetLastError());
HIP_CHECK(memcpy_func(host_alloc.ptr(), host_pitch, device_alloc.ptr(), device_alloc.pitch(),
device_alloc.width(), device_alloc.height(), kind));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(kernel_stream));
}
constexpr auto f = [](size_t x, size_t y, size_t z) { return z * cols * rows + y * cols + x; };
PitchedMemoryVerify(host_alloc.ptr(), host_pitch, device_alloc.width_logical(),
device_alloc.height(), 1, f);
}
template <bool should_synchronize, bool enable_peer_access, typename F>
void Memcpy2DDeviceToDeviceShell(F memcpy_func, const hipStream_t kernel_stream = nullptr) {
const auto kind = GENERATE(hipMemcpyDeviceToDevice, hipMemcpyDefault);
constexpr size_t cols = 127;
constexpr size_t rows = 128;
const auto device_count = HipTest::getDeviceCount();
const auto src_device = GENERATE_COPY(range(0, device_count));
const auto dst_device = GENERATE_COPY(range(0, device_count));
const size_t src_cols_mult = GENERATE(1, 2);
INFO("Src device: " << src_device << ", Dst device: " << dst_device);
HIP_CHECK(hipSetDevice(src_device));
if constexpr (enable_peer_access) {
if (src_device == dst_device) {
return;
}
int can_access_peer = 0;
HIP_CHECK(hipDeviceCanAccessPeer(&can_access_peer, src_device, dst_device));
if (!can_access_peer) {
INFO("Peer access cannot be enabled between devices " << src_device << " " << dst_device);
REQUIRE(can_access_peer);
}
HIP_CHECK(hipDeviceEnablePeerAccess(dst_device, 0));
}
LinearAllocGuard2D<int> src_alloc(cols * src_cols_mult, rows);
HIP_CHECK(hipSetDevice(src_device));
LinearAllocGuard2D<int> dst_alloc(cols, rows);
HIP_CHECK(hipSetDevice(src_device));
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, dst_alloc.width() * rows);
const dim3 threads_per_block(32, 32);
const dim3 blocks(cols / threads_per_block.x + 1, rows / threads_per_block.y + 1);
// Using dst_alloc width and height to set only the elements that will be copied over to
// dst_alloc
Iota<<<blocks, threads_per_block>>>(src_alloc.ptr(), src_alloc.pitch(), dst_alloc.width_logical(),
dst_alloc.height(), 1);
HIP_CHECK(hipGetLastError());
HIP_CHECK(memcpy_func(dst_alloc.ptr(), dst_alloc.pitch(), src_alloc.ptr(), src_alloc.pitch(),
dst_alloc.width(), dst_alloc.height(), kind));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(kernel_stream));
}
HIP_CHECK(hipMemcpy2D(host_alloc.ptr(), dst_alloc.width(), dst_alloc.ptr(), dst_alloc.pitch(),
dst_alloc.width(), dst_alloc.height(), hipMemcpyDeviceToHost));
constexpr auto f = [](size_t x, size_t y, size_t z) { return z * cols * rows + y * cols + x; };
PitchedMemoryVerify(host_alloc.ptr(), dst_alloc.width(), dst_alloc.width_logical(),
dst_alloc.height(), 1, f);
}
template <bool should_synchronize, typename F>
void Memcpy2DHostToDeviceShell(F memcpy_func, const hipStream_t kernel_stream = nullptr) {
const auto kind = GENERATE(hipMemcpyHostToDevice, hipMemcpyDefault);
constexpr size_t cols = 127;
constexpr size_t rows = 128;
LinearAllocGuard2D<int> device_alloc(cols, rows);
const size_t host_pitch = GENERATE_REF(device_alloc.pitch(), 2 * device_alloc.pitch());
LinearAllocGuard<int> src_host_alloc(LinearAllocs::hipHostMalloc, host_pitch * rows);
LinearAllocGuard<int> dst_host_alloc(LinearAllocs::hipHostMalloc, device_alloc.width() * rows);
constexpr auto f = [](size_t x, size_t y, size_t z) { return z * cols * rows + y * cols + x; };
PitchedMemorySet(src_host_alloc.ptr(), host_pitch, device_alloc.width_logical(),
device_alloc.height(), 1, f);
std::fill_n(dst_host_alloc.ptr(), device_alloc.width_logical() * rows, 0);
HIP_CHECK(memcpy_func(device_alloc.ptr(), device_alloc.pitch(), src_host_alloc.ptr(), host_pitch,
device_alloc.width(), device_alloc.height(), kind));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(kernel_stream));
}
HIP_CHECK(hipMemcpy2D(dst_host_alloc.ptr(), device_alloc.width(), device_alloc.ptr(),
device_alloc.pitch(), device_alloc.width(), device_alloc.height(),
hipMemcpyDeviceToHost));
PitchedMemoryVerify(dst_host_alloc.ptr(), device_alloc.width(), device_alloc.width_logical(),
device_alloc.height(), 1, f);
}
template <bool should_synchronize, typename F>
void Memcpy2DHostToHostShell(F memcpy_func, const hipStream_t kernel_stream = nullptr) {
const auto kind = GENERATE(hipMemcpyHostToHost, hipMemcpyDefault);
constexpr size_t cols = 127;
constexpr size_t rows = 128;
const size_t src_pitch = GENERATE_REF(cols * sizeof(int), cols * sizeof(int) + 64);
LinearAllocGuard<int> src_host(LinearAllocs::hipHostMalloc, src_pitch * rows);
LinearAllocGuard<int> dst_host(LinearAllocs::hipHostMalloc, cols * sizeof(int) * rows);
constexpr auto f = [](size_t x, size_t y, size_t z) { return z * cols * rows + y * cols + x; };
PitchedMemorySet(src_host.ptr(), src_pitch, cols, rows, 1, f);
HIP_CHECK(memcpy_func(dst_host.ptr(), cols * sizeof(int), src_host.ptr(), src_pitch,
cols * sizeof(int), rows, kind));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(kernel_stream));
}
PitchedMemoryVerify(dst_host.ptr(), cols * sizeof(int), cols, rows, 1, f);
}
// Synchronization behavior checks
template <typename F>
void MemcpySyncBehaviorCheck(F memcpy_func, const bool should_sync,
const hipStream_t kernel_stream) {
LaunchDelayKernel(std::chrono::milliseconds{300}, kernel_stream);
HIP_CHECK(memcpy_func());
if (should_sync) {
HIP_CHECK(hipStreamQuery(kernel_stream));
} else {
HIP_CHECK_ERROR(hipStreamQuery(kernel_stream), hipErrorNotReady);
}
}
template <typename F>
void Memcpy2DHtoDSyncBehavior(F memcpy_func, const bool should_sync,
const hipStream_t kernel_stream = nullptr) {
using LA = LinearAllocs;
const auto host_alloc_type = GENERATE(LA::malloc, LA::hipHostMalloc);
LinearAllocGuard<int> host_alloc(host_alloc_type, 32 * sizeof(int) * 32);
LinearAllocGuard2D<int> device_alloc(32, 32);
MemcpySyncBehaviorCheck(std::bind(memcpy_func, device_alloc.ptr(), device_alloc.pitch(),
host_alloc.ptr(), device_alloc.width(), device_alloc.width(),
device_alloc.height(), hipMemcpyHostToDevice),
should_sync, kernel_stream);
}
template <typename F>
void Memcpy2DDtoHPageableSyncBehavior(F memcpy_func, const bool should_sync,
const hipStream_t kernel_stream = nullptr) {
LinearAllocGuard<int> host_alloc(LinearAllocs::malloc, 32 * sizeof(int) * 32);
LinearAllocGuard2D<int> device_alloc(32, 32);
MemcpySyncBehaviorCheck(std::bind(memcpy_func, host_alloc.ptr(), device_alloc.width(),
device_alloc.ptr(), device_alloc.pitch(), device_alloc.width(),
device_alloc.height(), hipMemcpyDeviceToHost),
should_sync, kernel_stream);
}
template <typename F>
void Memcpy2DDtoHPinnedSyncBehavior(F memcpy_func, const bool should_sync,
const hipStream_t kernel_stream = nullptr) {
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, 32 * sizeof(int) * 32);
LinearAllocGuard2D<int> device_alloc(32, 32);
MemcpySyncBehaviorCheck(std::bind(memcpy_func, host_alloc.ptr(), device_alloc.width(),
device_alloc.ptr(), device_alloc.pitch(), device_alloc.width(),
device_alloc.height(), hipMemcpyDeviceToHost),
should_sync, kernel_stream);
}
template <typename F>
void Memcpy2DDtoDSyncBehavior(F memcpy_func, const bool should_sync,
const hipStream_t kernel_stream = nullptr) {
LinearAllocGuard2D<int> src_alloc(32, 32);
LinearAllocGuard2D<int> dst_alloc(32, 32);
MemcpySyncBehaviorCheck(
std::bind(memcpy_func, dst_alloc.ptr(), dst_alloc.pitch(), src_alloc.ptr(), src_alloc.pitch(),
dst_alloc.width(), dst_alloc.height(), hipMemcpyDeviceToDevice),
should_sync, kernel_stream);
}
template <typename F>
void Memcpy2DHtoHSyncBehavior(F memcpy_func, const bool should_sync,
const hipStream_t kernel_stream = nullptr) {
using LA = LinearAllocs;
const auto src_alloc_type = GENERATE(LA::malloc, LA::hipHostMalloc);
const auto dst_alloc_type = GENERATE(LA::malloc, LA::hipHostMalloc);
LinearAllocGuard<int> src_alloc(src_alloc_type, 32 * sizeof(int) * 32);
LinearAllocGuard<int> dst_alloc(dst_alloc_type, 32 * sizeof(int) * 32);
MemcpySyncBehaviorCheck(std::bind(memcpy_func, dst_alloc.ptr(), 32 * sizeof(int), src_alloc.ptr(),
32 * sizeof(int), 32 * sizeof(int), 32, hipMemcpyHostToHost),
should_sync, kernel_stream);
}
template <bool should_synchronize, typename F>
void Memcpy2DZeroWidthHeight(F memcpy_func, const hipStream_t stream = nullptr) {
constexpr size_t cols = 63;
constexpr size_t rows = 64;
const auto [width_mult, height_mult] =
GENERATE(std::make_pair(0, 1), std::make_pair(1, 0), std::make_pair(0, 0));
SECTION("Device to Host") {
LinearAllocGuard2D<uint8_t> device_alloc(cols, rows);
LinearAllocGuard<uint8_t> host_alloc(LinearAllocs::hipHostMalloc, device_alloc.width() * rows);
std::fill_n(host_alloc.ptr(), device_alloc.width_logical() * device_alloc.height(), 42);
HIP_CHECK(hipMemset2D(device_alloc.ptr(), device_alloc.pitch(), 1, device_alloc.width(),
device_alloc.height()));
HIP_CHECK(memcpy_func(host_alloc.ptr(), device_alloc.width(), device_alloc.ptr(),
device_alloc.pitch(), device_alloc.width() * width_mult,
device_alloc.height() * height_mult, hipMemcpyDeviceToHost));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(stream));
}
ArrayFindIfNot(host_alloc.ptr(), static_cast<uint8_t>(42),
device_alloc.width_logical() * device_alloc.height());
}
SECTION("Device to Device") {
LinearAllocGuard2D<uint8_t> src_alloc(cols, rows);
LinearAllocGuard2D<uint8_t> dst_alloc(cols, rows);
LinearAllocGuard<uint8_t> host_alloc(LinearAllocs::hipHostMalloc, dst_alloc.width() * rows);
HIP_CHECK(
hipMemset2D(src_alloc.ptr(), src_alloc.pitch(), 1, src_alloc.width(), src_alloc.height()));
HIP_CHECK(
hipMemset2D(dst_alloc.ptr(), dst_alloc.pitch(), 42, dst_alloc.width(), dst_alloc.height()));
HIP_CHECK(memcpy_func(dst_alloc.ptr(), dst_alloc.pitch(), src_alloc.ptr(), src_alloc.pitch(),
dst_alloc.width() * width_mult, dst_alloc.height() * height_mult,
hipMemcpyDeviceToDevice));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(stream));
}
HIP_CHECK(hipMemcpy2D(host_alloc.ptr(), dst_alloc.width(), dst_alloc.ptr(), dst_alloc.pitch(),
dst_alloc.width(), dst_alloc.height(), hipMemcpyDeviceToHost));
ArrayFindIfNot(host_alloc.ptr(), static_cast<uint8_t>(42),
dst_alloc.width_logical() * dst_alloc.height());
}
SECTION("Host to Device") {
LinearAllocGuard2D<uint8_t> device_alloc(cols, rows);
LinearAllocGuard<uint8_t> src_host_alloc(LinearAllocs::hipHostMalloc,
device_alloc.width() * rows);
LinearAllocGuard<uint8_t> dst_host_alloc(LinearAllocs::hipHostMalloc,
device_alloc.width() * rows);
std::fill_n(src_host_alloc.ptr(), device_alloc.width_logical() * device_alloc.height(), 1);
HIP_CHECK(hipMemset2D(device_alloc.ptr(), device_alloc.pitch(), 42, device_alloc.width(),
device_alloc.height()));
HIP_CHECK(memcpy_func(device_alloc.ptr(), device_alloc.pitch(), src_host_alloc.ptr(),
device_alloc.width(), device_alloc.width() * width_mult,
device_alloc.height() * height_mult, hipMemcpyHostToDevice));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(stream));
}
HIP_CHECK(hipMemcpy2D(dst_host_alloc.ptr(), device_alloc.width(), device_alloc.ptr(),
device_alloc.pitch(), device_alloc.width(), device_alloc.height(),
hipMemcpyDeviceToHost));
ArrayFindIfNot(dst_host_alloc.ptr(), static_cast<uint8_t>(42),
device_alloc.width_logical() * device_alloc.height());
}
SECTION("Host to Host") {
const auto alloc_size = cols * rows;
LinearAllocGuard<uint8_t> src_alloc(LinearAllocs::hipHostMalloc, alloc_size);
LinearAllocGuard<uint8_t> dst_alloc(LinearAllocs::hipHostMalloc, alloc_size);
std::fill_n(src_alloc.ptr(), alloc_size, 1);
std::fill_n(dst_alloc.ptr(), alloc_size, 42);
HIP_CHECK(memcpy_func(dst_alloc.ptr(), cols, src_alloc.ptr(), cols, cols * width_mult,
rows * height_mult, hipMemcpyHostToHost));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(stream));
}
ArrayFindIfNot(dst_alloc.ptr(), static_cast<uint8_t>(42), alloc_size);
}
}