AI/rocm-systems
2
0
Bifurcation 0
Code Tickets Demandes d'ajout Actions Paquets Projets Publications Wiki Activité

EXSWHTEC-74 - Implement tests for hipMemcpy and derivatives (#15)

Parcourir la source 
- Basic positive tests
- Negative parameter tests
Cette révision appartient à :
music-dino
2023-01-17 07:55:43 +01:00
révisé par GitHub
Parent fb919e7995
révision f5eb0da5c1
7 fichiers modifiés avec 1784 ajouts et 1141 suppressions
Télécharger le Fichier Patch Télécharger le Fichier des Différences Développer tous les fichiers Réduire tous les fichiers
catch/include/memcpy1d_tests_common.hh
+316
Voir le fichier
@@ -0,0 +1,316 @@
/*
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 <functional>
#include <hip_test_common.hh>
#include <hip/hip_runtime_api.h>
#include <utils.hh>
#include <resource_guards.hh>
static inline unsigned int GenerateLinearAllocationFlagCombinations(
const LinearAllocs allocation_type) {
switch (allocation_type) {
case LinearAllocs::hipHostMalloc:
return GENERATE(hipHostMallocDefault, hipHostMallocPortable, hipHostMallocMapped,
hipHostMallocWriteCombined);
case LinearAllocs::mallocAndRegister:
case LinearAllocs::hipMallocManaged:
case LinearAllocs::malloc:
case LinearAllocs::hipMalloc:
return 0u;
default:
assert("Invalid LinearAllocs enumerator");
throw std::invalid_argument("Invalid LinearAllocs enumerator");
}
}
template <bool should_synchronize, typename F>
void MemcpyDeviceToHostShell(F memcpy_func, const hipStream_t kernel_stream = nullptr) {
using LA = LinearAllocs;
const auto allocation_size = GENERATE(kPageSize / 2, kPageSize, kPageSize * 2);
const auto host_allocation_type = GENERATE(LA::malloc, LA::hipHostMalloc);
const auto host_allocation_flags = GenerateLinearAllocationFlagCombinations(host_allocation_type);
LinearAllocGuard<int> host_allocation(host_allocation_type, allocation_size,
host_allocation_flags);
LinearAllocGuard<int> device_allocation(LA::hipMalloc, allocation_size);
const auto element_count = allocation_size / sizeof(*device_allocation.ptr());
constexpr auto thread_count = 1024;
const auto block_count = element_count / thread_count + 1;
constexpr int expected_value = 42;
VectorSet<<<block_count, thread_count, 0, kernel_stream>>>(device_allocation.ptr(),
expected_value, element_count);
HIP_CHECK(hipGetLastError());
HIP_CHECK(memcpy_func(host_allocation.host_ptr(), device_allocation.ptr(), allocation_size));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(kernel_stream));
}
ArrayFindIfNot(host_allocation.host_ptr(), expected_value, element_count);
}
template <bool should_synchronize, typename F>
void MemcpyHostToDeviceShell(F memcpy_func, const hipStream_t kernel_stream = nullptr) {
using LA = LinearAllocs;
const auto allocation_size = GENERATE(kPageSize / 2, kPageSize, kPageSize * 2);
const auto host_allocation_type = GENERATE(LA::malloc, LA::hipHostMalloc);
const auto host_allocation_flags = GenerateLinearAllocationFlagCombinations(host_allocation_type);
LinearAllocGuard<int> src_host_allocation(host_allocation_type, allocation_size,
host_allocation_flags);
LinearAllocGuard<int> dst_host_allocation(LA::hipHostMalloc, allocation_size);
LinearAllocGuard<int> device_allocation(LA::hipMalloc, allocation_size);
const auto element_count = allocation_size / sizeof(*device_allocation.ptr());
constexpr int fill_value = 42;
std::fill_n(src_host_allocation.host_ptr(), element_count, fill_value);
std::fill_n(dst_host_allocation.host_ptr(), element_count, 0);
HIP_CHECK(memcpy_func(device_allocation.ptr(), src_host_allocation.host_ptr(), allocation_size));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(kernel_stream));
}
HIP_CHECK(hipMemcpy(dst_host_allocation.host_ptr(), device_allocation.ptr(), allocation_size,
hipMemcpyDeviceToHost));
ArrayFindIfNot(dst_host_allocation.host_ptr(), fill_value, element_count);
}
template <bool should_synchronize, typename F>
void MemcpyHostToHostShell(F memcpy_func, const hipStream_t kernel_stream = nullptr) {
using LA = LinearAllocs;
const auto allocation_size = GENERATE(kPageSize / 2, kPageSize, kPageSize * 2);
const auto src_allocation_type = GENERATE(LA::malloc, LA::hipHostMalloc);
const auto dst_allocation_type = GENERATE(LA::malloc, LA::hipHostMalloc);
const auto src_allocation_flags = GenerateLinearAllocationFlagCombinations(src_allocation_type);
const auto dst_allocation_flags = GenerateLinearAllocationFlagCombinations(dst_allocation_type);
LinearAllocGuard<int> src_allocation(src_allocation_type, allocation_size, src_allocation_flags);
LinearAllocGuard<int> dst_allocation(dst_allocation_type, allocation_size, dst_allocation_flags);
const auto element_count = allocation_size / sizeof(*src_allocation.host_ptr());
constexpr auto expected_value = 42;
std::fill_n(src_allocation.host_ptr(), element_count, expected_value);
HIP_CHECK(memcpy_func(dst_allocation.host_ptr(), src_allocation.host_ptr(), allocation_size));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(kernel_stream));
}
ArrayFindIfNot(dst_allocation.host_ptr(), expected_value, element_count);
}
template <bool should_synchronize, bool enable_peer_access, typename F>
void MemcpyDeviceToDeviceShell(F memcpy_func, const hipStream_t kernel_stream = nullptr) {
const auto allocation_size = GENERATE(kPageSize / 2, kPageSize, kPageSize * 2);
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));
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));
}
LinearAllocGuard<int> src_allocation(LinearAllocs::hipMalloc, allocation_size);
LinearAllocGuard<int> result(LinearAllocs::hipHostMalloc, allocation_size, hipHostMallocPortable);
HIP_CHECK(hipSetDevice(dst_device));
LinearAllocGuard<int> dst_allocation(LinearAllocs::hipMalloc, allocation_size);
const auto element_count = allocation_size / sizeof(*src_allocation.ptr());
constexpr auto thread_count = 1024;
const auto block_count = element_count / thread_count + 1;
constexpr int expected_value = 42;
HIP_CHECK(hipSetDevice(src_device));
VectorSet<<<block_count, thread_count, 0, kernel_stream>>>(src_allocation.ptr(), expected_value,
element_count);
HIP_CHECK(hipGetLastError());
HIP_CHECK(memcpy_func(dst_allocation.ptr(), src_allocation.ptr(), allocation_size));
if constexpr (should_synchronize) {
HIP_CHECK(hipStreamSynchronize(kernel_stream));
}
HIP_CHECK(
hipMemcpy(result.host_ptr(), dst_allocation.ptr(), allocation_size, hipMemcpyDeviceToHost));
if constexpr (enable_peer_access) {
// If we've gotten this far, EnablePeerAccess must have succeeded, so we only need to check this
// condition
HIP_CHECK(hipDeviceDisablePeerAccess(dst_device));
}
ArrayFindIfNot(result.host_ptr(), expected_value, element_count);
}
template <bool should_synchronize, typename F> void MemcpyWithDirectionCommonTests(F memcpy_func) {
using namespace std::placeholders;
SECTION("Device to host") {
MemcpyDeviceToHostShell<should_synchronize>(
std::bind(memcpy_func, _1, _2, _3, hipMemcpyDeviceToHost));
}
SECTION("Device to host with default kind") {
MemcpyDeviceToHostShell<should_synchronize>(
std::bind(memcpy_func, _1, _2, _3, hipMemcpyDefault));
}
SECTION("Host to device") {
MemcpyHostToDeviceShell<should_synchronize>(
std::bind(memcpy_func, _1, _2, _3, hipMemcpyHostToDevice));
}
SECTION("Host to device with default kind") {
MemcpyHostToDeviceShell<should_synchronize>(
std::bind(memcpy_func, _1, _2, _3, hipMemcpyDefault));
}
SECTION("Host to host") {
MemcpyHostToHostShell<should_synchronize>(
std::bind(memcpy_func, _1, _2, _3, hipMemcpyHostToHost));
}
SECTION("Host to host with default kind") {
MemcpyHostToHostShell<should_synchronize>(std::bind(memcpy_func, _1, _2, _3, hipMemcpyDefault));
}
SECTION("Device to device") {
SECTION("Peer access enabled") {
MemcpyDeviceToDeviceShell<should_synchronize, true>(
std::bind(memcpy_func, _1, _2, _3, hipMemcpyDeviceToDevice));
}
SECTION("Peer access disabled") {
MemcpyDeviceToDeviceShell<should_synchronize, false>(
std::bind(memcpy_func, _1, _2, _3, hipMemcpyDeviceToDevice));
}
}
SECTION("Device to device with default kind") {
SECTION("Peer access enabled") {
MemcpyDeviceToDeviceShell<should_synchronize, true>(
std::bind(memcpy_func, _1, _2, _3, hipMemcpyDefault));
}
SECTION("Peer access disabled") {
MemcpyDeviceToDeviceShell<should_synchronize, false>(
std::bind(memcpy_func, _1, _2, _3, hipMemcpyDefault));
}
}
}
// Synchronization behavior checks
template <typename F>
void MemcpySyncBehaviorCheck(F memcpy_func, const bool should_sync,
const hipStream_t kernel_stream) {
LaunchDelayKernel(std::chrono::milliseconds{100}, 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 MemcpyHtoDSyncBehavior(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, kPageSize);
LinearAllocGuard<int> device_alloc(LA::hipMalloc, kPageSize);
MemcpySyncBehaviorCheck(std::bind(memcpy_func, device_alloc.ptr(), host_alloc.ptr(), kPageSize),
should_sync, kernel_stream);
}
template <typename F>
void MemcpyDtoHPageableSyncBehavior(F memcpy_func, const bool should_sync,
const hipStream_t kernel_stream = nullptr) {
LinearAllocGuard<int> host_alloc(LinearAllocs::malloc, kPageSize);
LinearAllocGuard<int> device_alloc(LinearAllocs::hipMalloc, kPageSize);
MemcpySyncBehaviorCheck(std::bind(memcpy_func, host_alloc.ptr(), device_alloc.ptr(), kPageSize),
should_sync, kernel_stream);
}
template <typename F>
void MemcpyDtoHPinnedSyncBehavior(F memcpy_func, const bool should_sync,
const hipStream_t kernel_stream = nullptr) {
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, kPageSize);
LinearAllocGuard<int> device_alloc(LinearAllocs::hipMalloc, kPageSize);
MemcpySyncBehaviorCheck(std::bind(memcpy_func, host_alloc.ptr(), device_alloc.ptr(), kPageSize),
should_sync, kernel_stream);
}
template <typename F>
void MemcpyDtoDSyncBehavior(F memcpy_func, const bool should_sync,
const hipStream_t kernel_stream = nullptr) {
LinearAllocGuard<int> src_alloc(LinearAllocs::hipMalloc, kPageSize);
LinearAllocGuard<int> dst_alloc(LinearAllocs::hipMalloc, kPageSize);
MemcpySyncBehaviorCheck(std::bind(memcpy_func, dst_alloc.ptr(), src_alloc.ptr(), kPageSize),
should_sync, kernel_stream);
}
template <typename F>
void MemcpyHtoHSyncBehavior(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, kPageSize);
LinearAllocGuard<int> dst_alloc(dst_alloc_type, kPageSize);
MemcpySyncBehaviorCheck(std::bind(memcpy_func, dst_alloc.ptr(), src_alloc.ptr(), kPageSize),
should_sync, kernel_stream);
}
// Common negative tests
template <typename F> void MemcpyCommonNegativeTests(F f, void* dst, void* src, size_t count) {
SECTION("dst == nullptr") { HIP_CHECK_ERROR(f(nullptr, src, count), hipErrorInvalidValue); }
SECTION("src == nullptr") { HIP_CHECK_ERROR(f(dst, nullptr, count), hipErrorInvalidValue); }
}
template <typename F>
void MemcpyWithDirectionCommonNegativeTests(F f, void* dst, void* src, size_t count,
hipMemcpyKind kind) {
using namespace std::placeholders;
MemcpyCommonNegativeTests(std::bind(f, _1, _2, _3, kind), dst, src, count);
// Disabled on AMD due to defect - EXSWHTEC-128
#if HT_NVIDIA
SECTION("Invalid MemcpyKind") {
HIP_CHECK_ERROR(f(dst, src, count, static_cast<hipMemcpyKind>(-1)),
hipErrorInvalidMemcpyDirection);
}
#endif
}
catch/unit/memory/CMakeLists.txt
+6
Voir le fichier
@@ -65,6 +65,7 @@ set(TEST_SRC
hipMemPoolApi.cc
hipMemcpyPeer.cc
hipMemcpyPeerAsync.cc
hipMemcpyWithStream_old.cc
hipMemcpyWithStream.cc
hipMemcpyWithStreamMultiThread.cc
hipMemsetAsyncAndKernel.cc
@@ -74,7 +75,9 @@ set(TEST_SRC
hipMemcpyDtoD.cc
hipMemcpyDtoDAsync.cc
hipHostMalloc.cc
hipMemcpy_old.cc
hipMemcpy.cc
hipMemcpy_derivatives.cc
hipMemcpyAsync.cc
hipMemsetFunctional.cc
hipMallocPitch.cc
@@ -142,6 +145,7 @@ set(TEST_SRC
hipMemPoolApi.cc
hipMemcpyPeer.cc
hipMemcpyPeerAsync.cc
hipMemcpyWithStream_old.cc
hipMemcpyWithStream.cc
hipMemcpyWithStreamMultiThread.cc
hipMemsetAsyncAndKernel.cc
@@ -151,7 +155,9 @@ set(TEST_SRC
hipMemcpyDtoD.cc
hipMemcpyDtoDAsync.cc
hipHostMalloc.cc
hipMemcpy_old.cc
hipMemcpy.cc
hipMemcpy_derivatives.cc
hipMemcpyAsync.cc
hipMemsetFunctional.cc
hipMallocPitch.cc
catch/unit/memory/hipMemcpy.cc
+66 -584
Voir le fichier
@@ -1,13 +1,15 @@
/*
Copyright (c) 2021 - present 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 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
@@ -17,603 +19,83 @@ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.
*/
/*
This testcase verifies following scenarios
1. hipMemcpy API along with kernel launch with different data types
2. H2D-D2D-D2H scenarios for unpinned and pinned memory
3. Boundary checks with different sizes
4. Multithread scenario
5. device offset scenario
*/
#include <hip_test_common.hh>
#include <hip_test_kernels.hh>
#include <hip_test_checkers.hh>
#include <hip/hip_runtime_api.h>
#include <memcpy1d_tests_common.hh>
#include <resource_guards.hh>
#include <utils.hh>
#ifdef _WIN32
#define WIN32_LEAN_AND_MEAN
#include <windows.h>
#else
#include "sys/types.h"
#include "sys/sysinfo.h"
#endif
static constexpr auto NUM_ELM{4*1024 * 1024};
static unsigned blocksPerCU{6}; // to hide latency
static unsigned threadsPerBlock{256};
template<typename T>
class DeviceMemory {
public:
explicit DeviceMemory(size_t numElements);
DeviceMemory() = delete;
~DeviceMemory();
T* A_d() const { return _A_d + _offset; }
T* B_d() const { return _B_d + _offset; }
T* C_d() const { return _C_d + _offset; }
T* C_dd() const { return _C_dd + _offset; }
size_t maxNumElements() const { return _maxNumElements; }
void offset(int offset) { _offset = offset; }
int offset() const { return _offset; }
private:
T* _A_d;
T* _B_d;
T* _C_d;
T* _C_dd;
size_t _maxNumElements;
int _offset;
};
template <typename T>
DeviceMemory<T>::DeviceMemory(size_t numElements) :
_maxNumElements(numElements), _offset(0) {
T** np = nullptr;
HipTest::initArrays(&_A_d, &_B_d, &_C_d, np, np, np, numElements, 0);
size_t sizeElements = numElements * sizeof(T);
HIP_CHECK(hipMalloc(&_C_dd, sizeElements));
}
template <typename T>
DeviceMemory<T>::~DeviceMemory() {
T* np = nullptr;
HipTest::freeArrays<T>(_A_d, _B_d, _C_d, np, np, np, 0);
HIP_CHECK(hipFree(_C_dd));
_C_dd = NULL;
}
template <typename T>
class HostMemory {
public:
HostMemory(size_t numElements, bool usePinnedHost);
HostMemory() = delete;
void reset(size_t numElements, bool full = false);
~HostMemory();
T* A_h() const { return _A_h + _offset; }
T* B_h() const { return _B_h + _offset; }
T* C_h() const { return _C_h + _offset; }
size_t maxNumElements() const { return _maxNumElements; }
void offset(int offset) { _offset = offset; }
int offset() const { return _offset; }
// Host arrays, secondary copy
T* A_hh;
T* B_hh;
bool _usePinnedHost;
private:
size_t _maxNumElements;
int _offset;
// Host arrays
T* _A_h;
T* _B_h;
T* _C_h;
};
template <typename T>
HostMemory<T>::HostMemory(size_t numElements, bool usePinnedHost)
: _usePinnedHost(usePinnedHost), _maxNumElements(numElements), _offset(0) {
T** np = nullptr;
HipTest::initArrays(np, np, np, &_A_h, &_B_h, &_C_h,
numElements, usePinnedHost);
A_hh = NULL;
B_hh = NULL;
size_t sizeElements = numElements * sizeof(T);
if (usePinnedHost) {
HIP_CHECK(hipHostMalloc(reinterpret_cast<void**>(&A_hh), sizeElements,
hipHostMallocDefault));
HIP_CHECK(hipHostMalloc(reinterpret_cast<void**>(&B_hh), sizeElements,
hipHostMallocDefault));
} else {
A_hh = reinterpret_cast<T*>(malloc(sizeElements));
B_hh = reinterpret_cast<T*>(malloc(sizeElements));
}
}
template <typename T>
void HostMemory<T>::reset(size_t numElements, bool full) {
// Initialize the host data:
for (size_t i = 0; i < numElements; i++) {
(A_hh)[i] = 1097.0 + i;
(B_hh)[i] = 1492.0 + i; // Phi
if (full) {
(_A_h)[i] = 3.146f + i; // Pi
(_B_h)[i] = 1.618f + i; // Phi
}
}
}
template <typename T>
HostMemory<T>::~HostMemory() {
HipTest::freeArraysForHost(_A_h, _B_h, _C_h, _usePinnedHost);
if (_usePinnedHost) {
HIP_CHECK(hipHostFree(A_hh));
HIP_CHECK(hipHostFree(B_hh));
} else {
free(A_hh);
free(B_hh);
}
}
#ifdef _WIN32
void memcpytest2_get_host_memory(size_t *free, size_t *total) {
MEMORYSTATUSEX status;
status.dwLength = sizeof(status);
GlobalMemoryStatusEx(&status);
// Windows doesn't allow allocating more than half of system memory to the gpu
// Since the runtime also needs space for its internal allocations,
// we should not try to allocate more than 40% of reported system memory,
// otherwise we can run into OOM issues.
*free = static_cast<size_t>(0.4 * status.ullAvailPhys);
*total = static_cast<size_t>(0.4 * status.ullTotalPhys);
}
#else
struct sysinfo memInfo;
void memcpytest2_get_host_memory(size_t *free, size_t *total) {
sysinfo(&memInfo);
uint64_t freePhysMem = memInfo.freeram;
freePhysMem *= memInfo.mem_unit;
*free = freePhysMem;
uint64_t totalPhysMem = memInfo.totalram;
totalPhysMem *= memInfo.mem_unit;
*total = totalPhysMem;
}
#endif
//---
// Test many different kinds of memory copies.
// The subroutine allocates memory , copies to device, runs a vector
// add kernel, copies back, and
// checks the result.
//
// IN: numElements controls the number of elements used for allocations.
// IN: usePinnedHost : If true, allocate host with hipHostMalloc and is pinned
// else allocate host
// memory with malloc. IN: useHostToHost : If true, add an extra
// host-to-host copy. IN:
// useDeviceToDevice : If true, add an extra deviceto-device copy after
// result is produced. IN:
// useMemkindDefault : If true, use memkinddefault
// (runtime figures out direction). if false, use
// explicit memcpy direction.
//
template <typename T>
void memcpytest2(DeviceMemory<T>* dmem, HostMemory<T>* hmem,
size_t numElements, bool useHostToHost,
bool useDeviceToDevice, bool useMemkindDefault) {
size_t sizeElements = numElements * sizeof(T);
hmem->reset(numElements);
assert(numElements <= dmem->maxNumElements());
assert(numElements <= hmem->maxNumElements());
if (useHostToHost) {
// Do some extra host-to-host copies here to mix things up:
HIP_CHECK(hipMemcpy(hmem->A_hh, hmem->A_h(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToHost));
HIP_CHECK(hipMemcpy(hmem->B_hh, hmem->B_h(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToHost));
HIP_CHECK(hipMemcpy(dmem->A_d(), hmem->A_hh, sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy(dmem->B_d(), hmem->B_hh, sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice));
} else {
HIP_CHECK(hipMemcpy(dmem->A_d(), hmem->A_h(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy(dmem->B_d(), hmem->B_h(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice));
}
hipLaunchKernelGGL(HipTest::vectorADD, dim3(1), dim3(1), 0, 0,
static_cast<const T*>(dmem->A_d()), static_cast<const T*>(dmem->B_d()),
dmem->C_d(), numElements);
HIP_CHECK(hipGetLastError());
if (useDeviceToDevice) {
// Do an extra device-to-device copy here to mix things up:
HIP_CHECK(hipMemcpy(dmem->C_dd(), dmem->C_d(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyDeviceToDevice));
// Destroy the original dmem->C_d():
HIP_CHECK(hipMemset(dmem->C_d(), 0x5A, sizeElements));
HIP_CHECK(hipMemcpy(hmem->C_h(), dmem->C_dd(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyDeviceToHost));
} else {
HIP_CHECK(hipMemcpy(hmem->C_h(), dmem->C_d(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyDeviceToHost));
}
TEST_CASE("Unit_hipMemcpy_Positive_Basic") { MemcpyWithDirectionCommonTests<false>(hipMemcpy); }
TEST_CASE("Unit_hipMemcpy_Positive_Synchronization_Behavior") {
using namespace std::placeholders;
HIP_CHECK(hipDeviceSynchronize());
HipTest::checkVectorADD(hmem->A_h(), hmem->B_h(), hmem->C_h(), numElements);
// For transfers from pageable host memory to device memory, a stream sync is performed before
// the copy is initiated. The function will return once the pageable buffer has been copied to
// the staging memory for DMA transfer to device memory, but the DMA to final destination may
// not have completed.
// For transfers from pinned host memory to device memory, the function is synchronous with
// respect to the host
SECTION("Host memory to device memory") {
MemcpyHtoDSyncBehavior(std::bind(hipMemcpy, _1, _2, _3, hipMemcpyHostToDevice), true);
}
printf(" %s success\n", __func__);
}
// For transfers from device to either pageable or pinned host memory, the function returns only
// once the copy has completed
SECTION("Device memory to host memory") {
const auto f = std::bind(hipMemcpy, _1, _2, _3, hipMemcpyDeviceToHost);
MemcpyDtoHPageableSyncBehavior(f, true);
MemcpyDtoHPinnedSyncBehavior(f, true);
}
// Try all the 16 possible combinations to memcpytest2 - usePinnedHost,
// useHostToHost,
// useDeviceToDevice, useMemkindDefault
template <typename T>
void memcpytest2_for_type(size_t numElements) {
DeviceMemory<T> memD(numElements);
HostMemory<T> memU(numElements, 0 /*usePinnedHost*/);
HostMemory<T> memP(numElements, 1 /*usePinnedHost*/);
// For transfers from device memory to device memory, no host-side synchronization is performed.
SECTION("Device memory to device memory") {
// This behavior differs on NVIDIA and AMD, on AMD the hipMemcpy calls is synchronous with
// respect to the host
#if HT_AMD
HipTest::HIP_SKIP_TEST(
"EXSWCPHIPT-127 - Memcpy from device to device memory behavior differs on AMD and Nvidia");
return;
#endif
MemcpyDtoDSyncBehavior(std::bind(hipMemcpy, _1, _2, _3, hipMemcpyDeviceToDevice), false);
}
for (int usePinnedHost = 0; usePinnedHost <= 1; usePinnedHost++) {
for (int useHostToHost = 0; useHostToHost <= 1; useHostToHost++) {
for (int useDeviceToDevice = 0; useDeviceToDevice <= 1;
useDeviceToDevice++) {
for (int useMemkindDefault = 0; useMemkindDefault <= 1;
useMemkindDefault++) {
memcpytest2<T>(&memD, usePinnedHost ? &memP : &memU,
numElements, useHostToHost,
useDeviceToDevice, useMemkindDefault);
}
}
}
// For transfers from any host memory to any host memory, the function is fully synchronous with
// respect to the host
SECTION("Host memory to host memory") {
MemcpyHtoHSyncBehavior(std::bind(hipMemcpy, _1, _2, _3, hipMemcpyHostToHost), true);
}
}
// Try many different sizes to memory copy.
template <typename T>
void memcpytest2_sizes(size_t maxElem = 0) {
int deviceId;
HIP_CHECK(hipGetDevice(&deviceId));
TEST_CASE("Unit_hipMemcpy_Negative_Parameters") {
using namespace std::placeholders;
size_t free, total, freeCPU, totalCPU;
HIP_CHECK(hipMemGetInfo(&free, &total));
memcpytest2_get_host_memory(&freeCPU, &totalCPU);
if (maxElem == 0) {
// Use lesser maxElem if not enough host memory available
size_t maxElemGPU = free / sizeof(T) / 8;
size_t maxElemCPU = freeCPU / sizeof(T) / 8;
maxElem = maxElemGPU < maxElemCPU ? maxElemGPU : maxElemCPU;
SECTION("Host to device") {
LinearAllocGuard<int> device_alloc(LinearAllocs::hipMalloc, kPageSize);
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, kPageSize);
MemcpyWithDirectionCommonNegativeTests(hipMemcpy, device_alloc.ptr(), host_alloc.ptr(),
kPageSize, hipMemcpyHostToDevice);
}
HIP_CHECK(hipDeviceReset());
DeviceMemory<T> memD(maxElem);
HostMemory<T> memU(maxElem, 0 /*usePinnedHost*/);
HostMemory<T> memP(maxElem, 1 /*usePinnedHost*/);
SECTION("Device to host") {
LinearAllocGuard<int> device_alloc(LinearAllocs::hipMalloc, kPageSize);
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, kPageSize);
MemcpyWithDirectionCommonNegativeTests(hipMemcpy, host_alloc.ptr(), device_alloc.ptr(),
kPageSize, hipMemcpyDeviceToHost);
}
for (size_t elem = 1; elem <= maxElem; elem *= 2) {
memcpytest2<T>(&memD, &memU, elem, 1, 1, 0); // unpinned host
memcpytest2<T>(&memD, &memP, elem, 1, 1, 0); // pinned host
}
}
// Try many different sizes to memory copy.
template <typename T>
void memcpytest2_offsets(size_t maxElem, bool devOffsets, bool hostOffsets) {
int deviceId;
HIP_CHECK(hipGetDevice(&deviceId));
size_t free, total;
HIP_CHECK(hipMemGetInfo(&free, &total));
HIP_CHECK(hipDeviceReset());
DeviceMemory<T> memD(maxElem);
HostMemory<T> memU(maxElem, 0 /*usePinnedHost*/);
HostMemory<T> memP(maxElem, 1 /*usePinnedHost*/);
size_t elem = maxElem / 2;
for (size_t offset = 0; offset < 512; offset++) {
assert(elem + offset < maxElem);
if (devOffsets) {
memD.offset(offset);
}
if (hostOffsets) {
memU.offset(offset);
memP.offset(offset);
}
memcpytest2<T>(&memD, &memU, elem, 1, 1, 0); // unpinned host
memcpytest2<T>(&memD, &memP, elem, 1, 1, 0); // pinned host
}
for (size_t offset = 512; offset < elem; offset *= 2) {
assert(elem + offset < maxElem);
if (devOffsets) {
memD.offset(offset);
}
if (hostOffsets) {
memU.offset(offset);
memP.offset(offset);
}
memcpytest2<T>(&memD, &memU, elem, 1, 1, 0); // unpinned host
memcpytest2<T>(&memD, &memP, elem, 1, 1, 0); // pinned host
}
}
// Create multiple threads to stress multi-thread locking behavior in the
// allocation/deallocation/tracking logic:
template <typename T>
void multiThread_1(bool serialize, bool usePinnedHost) {
DeviceMemory<T> memD(NUM_ELM);
HostMemory<T> mem1(NUM_ELM, usePinnedHost);
HostMemory<T> mem2(NUM_ELM, usePinnedHost);
std::thread t1(memcpytest2<T>, &memD, &mem1, NUM_ELM, 0, 0, 0);
if (serialize) {
t1.join();
}
std::thread t2(memcpytest2<T>, &memD, &mem2, NUM_ELM, 0, 0, 0);
if (serialize) {
t2.join();
}
}
/*
This testcase verifies hipMemcpy API
Initializes device variables
Launches kernel and performs the sum of device variables
copies the result to host variable and validates the result.
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy_KernelLaunch", "", int, float,
double) {
size_t Nbytes = NUM_ELM * sizeof(TestType);
TestType *A_d{nullptr}, *B_d{nullptr}, *C_d{nullptr};
TestType *A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr};
HipTest::initArrays(&A_d, &B_d, &C_d, &A_h, &B_h, &C_h, NUM_ELM, false);
HIP_CHECK(hipMemcpy(A_d, A_h, Nbytes, hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy(B_d, B_h, Nbytes, hipMemcpyHostToDevice));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(1), dim3(1), 0, 0,
static_cast<const TestType*>(A_d),
static_cast<const TestType*>(B_d), C_d, NUM_ELM);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipMemcpy(C_h, C_d, Nbytes, hipMemcpyDeviceToHost));
HIP_CHECK(hipDeviceSynchronize());
HipTest::checkVectorADD(A_h, B_h, C_h, NUM_ELM);
HipTest::freeArrays<TestType>(A_d, B_d, C_d, A_h, B_h, C_h, false);
}
/*
This testcase verifies the following scenarios
1. H2H,H2PinMem and PinnedMem2Host
2. H2D-D2D-D2H in same GPU
3. Pinned Host Memory to device variables in same GPU
4. Device context change
5. H2D-D2D-D2H peer GPU
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy_H2H-H2D-D2H-H2PinMem", "", int,
float, double) {
TestType *A_d{nullptr}, *B_d{nullptr};
TestType *A_h{nullptr}, *B_h{nullptr};
TestType *A_Ph{nullptr}, *B_Ph{nullptr};
HIP_CHECK(hipSetDevice(0));
HipTest::initArrays<TestType>(&A_d, &B_d, nullptr,
&A_h, &B_h, nullptr,
NUM_ELM*sizeof(TestType));
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_Ph, &B_Ph, nullptr,
NUM_ELM*sizeof(TestType), true);
SECTION("H2H, H2PinMem and PinMem2H") {
HIP_CHECK(hipMemcpy(B_h, A_h, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(A_Ph, B_h, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_Ph, A_Ph, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HipTest::checkTest(A_h, B_Ph, NUM_ELM);
}
SECTION("H2D-D2D-D2H-SameGPU") {
HIP_CHECK(hipMemcpy(A_d, A_h, NUM_ELM*sizeof(TestType), hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_d, A_d, NUM_ELM*sizeof(TestType), hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_h, B_d, NUM_ELM*sizeof(TestType), hipMemcpyDefault));
HipTest::checkTest(A_h, B_h, NUM_ELM);
}
SECTION("pH2D-D2D-D2pH-SameGPU") {
HIP_CHECK(hipMemcpy(A_d, A_Ph, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_d, A_d, NUM_ELM*sizeof(TestType), hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_Ph, B_d, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HipTest::checkTest(A_Ph, B_Ph, NUM_ELM);
}
SECTION("H2D-D2D-D2H-DeviceContextChange") {
int deviceCount = 0;
HIP_CHECK(hipGetDeviceCount(&deviceCount));
if (deviceCount < 2) {
SUCCEED("deviceCount less then 2");
} else {
int canAccessPeer = 0;
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
if (canAccessPeer) {
HIP_CHECK(hipSetDevice(1));
HIP_CHECK(hipMemcpy(A_d, A_h, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_d, A_d, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_h, B_d, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HipTest::checkTest(A_h, B_h, NUM_ELM);
} else {
SUCCEED("P2P capability is not present");
}
}
}
SECTION("H2D-D2D-D2H-PeerGPU") {
int deviceCount = 0;
HIP_CHECK(hipGetDeviceCount(&deviceCount));
if (deviceCount < 2) {
SUCCEED("deviceCount less then 2");
} else {
int canAccessPeer = 0;
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
if (canAccessPeer) {
HIP_CHECK(hipSetDevice(1));
TestType *C_d{nullptr};
HipTest::initArrays<TestType>(nullptr, nullptr, &C_d,
nullptr, nullptr, nullptr,
NUM_ELM*sizeof(TestType));
HIP_CHECK(hipMemcpy(A_d, A_h, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(C_d, A_d, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_h, C_d, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HipTest::checkTest(A_h, B_h, NUM_ELM);
HIP_CHECK(hipFree(C_d));
} else {
SUCCEED("P2P capability is not present");
}
}
}
HipTest::freeArrays<TestType>(A_d, B_d, nullptr, A_h, B_h, nullptr, false);
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr, A_Ph,
B_Ph, nullptr, true);
}
/*
This testcase verifies the multi thread scenario
*/
TEST_CASE("Unit_hipMemcpy_MultiThreadWithSerialization") {
HIP_CHECK(hipDeviceReset());
// Simplest cases: serialize the threads, and also used pinned memory:
// This verifies that the sub-calls to memcpytest2 are correct.
multiThread_1<float>(true, true);
// Serialize, but use unpinned memory to stress the unpinned memory xfer path.
multiThread_1<float>(true, false);
}
/*
This testcase verifies hipMemcpy API with pinnedMemory and hostRegister
along with kernel launches
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy_PinnedRegMemWithKernelLaunch",
"", int, float, double) {
int numDevices = 0;
HIP_CHECK(hipGetDeviceCount(&numDevices));
if (numDevices < 2) {
SUCCEED("No of devices are less than 2");
} else {
// 1 refers to pinned Memory
// 2 refers to register Memory
int MallocPinType = GENERATE(0, 1);
size_t Nbytes = NUM_ELM * sizeof(TestType);
unsigned blocks = HipTest::setNumBlocks(blocksPerCU,
threadsPerBlock, NUM_ELM);
TestType *A_d{nullptr}, *B_d{nullptr}, *C_d{nullptr};
TestType *X_d{nullptr}, *Y_d{nullptr}, *Z_d{nullptr};
TestType *A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr};
if (MallocPinType) {
HipTest::initArrays(&A_d, &B_d, &C_d, &A_h, &B_h, &C_h, NUM_ELM, true);
} else {
A_h = reinterpret_cast<TestType*>(malloc(Nbytes));
HIP_CHECK(hipHostRegister(A_h, Nbytes, hipHostRegisterDefault));
B_h = reinterpret_cast<TestType*>(malloc(Nbytes));
HIP_CHECK(hipHostRegister(B_h, Nbytes, hipHostRegisterDefault));
C_h = reinterpret_cast<TestType*>(malloc(Nbytes));
HIP_CHECK(hipHostRegister(C_h, Nbytes, hipHostRegisterDefault));
HipTest::initArrays<TestType>(&A_d, &B_d, &C_d, nullptr, nullptr,
nullptr, NUM_ELM, false);
HipTest::setDefaultData<TestType>(NUM_ELM, A_h, B_h, C_h);
}
HIP_CHECK(hipMemcpy(A_d, A_h, Nbytes, hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy(B_d, B_h, Nbytes, hipMemcpyHostToDevice));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, 0, static_cast<const TestType*>(A_d),
static_cast<const TestType*>(B_d), C_d, NUM_ELM);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipMemcpy(C_h, C_d, Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h, B_h, C_h, NUM_ELM);
unsigned int seed = time(0);
HIP_CHECK(hipSetDevice(HipTest::RAND_R(&seed) % (numDevices-1)+1));
int device;
HIP_CHECK(hipGetDevice(&device));
std::cout <<"hipMemcpy is set to happen between device 0 and device "
<<device << std::endl;
HipTest::initArrays<TestType>(&X_d, &Y_d, &Z_d, nullptr,
nullptr, nullptr, NUM_ELM, false);
for (int j = 0; j < NUM_ELM; j++) {
A_h[j] = 0;
B_h[j] = 0;
C_h[j] = 0;
}
HIP_CHECK(hipMemcpy(A_h, A_d, Nbytes, hipMemcpyDeviceToHost));
HIP_CHECK(hipMemcpy(X_d, A_h, Nbytes, hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy(B_h, B_d, Nbytes, hipMemcpyDeviceToHost));
HIP_CHECK(hipMemcpy(Y_d, B_h, Nbytes, hipMemcpyHostToDevice));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, 0, static_cast<const TestType*>(X_d),
static_cast<const TestType*>(Y_d), Z_d, NUM_ELM);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipMemcpy(C_h, Z_d, Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h, B_h, C_h, NUM_ELM);
if (MallocPinType) {
HipTest::freeArrays<TestType>(A_d, B_d, C_d, A_h, B_h, C_h, true);
} else {
HIP_CHECK(hipHostUnregister(A_h));
free(A_h);
HIP_CHECK(hipHostUnregister(B_h));
free(B_h);
HIP_CHECK(hipHostUnregister(C_h));
free(C_h);
HipTest::freeArrays<TestType>(A_d, B_d, C_d, nullptr,
nullptr, nullptr, false);
}
HipTest::freeArrays<TestType>(X_d, Y_d, Z_d, nullptr,
nullptr, nullptr, false);
SECTION("Host to host") {
LinearAllocGuard<int> src_alloc(LinearAllocs::hipHostMalloc, kPageSize);
LinearAllocGuard<int> dst_alloc(LinearAllocs::hipHostMalloc, kPageSize);
MemcpyWithDirectionCommonNegativeTests(hipMemcpy, dst_alloc.ptr(), src_alloc.ptr(), kPageSize,
hipMemcpyHostToHost);
}
SECTION("Device to device") {
LinearAllocGuard<int> src_alloc(LinearAllocs::hipMalloc, kPageSize);
LinearAllocGuard<int> dst_alloc(LinearAllocs::hipMalloc, kPageSize);
MemcpyWithDirectionCommonNegativeTests(hipMemcpy, dst_alloc.ptr(), src_alloc.ptr(), kPageSize,
hipMemcpyDeviceToDevice);
}
}
catch/unit/memory/hipMemcpyWithStream.cc
+72 -557
Voir le fichier
@@ -1,13 +1,15 @@
/*
Copyright (c) 2021-22-present 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 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
@@ -17,570 +19,83 @@ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.
*/
/*
* Different test for checking functionality of
* hipError_t hipMemcpyWithStream(void* dst, const void* src, size_t sizeBytes,hipMemcpyKind kind,
* hipStream_t stream);
*/
/*
This testfile verifies the following scenarios
1. hipMemcpyWithStream with one stream
2. hipMemcpyWithStream with two streams
3. Multi GPU and single stream
4. hipMemcpyWithStream API with testkind DtoH
5. hipMemcpyWithStream API with testkind DtoD
6. hipMemcpyWithStream API with testkind HtoH
7. hipMemcpyWithStream API with testkind TestkindDefault
8. hipMemcpyWithStream API with testkind TestkindDefaultForDtoD
9. hipMemcpyWithStream API DtoD on same device
*/
#include <hip_test_common.hh>
#include <hip_test_kernels.hh>
#include <hip_test_checkers.hh>
#include <hip/hip_runtime_api.h>
#include <memcpy1d_tests_common.hh>
#include <resource_guards.hh>
#include <utils.hh>
#include<vector>
#include<thread>
#include<chrono>
TEST_CASE("Unit_hipMemcpy_Positive_Basic") { MemcpyWithDirectionCommonTests<false>(hipMemcpy); }
#define LEN 64
#define SIZE LEN << 2
#define THREADS 2
#define MAX_THREADS 16
TEST_CASE("Unit_hipMemcpy_Positive_Synchronization_Behavior") {
using namespace std::placeholders;
HIP_CHECK(hipDeviceSynchronize());
static constexpr size_t N{4 * 1024 * 1024};
static const auto MaxGPUDevices{256};
static constexpr unsigned blocksPerCU{6}; // to hide latency
static constexpr unsigned threadsPerBlock{256};
enum class ops
{ TestwithOnestream,
TestwithTwoStream,
TestOnMultiGPUwithOneStream,
TestkindDtoH,
TestkindDtoD,
TestkindHtoH,
TestkindDefault,
TestkindDefaultForDtoD,
TestDtoDonSameDevice,
END_OF_LIST
};
struct joinable_thread : std::thread {
template <class... Xs>
explicit joinable_thread(Xs&&... xs) : std::thread(std::forward<Xs>(xs)...)
{} // NOLINT
joinable_thread& operator=(joinable_thread&& other) = default;
joinable_thread(joinable_thread&& other) = default;
~joinable_thread() {
if (this->joinable())
this->join();
}
};
void TestwithOnestream(void) {
size_t Nbytes = N * sizeof(int);
int *A_d, *B_d, *C_d;
int *A_h, *B_h, *C_h;
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HipTest::initArrays(&A_d, &B_d, &C_d, &A_h, &B_h, &C_h, N, false);
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
HIP_CHECK(hipMemcpyWithStream(A_d, A_h, Nbytes,
hipMemcpyHostToDevice, stream));
HIP_CHECK(hipMemcpyWithStream(B_d, B_h, Nbytes,
hipMemcpyHostToDevice, stream));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, stream, static_cast<const int*>(A_d),
static_cast<const int*>(B_d), C_d, N);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipStreamSynchronize(stream));
HIP_CHECK(hipMemcpy(C_h, C_d, Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h, B_h, C_h, N);
HipTest::freeArrays(A_d, B_d, C_d, A_h, B_h, C_h, false);
HIP_CHECK(hipStreamDestroy(stream));
}
void TestwithTwoStream(void) {
size_t Nbytes = N * sizeof(int);
const int NUM_STREAMS = 2;
int *A_d[NUM_STREAMS], *B_d[NUM_STREAMS], *C_d[NUM_STREAMS];
int *A_h[NUM_STREAMS], *B_h[NUM_STREAMS], *C_h[NUM_STREAMS];
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
for (int i=0; i < NUM_STREAMS; ++i) {
HipTest::initArrays(&A_d[i], &B_d[i], &C_d[i],
&A_h[i], &B_h[i], &C_h[i], N, false);
// For transfers from pageable host memory to device memory, a stream sync is performed before
// the copy is initiated. The function will return once the pageable buffer has been copied to
// the staging memory for DMA transfer to device memory, but the DMA to final destination may
// not have completed.
// For transfers from pinned host memory to device memory, the function is synchronous with
// respect to the host
SECTION("Host memory to device memory") {
MemcpyHtoDSyncBehavior(std::bind(hipMemcpy, _1, _2, _3, hipMemcpyHostToDevice), true);
}
hipStream_t stream[NUM_STREAMS];
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipStreamCreate(&stream[i]));
// For transfers from device to either pageable or pinned host memory, the function returns only
// once the copy has completed
SECTION("Device memory to host memory") {
const auto f = std::bind(hipMemcpy, _1, _2, _3, hipMemcpyDeviceToHost);
MemcpyDtoHPageableSyncBehavior(f, true);
MemcpyDtoHPinnedSyncBehavior(f, true);
}
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipMemcpyWithStream(A_d[i], A_h[i], Nbytes,
hipMemcpyHostToDevice, stream[i]));
HIP_CHECK(hipMemcpyWithStream(B_d[i], B_h[i], Nbytes,
hipMemcpyHostToDevice, stream[i]));
}
for (int i=0; i < NUM_STREAMS; ++i) {
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, stream[i], static_cast<const int*>(A_d[i]),
static_cast<const int*>(B_d[i]), C_d[i], N);
HIP_CHECK(hipGetLastError());
}
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipStreamSynchronize(stream[i]));
HIP_CHECK(hipMemcpy(C_h[i], C_d[i], Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h[i], B_h[i], C_h[i], N);
}
for (int i=0; i < NUM_STREAMS; ++i) {
HipTest::freeArrays(A_d[i], B_d[i], C_d[i], A_h[i], B_h[i], C_h[i], false);
HIP_CHECK(hipStreamDestroy(stream[i]));
}
}
void TestDtoDonSameDevice(void) {
size_t Nbytes = N * sizeof(int);
const int NUM_STREAMS = 2;
int *A_d[NUM_STREAMS], *B_d[NUM_STREAMS], *C_d[NUM_STREAMS];
int *A_h[NUM_STREAMS], *B_h[NUM_STREAMS], *C_h[NUM_STREAMS];
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HipTest::initArrays(&A_d[0], &B_d[0], &C_d[0],
&A_h[0], &B_h[0], &C_h[0], N, false);
hipStream_t stream[NUM_STREAMS];
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipSetDevice(0));
HIP_CHECK(hipStreamCreate(&stream[i]));
}
HIP_CHECK(hipSetDevice(0));
HIP_CHECK(hipMalloc(&A_d[1], Nbytes));
HIP_CHECK(hipMalloc(&B_d[1], Nbytes));
HIP_CHECK(hipMalloc(&C_d[1], Nbytes));
C_h[1] = reinterpret_cast<int*>(malloc(Nbytes));
HIP_ASSERT(C_h[1] != NULL);
HIP_CHECK(hipMemcpyWithStream(A_d[0], A_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
HIP_CHECK(hipMemcpyWithStream(B_d[0], B_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
HIP_CHECK(hipMemcpyWithStream(A_d[1], A_d[0], Nbytes,
hipMemcpyDeviceToDevice, stream[1]));
HIP_CHECK(hipMemcpyWithStream(B_d[1], B_d[0], Nbytes,
hipMemcpyDeviceToDevice, stream[1]));
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipSetDevice(0));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, stream[i], static_cast<const int*>(A_d[i]),
static_cast<const int*>(B_d[i]), C_d[i], N);
HIP_CHECK(hipGetLastError());
}
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipSetDevice(0));
HIP_CHECK(hipStreamSynchronize(stream[i]));
HIP_CHECK(hipMemcpy(C_h[i], C_d[i], Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h[0], B_h[0], C_h[i], N);
}
HipTest::freeArrays(A_d[0], B_d[0], C_d[0], A_h[0], B_h[0], C_h[0], false);
if (A_d[1]) {
HIP_CHECK(hipFree(A_d[1]));
}
if (B_d[1]) {
HIP_CHECK(hipFree(B_d[1]));
}
if (C_d[1]) {
HIP_CHECK(hipFree(C_d[1]));
}
if (C_h[1]) {
free(C_h[1]);
}
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipStreamDestroy(stream[i]));
}
}
void TestOnMultiGPUwithOneStream(void) {
size_t Nbytes = N * sizeof(int);
int NumDevices = 0;
HIP_CHECK(hipGetDeviceCount(&NumDevices));
// If you have single GPU machine the return
if (NumDevices <= 1) {
SUCCEED("NumDevices <2");
} else {
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
int *A_d[MaxGPUDevices], *B_d[MaxGPUDevices], *C_d[MaxGPUDevices];
int *A_h[MaxGPUDevices], *B_h[MaxGPUDevices], *C_h[MaxGPUDevices];
hipStream_t stream[MaxGPUDevices];
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipStreamCreate(&stream[i]));
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HipTest::initArrays(&A_d[i], &B_d[i], &C_d[i],
&A_h[i], &B_h[i], &C_h[i], N, false);
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipMemcpyWithStream(A_d[i], A_h[i], Nbytes,
hipMemcpyHostToDevice, stream[i]));
HIP_CHECK(hipMemcpyWithStream(B_d[i], B_h[i], Nbytes,
hipMemcpyHostToDevice, stream[i]));
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks),
dim3(threadsPerBlock), 0, stream[i],
static_cast<const int*>(A_d[i]),
static_cast<const int*>(B_d[i]), C_d[i], N);
HIP_CHECK(hipGetLastError());
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipStreamSynchronize(stream[i]));
HIP_CHECK(hipMemcpy(C_h[i], C_d[i], Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h[i], B_h[i], C_h[i], N);
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HipTest::freeArrays(A_d[i], B_d[i], C_d[i],
A_h[i], B_h[i], C_h[i], false);
HIP_CHECK(hipStreamDestroy(stream[i]));
}
}
}
void TestkindDtoH(void) {
size_t Nbytes = N * sizeof(int);
int *A_d, *B_d, *C_d;
int *A_h, *B_h, *C_h;
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HipTest::initArrays(&A_d, &B_d, &C_d, &A_h, &B_h, &C_h, N, false);
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
HIP_CHECK(hipMemcpyWithStream(A_d, A_h, Nbytes,
hipMemcpyHostToDevice, stream));
HIP_CHECK(hipMemcpyWithStream(B_d, B_h, Nbytes,
hipMemcpyHostToDevice, stream));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, stream, static_cast<const int*>(A_d),
static_cast<const int*>(B_d), C_d, N);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipStreamSynchronize(stream));
HIP_CHECK(hipMemcpyWithStream(C_h, C_d, Nbytes,
hipMemcpyDeviceToHost, stream));
HipTest::checkVectorADD(A_h, B_h, C_h, N);
HipTest::freeArrays(A_d, B_d, C_d, A_h, B_h, C_h, false);
HIP_CHECK(hipStreamDestroy(stream));
}
void TestkindDtoD(void) {
size_t Nbytes = N * sizeof(int);
int NumDevices = 0;
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HIP_CHECK(hipGetDeviceCount(&NumDevices));
// If you have single GPU machine the return
if (NumDevices <= 1) {
SUCCEED("NumDevices are less than 2");
} else {
int *A_d[MaxGPUDevices], *B_d[MaxGPUDevices], *C_d[MaxGPUDevices];
int *A_h[MaxGPUDevices], *B_h[MaxGPUDevices], *C_h[MaxGPUDevices];
hipStream_t stream[MaxGPUDevices];
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipStreamCreate(&stream[i]));
}
// Initialize and create the host and device elements for first device
HIP_CHECK(hipSetDevice(0));
HipTest::initArrays(&A_d[0], &B_d[0], &C_d[0],
&A_h[0], &B_h[0], &C_h[0], N, false);
for (int i=1; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i))
HIP_CHECK(hipMalloc(&A_d[i], Nbytes));
HIP_CHECK(hipMalloc(&B_d[i], Nbytes));
HIP_CHECK(hipMalloc(&C_d[i], Nbytes));
C_h[i] = reinterpret_cast<int*>(malloc(Nbytes));
HIP_ASSERT(C_h[i] != NULL);
}
HIP_CHECK(hipSetDevice(0));
HIP_CHECK(hipMemcpyWithStream(A_d[0], A_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
HIP_CHECK(hipMemcpyWithStream(B_d[0], B_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
// Copying device data from 1st GPU to the rest of the the GPUs that is
// NumDevices in the setup. 1st GPU start numbering from 0,1,2..n etc.
for (int i=1; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipMemcpyWithStream(A_d[i], A_d[0], Nbytes,
hipMemcpyDeviceToDevice, stream[i]));
HIP_CHECK(hipMemcpyWithStream(B_d[i], B_d[0], Nbytes,
hipMemcpyDeviceToDevice, stream[i]));
}
// Launching the kernel including the 1st GPU to the no of GPUs present
// in the setup. 1st GPU start numbering from 0,1,2..n etc.
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks),
dim3(threadsPerBlock),
0, stream[i], static_cast<const int*>(A_d[i]),
static_cast<const int*>(B_d[i]), C_d[i], N);
HIP_CHECK(hipGetLastError());
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipStreamSynchronize(stream[i]));
HIP_CHECK(hipMemcpy(C_h[i], C_d[i], Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h[0], B_h[0], C_h[i], N);
}
HipTest::freeArrays(A_d[0], B_d[0], C_d[0], A_h[0], B_h[0], C_h[0], false);
HIP_CHECK(hipStreamDestroy(stream[0]));
for (int i=1; i < NumDevices; ++i) {
if (A_d[i]) {
HIP_CHECK(hipFree(A_d[i]));
}
if (B_d[i]) {
HIP_CHECK(hipFree(B_d[i]));
}
if (C_d[i]) {
HIP_CHECK(hipFree(C_d[i]));
}
if (C_h[i]) {
free(C_h[i]);
}
HIP_CHECK(hipStreamDestroy(stream[i]));
}
}
}
void TestkindDefault(void) {
size_t Nbytes = N * sizeof(int);
int *A_d, *B_d, *C_d;
int *A_h, *B_h, *C_h;
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HipTest::initArrays(&A_d, &B_d, &C_d, &A_h, &B_h, &C_h, N, false);
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
HIP_CHECK(hipMemcpyWithStream(A_d, A_h, Nbytes, hipMemcpyDefault, stream));
HIP_CHECK(hipMemcpyWithStream(B_d, B_h, Nbytes, hipMemcpyDefault, stream));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, stream, static_cast<const int*>(A_d),
static_cast<const int*>(B_d), C_d, N);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipStreamSynchronize(stream));
HIP_CHECK(hipMemcpyWithStream(C_h, C_d, Nbytes, hipMemcpyDefault, stream));
HipTest::checkVectorADD(A_h, B_h, C_h, N);
HipTest::freeArrays(A_d, B_d, C_d, A_h, B_h, C_h, false);
HIP_CHECK(hipStreamDestroy(stream));
}
void TestkindDefaultForDtoD(void) {
size_t Nbytes = N * sizeof(int);
int NumDevices = 0;
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HIP_CHECK(hipGetDeviceCount(&NumDevices));
// Test case will not run on single GPU setup.
if (NumDevices <= 1) {
SUCCEED("No of Devices < 2");
} else {
int *A_d[MaxGPUDevices], *B_d[MaxGPUDevices], *C_d[MaxGPUDevices];
int *A_h[MaxGPUDevices], *B_h[MaxGPUDevices], *C_h[MaxGPUDevices];
// Initialize and create the host and device elements for first device
HIP_CHECK(hipSetDevice(0));
HipTest::initArrays(&A_d[0], &B_d[0], &C_d[0],
&A_h[0], &B_h[0], &C_h[0], N, false);
for (int i=1; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipMalloc(&A_d[i], Nbytes));
HIP_CHECK(hipMalloc(&B_d[i], Nbytes));
HIP_CHECK(hipMalloc(&C_d[i], Nbytes));
C_h[i] = reinterpret_cast<int*>(malloc(Nbytes));
HIP_ASSERT(C_h[i] != NULL);
}
hipStream_t stream[MaxGPUDevices];
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipStreamCreate(&stream[i]));
}
HIP_CHECK(hipMemcpyWithStream(A_d[0], A_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
HIP_CHECK(hipMemcpyWithStream(B_d[0], B_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
// Copying device data from 1st GPU to the rest of the the GPUs
// using hipMemcpyDefault kind that is NumDevices in the setup.
// 1st GPU start numbering from 0,1,2..n etc.
for (int i=1; i < NumDevices; ++i) {
HIP_CHECK(hipMemcpyWithStream(A_d[i], A_d[0], Nbytes,
hipMemcpyDefault, stream[i]));
HIP_CHECK(hipMemcpyWithStream(B_d[i], B_d[0], Nbytes,
hipMemcpyDefault, stream[i]));
}
for (int i=0; i < NumDevices; ++i) {
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks),
dim3(threadsPerBlock),
0, stream[i], static_cast<const int*>(A_d[i]),
static_cast<const int*>(B_d[i]), C_d[i], N);
HIP_CHECK(hipGetLastError());
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i)); // hipMemcpy will be on this device
HIP_CHECK(hipStreamSynchronize(stream[i]));
HIP_CHECK(hipMemcpy(C_h[i], C_d[i], Nbytes, hipMemcpyDeviceToHost));
// Output of each GPU is getting validated with input of 1st GPU.
HipTest::checkVectorADD(A_h[0], B_h[0], C_h[i], N);
}
HipTest::freeArrays(A_d[0], B_d[0], C_d[0], A_h[0], B_h[0], C_h[0], false);
HIP_CHECK(hipStreamDestroy(stream[0]));
for (int i=1; i < NumDevices; ++i) {
if (A_d[i]) {
HIP_CHECK(hipFree(A_d[i]));
}
if (B_d[i]) {
HIP_CHECK(hipFree(B_d[i]));
}
if (C_d[i]) {
HIP_CHECK(hipFree(C_d[i]));
}
if (C_h[i]) {
free(C_h[i]);
}
HIP_CHECK(hipStreamDestroy(stream[i]));
}
}
}
void TestkindHtoH(void) {
size_t Nbytes = N * sizeof(int);
int *A_h, *B_h;
// Allocate memory to A_h and B_h
A_h = static_cast<int*>(malloc(Nbytes));
HIP_ASSERT(A_h != NULL);
B_h = static_cast<int*>(malloc(Nbytes));
HIP_ASSERT(B_h != NULL);
for (size_t i = 0; i < N; ++i) {
if (A_h) {
(A_h)[i] = 3.146f + i; // Pi
}
}
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
HIP_CHECK(hipMemcpyWithStream(B_h, A_h, Nbytes, hipMemcpyHostToHost, stream));
HIP_CHECK(hipStreamSynchronize(stream));
for (size_t i = 0; i < N; i++) {
HIP_ASSERT(A_h[i] == B_h[i]);
}
if (A_h) {
free(A_h);
}
if (B_h) {
free(B_h);
}
HIP_CHECK(hipStreamDestroy(stream));
}
TEST_CASE("Unit_hipMemcpyWithStream_TestWithOneStream") {
TestwithOnestream();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestwithTwoStream") {
TestwithTwoStream();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestkindDtoH") {
TestkindDtoH();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestkindHtoH") {
TestkindHtoH();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestkindDtoD") {
TestkindDtoD();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestOnMultiGPUwithOneStream") {
TestOnMultiGPUwithOneStream();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestkindDefault") {
TestkindDefault();
}
#ifndef __HIP_PLATFORM_NVCC__
TEST_CASE("Unit_hipMemcpyWithStream_TestkindDefaultForDtoD") {
TestkindDefaultForDtoD();
}
// For transfers from device memory to device memory, no host-side synchronization is performed.
SECTION("Device memory to device memory") {
// This behavior differs on NVIDIA and AMD, on AMD the hipMemcpy calls is synchronous with
// respect to the host
#if HT_AMD
HipTest::HIP_SKIP_TEST(
"EXSWCPHIPT-127 - Memcpy from device to device memory behavior differs on AMD and Nvidia");
return;
#endif
MemcpyDtoDSyncBehavior(std::bind(hipMemcpy, _1, _2, _3, hipMemcpyDeviceToDevice), false);
}
TEST_CASE("Unit_hipMemcpyWithStream_TestDtoDonSameDevice") {
TestDtoDonSameDevice();
// For transfers from any host memory to any host memory, the function is fully synchronous with
// respect to the host
SECTION("Host memory to host memory") {
MemcpyHtoHSyncBehavior(std::bind(hipMemcpy, _1, _2, _3, hipMemcpyHostToHost), true);
}
}
TEST_CASE("Unit_hipMemcpy_Negative_Parameters") {
using namespace std::placeholders;
SECTION("Host to device") {
LinearAllocGuard<int> device_alloc(LinearAllocs::hipMalloc, kPageSize);
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, kPageSize);
MemcpyWithDirectionCommonNegativeTests(hipMemcpy, device_alloc.ptr(), host_alloc.ptr(),
kPageSize, hipMemcpyHostToDevice);
}
SECTION("Device to host") {
LinearAllocGuard<int> device_alloc(LinearAllocs::hipMalloc, kPageSize);
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, kPageSize);
MemcpyWithDirectionCommonNegativeTests(hipMemcpy, host_alloc.ptr(), device_alloc.ptr(),
kPageSize, hipMemcpyDeviceToHost);
}
SECTION("Host to host") {
LinearAllocGuard<int> src_alloc(LinearAllocs::hipHostMalloc, kPageSize);
LinearAllocGuard<int> dst_alloc(LinearAllocs::hipHostMalloc, kPageSize);
MemcpyWithDirectionCommonNegativeTests(hipMemcpy, dst_alloc.ptr(), src_alloc.ptr(), kPageSize,
hipMemcpyHostToHost);
}
SECTION("Device to device") {
LinearAllocGuard<int> src_alloc(LinearAllocs::hipMalloc, kPageSize);
LinearAllocGuard<int> dst_alloc(LinearAllocs::hipMalloc, kPageSize);
MemcpyWithDirectionCommonNegativeTests(hipMemcpy, dst_alloc.ptr(), src_alloc.ptr(), kPageSize,
hipMemcpyDeviceToDevice);
}
}
catch/unit/memory/hipMemcpyWithStream_old.cc
+586
Voir le fichier
@@ -0,0 +1,586 @@
/*
Copyright (c) 2021-22-present 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.
*/
/*
* Different test for checking functionality of
* hipError_t hipMemcpyWithStream(void* dst, const void* src, size_t sizeBytes,hipMemcpyKind kind,
* hipStream_t stream);
*/
/*
This testfile verifies the following scenarios
1. hipMemcpyWithStream with one stream
2. hipMemcpyWithStream with two streams
3. Multi GPU and single stream
4. hipMemcpyWithStream API with testkind DtoH
5. hipMemcpyWithStream API with testkind DtoD
6. hipMemcpyWithStream API with testkind HtoH
7. hipMemcpyWithStream API with testkind TestkindDefault
8. hipMemcpyWithStream API with testkind TestkindDefaultForDtoD
9. hipMemcpyWithStream API DtoD on same device
*/
#include <hip_test_common.hh>
#include <hip_test_kernels.hh>
#include <hip_test_checkers.hh>
#include<vector>
#include<thread>
#include<chrono>
#define LEN 64
#define SIZE LEN << 2
#define THREADS 2
#define MAX_THREADS 16
static constexpr size_t N{4 * 1024 * 1024};
static const auto MaxGPUDevices{256};
static constexpr unsigned blocksPerCU{6}; // to hide latency
static constexpr unsigned threadsPerBlock{256};
enum class ops
{ TestwithOnestream,
TestwithTwoStream,
TestOnMultiGPUwithOneStream,
TestkindDtoH,
TestkindDtoD,
TestkindHtoH,
TestkindDefault,
TestkindDefaultForDtoD,
TestDtoDonSameDevice,
END_OF_LIST
};
struct joinable_thread : std::thread {
template <class... Xs>
explicit joinable_thread(Xs&&... xs) : std::thread(std::forward<Xs>(xs)...)
{} // NOLINT
joinable_thread& operator=(joinable_thread&& other) = default;
joinable_thread(joinable_thread&& other) = default;
~joinable_thread() {
if (this->joinable())
this->join();
}
};
void TestwithOnestream(void) {
size_t Nbytes = N * sizeof(int);
int *A_d, *B_d, *C_d;
int *A_h, *B_h, *C_h;
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HipTest::initArrays(&A_d, &B_d, &C_d, &A_h, &B_h, &C_h, N, false);
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
HIP_CHECK(hipMemcpyWithStream(A_d, A_h, Nbytes,
hipMemcpyHostToDevice, stream));
HIP_CHECK(hipMemcpyWithStream(B_d, B_h, Nbytes,
hipMemcpyHostToDevice, stream));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, stream, static_cast<const int*>(A_d),
static_cast<const int*>(B_d), C_d, N);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipStreamSynchronize(stream));
HIP_CHECK(hipMemcpy(C_h, C_d, Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h, B_h, C_h, N);
HipTest::freeArrays(A_d, B_d, C_d, A_h, B_h, C_h, false);
HIP_CHECK(hipStreamDestroy(stream));
}
void TestwithTwoStream(void) {
size_t Nbytes = N * sizeof(int);
const int NUM_STREAMS = 2;
int *A_d[NUM_STREAMS], *B_d[NUM_STREAMS], *C_d[NUM_STREAMS];
int *A_h[NUM_STREAMS], *B_h[NUM_STREAMS], *C_h[NUM_STREAMS];
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
for (int i=0; i < NUM_STREAMS; ++i) {
HipTest::initArrays(&A_d[i], &B_d[i], &C_d[i],
&A_h[i], &B_h[i], &C_h[i], N, false);
}
hipStream_t stream[NUM_STREAMS];
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipStreamCreate(&stream[i]));
}
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipMemcpyWithStream(A_d[i], A_h[i], Nbytes,
hipMemcpyHostToDevice, stream[i]));
HIP_CHECK(hipMemcpyWithStream(B_d[i], B_h[i], Nbytes,
hipMemcpyHostToDevice, stream[i]));
}
for (int i=0; i < NUM_STREAMS; ++i) {
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, stream[i], static_cast<const int*>(A_d[i]),
static_cast<const int*>(B_d[i]), C_d[i], N);
HIP_CHECK(hipGetLastError());
}
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipStreamSynchronize(stream[i]));
HIP_CHECK(hipMemcpy(C_h[i], C_d[i], Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h[i], B_h[i], C_h[i], N);
}
for (int i=0; i < NUM_STREAMS; ++i) {
HipTest::freeArrays(A_d[i], B_d[i], C_d[i], A_h[i], B_h[i], C_h[i], false);
HIP_CHECK(hipStreamDestroy(stream[i]));
}
}
void TestDtoDonSameDevice(void) {
size_t Nbytes = N * sizeof(int);
const int NUM_STREAMS = 2;
int *A_d[NUM_STREAMS], *B_d[NUM_STREAMS], *C_d[NUM_STREAMS];
int *A_h[NUM_STREAMS], *B_h[NUM_STREAMS], *C_h[NUM_STREAMS];
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HipTest::initArrays(&A_d[0], &B_d[0], &C_d[0],
&A_h[0], &B_h[0], &C_h[0], N, false);
hipStream_t stream[NUM_STREAMS];
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipSetDevice(0));
HIP_CHECK(hipStreamCreate(&stream[i]));
}
HIP_CHECK(hipSetDevice(0));
HIP_CHECK(hipMalloc(&A_d[1], Nbytes));
HIP_CHECK(hipMalloc(&B_d[1], Nbytes));
HIP_CHECK(hipMalloc(&C_d[1], Nbytes));
C_h[1] = reinterpret_cast<int*>(malloc(Nbytes));
HIP_ASSERT(C_h[1] != NULL);
HIP_CHECK(hipMemcpyWithStream(A_d[0], A_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
HIP_CHECK(hipMemcpyWithStream(B_d[0], B_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
HIP_CHECK(hipMemcpyWithStream(A_d[1], A_d[0], Nbytes,
hipMemcpyDeviceToDevice, stream[1]));
HIP_CHECK(hipMemcpyWithStream(B_d[1], B_d[0], Nbytes,
hipMemcpyDeviceToDevice, stream[1]));
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipSetDevice(0));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, stream[i], static_cast<const int*>(A_d[i]),
static_cast<const int*>(B_d[i]), C_d[i], N);
HIP_CHECK(hipGetLastError());
}
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipSetDevice(0));
HIP_CHECK(hipStreamSynchronize(stream[i]));
HIP_CHECK(hipMemcpy(C_h[i], C_d[i], Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h[0], B_h[0], C_h[i], N);
}
HipTest::freeArrays(A_d[0], B_d[0], C_d[0], A_h[0], B_h[0], C_h[0], false);
if (A_d[1]) {
HIP_CHECK(hipFree(A_d[1]));
}
if (B_d[1]) {
HIP_CHECK(hipFree(B_d[1]));
}
if (C_d[1]) {
HIP_CHECK(hipFree(C_d[1]));
}
if (C_h[1]) {
free(C_h[1]);
}
for (int i=0; i < NUM_STREAMS; ++i) {
HIP_CHECK(hipStreamDestroy(stream[i]));
}
}
void TestOnMultiGPUwithOneStream(void) {
size_t Nbytes = N * sizeof(int);
int NumDevices = 0;
HIP_CHECK(hipGetDeviceCount(&NumDevices));
// If you have single GPU machine the return
if (NumDevices <= 1) {
SUCCEED("NumDevices <2");
} else {
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
int *A_d[MaxGPUDevices], *B_d[MaxGPUDevices], *C_d[MaxGPUDevices];
int *A_h[MaxGPUDevices], *B_h[MaxGPUDevices], *C_h[MaxGPUDevices];
hipStream_t stream[MaxGPUDevices];
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipStreamCreate(&stream[i]));
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HipTest::initArrays(&A_d[i], &B_d[i], &C_d[i],
&A_h[i], &B_h[i], &C_h[i], N, false);
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipMemcpyWithStream(A_d[i], A_h[i], Nbytes,
hipMemcpyHostToDevice, stream[i]));
HIP_CHECK(hipMemcpyWithStream(B_d[i], B_h[i], Nbytes,
hipMemcpyHostToDevice, stream[i]));
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks),
dim3(threadsPerBlock), 0, stream[i],
static_cast<const int*>(A_d[i]),
static_cast<const int*>(B_d[i]), C_d[i], N);
HIP_CHECK(hipGetLastError());
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipStreamSynchronize(stream[i]));
HIP_CHECK(hipMemcpy(C_h[i], C_d[i], Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h[i], B_h[i], C_h[i], N);
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HipTest::freeArrays(A_d[i], B_d[i], C_d[i],
A_h[i], B_h[i], C_h[i], false);
HIP_CHECK(hipStreamDestroy(stream[i]));
}
}
}
void TestkindDtoH(void) {
size_t Nbytes = N * sizeof(int);
int *A_d, *B_d, *C_d;
int *A_h, *B_h, *C_h;
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HipTest::initArrays(&A_d, &B_d, &C_d, &A_h, &B_h, &C_h, N, false);
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
HIP_CHECK(hipMemcpyWithStream(A_d, A_h, Nbytes,
hipMemcpyHostToDevice, stream));
HIP_CHECK(hipMemcpyWithStream(B_d, B_h, Nbytes,
hipMemcpyHostToDevice, stream));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, stream, static_cast<const int*>(A_d),
static_cast<const int*>(B_d), C_d, N);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipStreamSynchronize(stream));
HIP_CHECK(hipMemcpyWithStream(C_h, C_d, Nbytes,
hipMemcpyDeviceToHost, stream));
HipTest::checkVectorADD(A_h, B_h, C_h, N);
HipTest::freeArrays(A_d, B_d, C_d, A_h, B_h, C_h, false);
HIP_CHECK(hipStreamDestroy(stream));
}
void TestkindDtoD(void) {
size_t Nbytes = N * sizeof(int);
int NumDevices = 0;
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HIP_CHECK(hipGetDeviceCount(&NumDevices));
// If you have single GPU machine the return
if (NumDevices <= 1) {
SUCCEED("NumDevices are less than 2");
} else {
int *A_d[MaxGPUDevices], *B_d[MaxGPUDevices], *C_d[MaxGPUDevices];
int *A_h[MaxGPUDevices], *B_h[MaxGPUDevices], *C_h[MaxGPUDevices];
hipStream_t stream[MaxGPUDevices];
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipStreamCreate(&stream[i]));
}
// Initialize and create the host and device elements for first device
HIP_CHECK(hipSetDevice(0));
HipTest::initArrays(&A_d[0], &B_d[0], &C_d[0],
&A_h[0], &B_h[0], &C_h[0], N, false);
for (int i=1; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i))
HIP_CHECK(hipMalloc(&A_d[i], Nbytes));
HIP_CHECK(hipMalloc(&B_d[i], Nbytes));
HIP_CHECK(hipMalloc(&C_d[i], Nbytes));
C_h[i] = reinterpret_cast<int*>(malloc(Nbytes));
HIP_ASSERT(C_h[i] != NULL);
}
HIP_CHECK(hipSetDevice(0));
HIP_CHECK(hipMemcpyWithStream(A_d[0], A_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
HIP_CHECK(hipMemcpyWithStream(B_d[0], B_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
// Copying device data from 1st GPU to the rest of the the GPUs that is
// NumDevices in the setup. 1st GPU start numbering from 0,1,2..n etc.
for (int i=1; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipMemcpyWithStream(A_d[i], A_d[0], Nbytes,
hipMemcpyDeviceToDevice, stream[i]));
HIP_CHECK(hipMemcpyWithStream(B_d[i], B_d[0], Nbytes,
hipMemcpyDeviceToDevice, stream[i]));
}
// Launching the kernel including the 1st GPU to the no of GPUs present
// in the setup. 1st GPU start numbering from 0,1,2..n etc.
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks),
dim3(threadsPerBlock),
0, stream[i], static_cast<const int*>(A_d[i]),
static_cast<const int*>(B_d[i]), C_d[i], N);
HIP_CHECK(hipGetLastError());
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipStreamSynchronize(stream[i]));
HIP_CHECK(hipMemcpy(C_h[i], C_d[i], Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h[0], B_h[0], C_h[i], N);
}
HipTest::freeArrays(A_d[0], B_d[0], C_d[0], A_h[0], B_h[0], C_h[0], false);
HIP_CHECK(hipStreamDestroy(stream[0]));
for (int i=1; i < NumDevices; ++i) {
if (A_d[i]) {
HIP_CHECK(hipFree(A_d[i]));
}
if (B_d[i]) {
HIP_CHECK(hipFree(B_d[i]));
}
if (C_d[i]) {
HIP_CHECK(hipFree(C_d[i]));
}
if (C_h[i]) {
free(C_h[i]);
}
HIP_CHECK(hipStreamDestroy(stream[i]));
}
}
}
void TestkindDefault(void) {
size_t Nbytes = N * sizeof(int);
int *A_d, *B_d, *C_d;
int *A_h, *B_h, *C_h;
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HipTest::initArrays(&A_d, &B_d, &C_d, &A_h, &B_h, &C_h, N, false);
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
HIP_CHECK(hipMemcpyWithStream(A_d, A_h, Nbytes, hipMemcpyDefault, stream));
HIP_CHECK(hipMemcpyWithStream(B_d, B_h, Nbytes, hipMemcpyDefault, stream));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, stream, static_cast<const int*>(A_d),
static_cast<const int*>(B_d), C_d, N);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipStreamSynchronize(stream));
HIP_CHECK(hipMemcpyWithStream(C_h, C_d, Nbytes, hipMemcpyDefault, stream));
HipTest::checkVectorADD(A_h, B_h, C_h, N);
HipTest::freeArrays(A_d, B_d, C_d, A_h, B_h, C_h, false);
HIP_CHECK(hipStreamDestroy(stream));
}
void TestkindDefaultForDtoD(void) {
size_t Nbytes = N * sizeof(int);
int NumDevices = 0;
unsigned blocks = HipTest::setNumBlocks(blocksPerCU, threadsPerBlock, N);
HIP_CHECK(hipGetDeviceCount(&NumDevices));
// Test case will not run on single GPU setup.
if (NumDevices <= 1) {
SUCCEED("No of Devices < 2");
} else {
int *A_d[MaxGPUDevices], *B_d[MaxGPUDevices], *C_d[MaxGPUDevices];
int *A_h[MaxGPUDevices], *B_h[MaxGPUDevices], *C_h[MaxGPUDevices];
// Initialize and create the host and device elements for first device
HIP_CHECK(hipSetDevice(0));
HipTest::initArrays(&A_d[0], &B_d[0], &C_d[0],
&A_h[0], &B_h[0], &C_h[0], N, false);
for (int i=1; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipMalloc(&A_d[i], Nbytes));
HIP_CHECK(hipMalloc(&B_d[i], Nbytes));
HIP_CHECK(hipMalloc(&C_d[i], Nbytes));
C_h[i] = reinterpret_cast<int*>(malloc(Nbytes));
HIP_ASSERT(C_h[i] != NULL);
}
hipStream_t stream[MaxGPUDevices];
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i));
HIP_CHECK(hipStreamCreate(&stream[i]));
}
HIP_CHECK(hipMemcpyWithStream(A_d[0], A_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
HIP_CHECK(hipMemcpyWithStream(B_d[0], B_h[0], Nbytes,
hipMemcpyHostToDevice, stream[0]));
// Copying device data from 1st GPU to the rest of the the GPUs
// using hipMemcpyDefault kind that is NumDevices in the setup.
// 1st GPU start numbering from 0,1,2..n etc.
for (int i=1; i < NumDevices; ++i) {
HIP_CHECK(hipMemcpyWithStream(A_d[i], A_d[0], Nbytes,
hipMemcpyDefault, stream[i]));
HIP_CHECK(hipMemcpyWithStream(B_d[i], B_d[0], Nbytes,
hipMemcpyDefault, stream[i]));
}
for (int i=0; i < NumDevices; ++i) {
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks),
dim3(threadsPerBlock),
0, stream[i], static_cast<const int*>(A_d[i]),
static_cast<const int*>(B_d[i]), C_d[i], N);
HIP_CHECK(hipGetLastError());
}
for (int i=0; i < NumDevices; ++i) {
HIP_CHECK(hipSetDevice(i)); // hipMemcpy will be on this device
HIP_CHECK(hipStreamSynchronize(stream[i]));
HIP_CHECK(hipMemcpy(C_h[i], C_d[i], Nbytes, hipMemcpyDeviceToHost));
// Output of each GPU is getting validated with input of 1st GPU.
HipTest::checkVectorADD(A_h[0], B_h[0], C_h[i], N);
}
HipTest::freeArrays(A_d[0], B_d[0], C_d[0], A_h[0], B_h[0], C_h[0], false);
HIP_CHECK(hipStreamDestroy(stream[0]));
for (int i=1; i < NumDevices; ++i) {
if (A_d[i]) {
HIP_CHECK(hipFree(A_d[i]));
}
if (B_d[i]) {
HIP_CHECK(hipFree(B_d[i]));
}
if (C_d[i]) {
HIP_CHECK(hipFree(C_d[i]));
}
if (C_h[i]) {
free(C_h[i]);
}
HIP_CHECK(hipStreamDestroy(stream[i]));
}
}
}
void TestkindHtoH(void) {
size_t Nbytes = N * sizeof(int);
int *A_h, *B_h;
// Allocate memory to A_h and B_h
A_h = static_cast<int*>(malloc(Nbytes));
HIP_ASSERT(A_h != NULL);
B_h = static_cast<int*>(malloc(Nbytes));
HIP_ASSERT(B_h != NULL);
for (size_t i = 0; i < N; ++i) {
if (A_h) {
(A_h)[i] = 3.146f + i; // Pi
}
}
hipStream_t stream;
HIP_CHECK(hipStreamCreate(&stream));
HIP_CHECK(hipMemcpyWithStream(B_h, A_h, Nbytes, hipMemcpyHostToHost, stream));
HIP_CHECK(hipStreamSynchronize(stream));
for (size_t i = 0; i < N; i++) {
HIP_ASSERT(A_h[i] == B_h[i]);
}
if (A_h) {
free(A_h);
}
if (B_h) {
free(B_h);
}
HIP_CHECK(hipStreamDestroy(stream));
}
TEST_CASE("Unit_hipMemcpyWithStream_TestWithOneStream") {
TestwithOnestream();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestwithTwoStream") {
TestwithTwoStream();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestkindDtoH") {
TestkindDtoH();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestkindHtoH") {
TestkindHtoH();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestkindDtoD") {
TestkindDtoD();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestOnMultiGPUwithOneStream") {
TestOnMultiGPUwithOneStream();
}
TEST_CASE("Unit_hipMemcpyWithStream_TestkindDefault") {
TestkindDefault();
}
#ifndef __HIP_PLATFORM_NVCC__
TEST_CASE("Unit_hipMemcpyWithStream_TestkindDefaultForDtoD") {
TestkindDefaultForDtoD();
}
#endif
TEST_CASE("Unit_hipMemcpyWithStream_TestDtoDonSameDevice") {
TestDtoDonSameDevice();
}
catch/unit/memory/hipMemcpy_derivatives.cc
+119
Voir le fichier
@@ -0,0 +1,119 @@
/*
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.
*/
#include <hip_test_common.hh>
#include <hip/hip_runtime_api.h>
#include <memcpy1d_tests_common.hh>
#include <resource_guards.hh>
#include <utils.hh>
// hipMemcpyDtoH
TEST_CASE("Unit_hipMemcpyDtoH_Positive_Basic") {
MemcpyDeviceToHostShell<false>([](void* dst, void* src, size_t count) {
return hipMemcpyDtoH(dst, reinterpret_cast<hipDeviceptr_t>(src), count);
});
}
TEST_CASE("Unit_hipMemcpyDtoH_Positive_Synchronization_Behavior") {
const auto f = [](void* dst, void* src, size_t count) {
return hipMemcpyDtoH(dst, reinterpret_cast<hipDeviceptr_t>(src), count);
};
MemcpyDtoHPageableSyncBehavior(f, true);
MemcpyDtoHPinnedSyncBehavior(f, true);
}
TEST_CASE("Unit_hipMemcpyDtoH_Negative_Parameters") {
using namespace std::placeholders;
LinearAllocGuard<int> device_alloc(LinearAllocs::hipMalloc, kPageSize);
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, kPageSize);
MemcpyCommonNegativeTests(
[](void* dst, void* src, size_t count) {
return hipMemcpyDtoH(dst, reinterpret_cast<hipDeviceptr_t>(src), count);
},
host_alloc.ptr(), device_alloc.ptr(), kPageSize);
}
// hipMemcpyHtoD
TEST_CASE("Unit_hipMemcpyHtoD_Positive_Basic") {
MemcpyHostToDeviceShell<false>([](void* dst, void* src, size_t count) {
return hipMemcpyHtoD(reinterpret_cast<hipDeviceptr_t>(dst), src, count);
});
}
TEST_CASE("Unit_hipMemcpyHtoD_Positive_Synchronization_Behavior") {
MemcpyHtoDSyncBehavior(
[](void* dst, void* src, size_t count) {
return hipMemcpyHtoD(reinterpret_cast<hipDeviceptr_t>(dst), src, count);
},
true);
}
TEST_CASE("Unit_hipMemcpyHtoD_Negative_Parameters") {
using namespace std::placeholders;
LinearAllocGuard<int> device_alloc(LinearAllocs::hipMalloc, kPageSize);
LinearAllocGuard<int> host_alloc(LinearAllocs::hipHostMalloc, kPageSize);
MemcpyCommonNegativeTests(
[](void* dst, void* src, size_t count) {
return hipMemcpyHtoD(reinterpret_cast<hipDeviceptr_t>(dst), src, count);
},
device_alloc.ptr(), host_alloc.ptr(), kPageSize);
}
// hipMemcpyDtoD
TEST_CASE("Unit_hipMemcpyDtoD_Positive_Basic") {
const auto f = [](void* dst, void* src, size_t count) {
return hipMemcpyDtoD(reinterpret_cast<hipDeviceptr_t>(dst),
reinterpret_cast<hipDeviceptr_t>(src), count);
};
SECTION("Peer access enabled") { MemcpyDeviceToDeviceShell<false, true>(f); }
SECTION("Peer access disabled") { MemcpyDeviceToDeviceShell<false, false>(f); }
}
TEST_CASE("Unit_hipMemcpyDtoD_Positive_Synchronization_Behavior") {
// This behavior differs on NVIDIA and AMD, on AMD the hipMemcpy calls is synchronous with
// respect to the host
#if HT_AMD
HipTest::HIP_SKIP_TEST(
"EXSWCPHIPT-127 - Memcpy from device to device memory behavior differs on AMD and Nvidia");
return;
#endif
MemcpyDtoDSyncBehavior(
[](void* dst, void* src, size_t count) {
return hipMemcpyDtoD(reinterpret_cast<hipDeviceptr_t>(dst),
reinterpret_cast<hipDeviceptr_t>(src), count);
},
false);
}
TEST_CASE("Unit_hipMemcpyDtoD_Negative_Parameters") {
using namespace std::placeholders;
LinearAllocGuard<int> src_alloc(LinearAllocs::hipMalloc, kPageSize);
LinearAllocGuard<int> dst_alloc(LinearAllocs::hipMalloc, kPageSize);
MemcpyCommonNegativeTests(
[](void* dst, void* src, size_t count) {
return hipMemcpyDtoD(reinterpret_cast<hipDeviceptr_t>(dst),
reinterpret_cast<hipDeviceptr_t>(src), count);
},
dst_alloc.ptr(), src_alloc.ptr(), kPageSize);
}
catch/unit/memory/hipMemcpy_old.cc
+619
Voir le fichier
@@ -0,0 +1,619 @@
/*
Copyright (c) 2021 - present 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 testcase verifies following scenarios
1. hipMemcpy API along with kernel launch with different data types
2. H2D-D2D-D2H scenarios for unpinned and pinned memory
3. Boundary checks with different sizes
4. Multithread scenario
5. device offset scenario
*/
#include <hip_test_common.hh>
#include <hip_test_kernels.hh>
#include <hip_test_checkers.hh>
#ifdef _WIN32
#define WIN32_LEAN_AND_MEAN
#include <windows.h>
#else
#include "sys/types.h"
#include "sys/sysinfo.h"
#endif
static constexpr auto NUM_ELM{4*1024 * 1024};
static unsigned blocksPerCU{6}; // to hide latency
static unsigned threadsPerBlock{256};
template<typename T>
class DeviceMemory {
public:
explicit DeviceMemory(size_t numElements);
DeviceMemory() = delete;
~DeviceMemory();
T* A_d() const { return _A_d + _offset; }
T* B_d() const { return _B_d + _offset; }
T* C_d() const { return _C_d + _offset; }
T* C_dd() const { return _C_dd + _offset; }
size_t maxNumElements() const { return _maxNumElements; }
void offset(int offset) { _offset = offset; }
int offset() const { return _offset; }
private:
T* _A_d;
T* _B_d;
T* _C_d;
T* _C_dd;
size_t _maxNumElements;
int _offset;
};
template <typename T>
DeviceMemory<T>::DeviceMemory(size_t numElements) :
_maxNumElements(numElements), _offset(0) {
T** np = nullptr;
HipTest::initArrays(&_A_d, &_B_d, &_C_d, np, np, np, numElements, 0);
size_t sizeElements = numElements * sizeof(T);
HIP_CHECK(hipMalloc(&_C_dd, sizeElements));
}
template <typename T>
DeviceMemory<T>::~DeviceMemory() {
T* np = nullptr;
HipTest::freeArrays<T>(_A_d, _B_d, _C_d, np, np, np, 0);
HIP_CHECK(hipFree(_C_dd));
_C_dd = NULL;
}
template <typename T>
class HostMemory {
public:
HostMemory(size_t numElements, bool usePinnedHost);
HostMemory() = delete;
void reset(size_t numElements, bool full = false);
~HostMemory();
T* A_h() const { return _A_h + _offset; }
T* B_h() const { return _B_h + _offset; }
T* C_h() const { return _C_h + _offset; }
size_t maxNumElements() const { return _maxNumElements; }
void offset(int offset) { _offset = offset; }
int offset() const { return _offset; }
// Host arrays, secondary copy
T* A_hh;
T* B_hh;
bool _usePinnedHost;
private:
size_t _maxNumElements;
int _offset;
// Host arrays
T* _A_h;
T* _B_h;
T* _C_h;
};
template <typename T>
HostMemory<T>::HostMemory(size_t numElements, bool usePinnedHost)
: _usePinnedHost(usePinnedHost), _maxNumElements(numElements), _offset(0) {
T** np = nullptr;
HipTest::initArrays(np, np, np, &_A_h, &_B_h, &_C_h,
numElements, usePinnedHost);
A_hh = NULL;
B_hh = NULL;
size_t sizeElements = numElements * sizeof(T);
if (usePinnedHost) {
HIP_CHECK(hipHostMalloc(reinterpret_cast<void**>(&A_hh), sizeElements,
hipHostMallocDefault));
HIP_CHECK(hipHostMalloc(reinterpret_cast<void**>(&B_hh), sizeElements,
hipHostMallocDefault));
} else {
A_hh = reinterpret_cast<T*>(malloc(sizeElements));
B_hh = reinterpret_cast<T*>(malloc(sizeElements));
}
}
template <typename T>
void HostMemory<T>::reset(size_t numElements, bool full) {
// Initialize the host data:
for (size_t i = 0; i < numElements; i++) {
(A_hh)[i] = 1097.0 + i;
(B_hh)[i] = 1492.0 + i; // Phi
if (full) {
(_A_h)[i] = 3.146f + i; // Pi
(_B_h)[i] = 1.618f + i; // Phi
}
}
}
template <typename T>
HostMemory<T>::~HostMemory() {
HipTest::freeArraysForHost(_A_h, _B_h, _C_h, _usePinnedHost);
if (_usePinnedHost) {
HIP_CHECK(hipHostFree(A_hh));
HIP_CHECK(hipHostFree(B_hh));
} else {
free(A_hh);
free(B_hh);
}
}
#ifdef _WIN32
void memcpytest2_get_host_memory(size_t *free, size_t *total) {
MEMORYSTATUSEX status;
status.dwLength = sizeof(status);
GlobalMemoryStatusEx(&status);
// Windows doesn't allow allocating more than half of system memory to the gpu
// Since the runtime also needs space for its internal allocations,
// we should not try to allocate more than 40% of reported system memory,
// otherwise we can run into OOM issues.
*free = static_cast<size_t>(0.4 * status.ullAvailPhys);
*total = static_cast<size_t>(0.4 * status.ullTotalPhys);
}
#else
struct sysinfo memInfo;
void memcpytest2_get_host_memory(size_t *free, size_t *total) {
sysinfo(&memInfo);
uint64_t freePhysMem = memInfo.freeram;
freePhysMem *= memInfo.mem_unit;
*free = freePhysMem;
uint64_t totalPhysMem = memInfo.totalram;
totalPhysMem *= memInfo.mem_unit;
*total = totalPhysMem;
}
#endif
//---
// Test many different kinds of memory copies.
// The subroutine allocates memory , copies to device, runs a vector
// add kernel, copies back, and
// checks the result.
//
// IN: numElements controls the number of elements used for allocations.
// IN: usePinnedHost : If true, allocate host with hipHostMalloc and is pinned
// else allocate host
// memory with malloc. IN: useHostToHost : If true, add an extra
// host-to-host copy. IN:
// useDeviceToDevice : If true, add an extra deviceto-device copy after
// result is produced. IN:
// useMemkindDefault : If true, use memkinddefault
// (runtime figures out direction). if false, use
// explicit memcpy direction.
//
template <typename T>
void memcpytest2(DeviceMemory<T>* dmem, HostMemory<T>* hmem,
size_t numElements, bool useHostToHost,
bool useDeviceToDevice, bool useMemkindDefault) {
size_t sizeElements = numElements * sizeof(T);
hmem->reset(numElements);
assert(numElements <= dmem->maxNumElements());
assert(numElements <= hmem->maxNumElements());
if (useHostToHost) {
// Do some extra host-to-host copies here to mix things up:
HIP_CHECK(hipMemcpy(hmem->A_hh, hmem->A_h(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToHost));
HIP_CHECK(hipMemcpy(hmem->B_hh, hmem->B_h(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToHost));
HIP_CHECK(hipMemcpy(dmem->A_d(), hmem->A_hh, sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy(dmem->B_d(), hmem->B_hh, sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice));
} else {
HIP_CHECK(hipMemcpy(dmem->A_d(), hmem->A_h(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy(dmem->B_d(), hmem->B_h(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyHostToDevice));
}
hipLaunchKernelGGL(HipTest::vectorADD, dim3(1), dim3(1), 0, 0,
static_cast<const T*>(dmem->A_d()), static_cast<const T*>(dmem->B_d()),
dmem->C_d(), numElements);
HIP_CHECK(hipGetLastError());
if (useDeviceToDevice) {
// Do an extra device-to-device copy here to mix things up:
HIP_CHECK(hipMemcpy(dmem->C_dd(), dmem->C_d(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyDeviceToDevice));
// Destroy the original dmem->C_d():
HIP_CHECK(hipMemset(dmem->C_d(), 0x5A, sizeElements));
HIP_CHECK(hipMemcpy(hmem->C_h(), dmem->C_dd(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyDeviceToHost));
} else {
HIP_CHECK(hipMemcpy(hmem->C_h(), dmem->C_d(), sizeElements,
useMemkindDefault ? hipMemcpyDefault : hipMemcpyDeviceToHost));
}
HIP_CHECK(hipDeviceSynchronize());
HipTest::checkVectorADD(hmem->A_h(), hmem->B_h(), hmem->C_h(), numElements);
printf(" %s success\n", __func__);
}
// Try all the 16 possible combinations to memcpytest2 - usePinnedHost,
// useHostToHost,
// useDeviceToDevice, useMemkindDefault
template <typename T>
void memcpytest2_for_type(size_t numElements) {
DeviceMemory<T> memD(numElements);
HostMemory<T> memU(numElements, 0 /*usePinnedHost*/);
HostMemory<T> memP(numElements, 1 /*usePinnedHost*/);
for (int usePinnedHost = 0; usePinnedHost <= 1; usePinnedHost++) {
for (int useHostToHost = 0; useHostToHost <= 1; useHostToHost++) {
for (int useDeviceToDevice = 0; useDeviceToDevice <= 1;
useDeviceToDevice++) {
for (int useMemkindDefault = 0; useMemkindDefault <= 1;
useMemkindDefault++) {
memcpytest2<T>(&memD, usePinnedHost ? &memP : &memU,
numElements, useHostToHost,
useDeviceToDevice, useMemkindDefault);
}
}
}
}
}
// Try many different sizes to memory copy.
template <typename T>
void memcpytest2_sizes(size_t maxElem = 0) {
int deviceId;
HIP_CHECK(hipGetDevice(&deviceId));
size_t free, total, freeCPU, totalCPU;
HIP_CHECK(hipMemGetInfo(&free, &total));
memcpytest2_get_host_memory(&freeCPU, &totalCPU);
if (maxElem == 0) {
// Use lesser maxElem if not enough host memory available
size_t maxElemGPU = free / sizeof(T) / 8;
size_t maxElemCPU = freeCPU / sizeof(T) / 8;
maxElem = maxElemGPU < maxElemCPU ? maxElemGPU : maxElemCPU;
}
HIP_CHECK(hipDeviceReset());
DeviceMemory<T> memD(maxElem);
HostMemory<T> memU(maxElem, 0 /*usePinnedHost*/);
HostMemory<T> memP(maxElem, 1 /*usePinnedHost*/);
for (size_t elem = 1; elem <= maxElem; elem *= 2) {
memcpytest2<T>(&memD, &memU, elem, 1, 1, 0); // unpinned host
memcpytest2<T>(&memD, &memP, elem, 1, 1, 0); // pinned host
}
}
// Try many different sizes to memory copy.
template <typename T>
void memcpytest2_offsets(size_t maxElem, bool devOffsets, bool hostOffsets) {
int deviceId;
HIP_CHECK(hipGetDevice(&deviceId));
size_t free, total;
HIP_CHECK(hipMemGetInfo(&free, &total));
HIP_CHECK(hipDeviceReset());
DeviceMemory<T> memD(maxElem);
HostMemory<T> memU(maxElem, 0 /*usePinnedHost*/);
HostMemory<T> memP(maxElem, 1 /*usePinnedHost*/);
size_t elem = maxElem / 2;
for (size_t offset = 0; offset < 512; offset++) {
assert(elem + offset < maxElem);
if (devOffsets) {
memD.offset(offset);
}
if (hostOffsets) {
memU.offset(offset);
memP.offset(offset);
}
memcpytest2<T>(&memD, &memU, elem, 1, 1, 0); // unpinned host
memcpytest2<T>(&memD, &memP, elem, 1, 1, 0); // pinned host
}
for (size_t offset = 512; offset < elem; offset *= 2) {
assert(elem + offset < maxElem);
if (devOffsets) {
memD.offset(offset);
}
if (hostOffsets) {
memU.offset(offset);
memP.offset(offset);
}
memcpytest2<T>(&memD, &memU, elem, 1, 1, 0); // unpinned host
memcpytest2<T>(&memD, &memP, elem, 1, 1, 0); // pinned host
}
}
// Create multiple threads to stress multi-thread locking behavior in the
// allocation/deallocation/tracking logic:
template <typename T>
void multiThread_1(bool serialize, bool usePinnedHost) {
DeviceMemory<T> memD(NUM_ELM);
HostMemory<T> mem1(NUM_ELM, usePinnedHost);
HostMemory<T> mem2(NUM_ELM, usePinnedHost);
std::thread t1(memcpytest2<T>, &memD, &mem1, NUM_ELM, 0, 0, 0);
if (serialize) {
t1.join();
}
std::thread t2(memcpytest2<T>, &memD, &mem2, NUM_ELM, 0, 0, 0);
if (serialize) {
t2.join();
}
}
/*
This testcase verifies hipMemcpy API
Initializes device variables
Launches kernel and performs the sum of device variables
copies the result to host variable and validates the result.
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy_KernelLaunch", "", int, float,
double) {
size_t Nbytes = NUM_ELM * sizeof(TestType);
TestType *A_d{nullptr}, *B_d{nullptr}, *C_d{nullptr};
TestType *A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr};
HipTest::initArrays(&A_d, &B_d, &C_d, &A_h, &B_h, &C_h, NUM_ELM, false);
HIP_CHECK(hipMemcpy(A_d, A_h, Nbytes, hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy(B_d, B_h, Nbytes, hipMemcpyHostToDevice));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(1), dim3(1), 0, 0,
static_cast<const TestType*>(A_d),
static_cast<const TestType*>(B_d), C_d, NUM_ELM);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipMemcpy(C_h, C_d, Nbytes, hipMemcpyDeviceToHost));
HIP_CHECK(hipDeviceSynchronize());
HipTest::checkVectorADD(A_h, B_h, C_h, NUM_ELM);
HipTest::freeArrays<TestType>(A_d, B_d, C_d, A_h, B_h, C_h, false);
}
/*
This testcase verifies the following scenarios
1. H2H,H2PinMem and PinnedMem2Host
2. H2D-D2D-D2H in same GPU
3. Pinned Host Memory to device variables in same GPU
4. Device context change
5. H2D-D2D-D2H peer GPU
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy_H2H-H2D-D2H-H2PinMem", "", int,
float, double) {
TestType *A_d{nullptr}, *B_d{nullptr};
TestType *A_h{nullptr}, *B_h{nullptr};
TestType *A_Ph{nullptr}, *B_Ph{nullptr};
HIP_CHECK(hipSetDevice(0));
HipTest::initArrays<TestType>(&A_d, &B_d, nullptr,
&A_h, &B_h, nullptr,
NUM_ELM*sizeof(TestType));
HipTest::initArrays<TestType>(nullptr, nullptr, nullptr,
&A_Ph, &B_Ph, nullptr,
NUM_ELM*sizeof(TestType), true);
SECTION("H2H, H2PinMem and PinMem2H") {
HIP_CHECK(hipMemcpy(B_h, A_h, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(A_Ph, B_h, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_Ph, A_Ph, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HipTest::checkTest(A_h, B_Ph, NUM_ELM);
}
SECTION("H2D-D2D-D2H-SameGPU") {
HIP_CHECK(hipMemcpy(A_d, A_h, NUM_ELM*sizeof(TestType), hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_d, A_d, NUM_ELM*sizeof(TestType), hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_h, B_d, NUM_ELM*sizeof(TestType), hipMemcpyDefault));
HipTest::checkTest(A_h, B_h, NUM_ELM);
}
SECTION("pH2D-D2D-D2pH-SameGPU") {
HIP_CHECK(hipMemcpy(A_d, A_Ph, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_d, A_d, NUM_ELM*sizeof(TestType), hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_Ph, B_d, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HipTest::checkTest(A_Ph, B_Ph, NUM_ELM);
}
SECTION("H2D-D2D-D2H-DeviceContextChange") {
int deviceCount = 0;
HIP_CHECK(hipGetDeviceCount(&deviceCount));
if (deviceCount < 2) {
SUCCEED("deviceCount less then 2");
} else {
int canAccessPeer = 0;
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
if (canAccessPeer) {
HIP_CHECK(hipSetDevice(1));
HIP_CHECK(hipMemcpy(A_d, A_h, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_d, A_d, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_h, B_d, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HipTest::checkTest(A_h, B_h, NUM_ELM);
} else {
SUCCEED("P2P capability is not present");
}
}
}
SECTION("H2D-D2D-D2H-PeerGPU") {
int deviceCount = 0;
HIP_CHECK(hipGetDeviceCount(&deviceCount));
if (deviceCount < 2) {
SUCCEED("deviceCount less then 2");
} else {
int canAccessPeer = 0;
HIP_CHECK(hipDeviceCanAccessPeer(&canAccessPeer, 0, 1));
if (canAccessPeer) {
HIP_CHECK(hipSetDevice(1));
TestType *C_d{nullptr};
HipTest::initArrays<TestType>(nullptr, nullptr, &C_d,
nullptr, nullptr, nullptr,
NUM_ELM*sizeof(TestType));
HIP_CHECK(hipMemcpy(A_d, A_h, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(C_d, A_d, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HIP_CHECK(hipMemcpy(B_h, C_d, NUM_ELM*sizeof(TestType),
hipMemcpyDefault));
HipTest::checkTest(A_h, B_h, NUM_ELM);
HIP_CHECK(hipFree(C_d));
} else {
SUCCEED("P2P capability is not present");
}
}
}
HipTest::freeArrays<TestType>(A_d, B_d, nullptr, A_h, B_h, nullptr, false);
HipTest::freeArrays<TestType>(nullptr, nullptr, nullptr, A_Ph,
B_Ph, nullptr, true);
}
/*
This testcase verifies the multi thread scenario
*/
TEST_CASE("Unit_hipMemcpy_MultiThreadWithSerialization") {
HIP_CHECK(hipDeviceReset());
// Simplest cases: serialize the threads, and also used pinned memory:
// This verifies that the sub-calls to memcpytest2 are correct.
multiThread_1<float>(true, true);
// Serialize, but use unpinned memory to stress the unpinned memory xfer path.
multiThread_1<float>(true, false);
}
/*
This testcase verifies hipMemcpy API with pinnedMemory and hostRegister
along with kernel launches
*/
TEMPLATE_TEST_CASE("Unit_hipMemcpy_PinnedRegMemWithKernelLaunch",
"", int, float, double) {
int numDevices = 0;
HIP_CHECK(hipGetDeviceCount(&numDevices));
if (numDevices < 2) {
SUCCEED("No of devices are less than 2");
} else {
// 1 refers to pinned Memory
// 2 refers to register Memory
int MallocPinType = GENERATE(0, 1);
size_t Nbytes = NUM_ELM * sizeof(TestType);
unsigned blocks = HipTest::setNumBlocks(blocksPerCU,
threadsPerBlock, NUM_ELM);
TestType *A_d{nullptr}, *B_d{nullptr}, *C_d{nullptr};
TestType *X_d{nullptr}, *Y_d{nullptr}, *Z_d{nullptr};
TestType *A_h{nullptr}, *B_h{nullptr}, *C_h{nullptr};
if (MallocPinType) {
HipTest::initArrays(&A_d, &B_d, &C_d, &A_h, &B_h, &C_h, NUM_ELM, true);
} else {
A_h = reinterpret_cast<TestType*>(malloc(Nbytes));
HIP_CHECK(hipHostRegister(A_h, Nbytes, hipHostRegisterDefault));
B_h = reinterpret_cast<TestType*>(malloc(Nbytes));
HIP_CHECK(hipHostRegister(B_h, Nbytes, hipHostRegisterDefault));
C_h = reinterpret_cast<TestType*>(malloc(Nbytes));
HIP_CHECK(hipHostRegister(C_h, Nbytes, hipHostRegisterDefault));
HipTest::initArrays<TestType>(&A_d, &B_d, &C_d, nullptr, nullptr,
nullptr, NUM_ELM, false);
HipTest::setDefaultData<TestType>(NUM_ELM, A_h, B_h, C_h);
}
HIP_CHECK(hipMemcpy(A_d, A_h, Nbytes, hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy(B_d, B_h, Nbytes, hipMemcpyHostToDevice));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, 0, static_cast<const TestType*>(A_d),
static_cast<const TestType*>(B_d), C_d, NUM_ELM);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipMemcpy(C_h, C_d, Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h, B_h, C_h, NUM_ELM);
unsigned int seed = time(0);
HIP_CHECK(hipSetDevice(HipTest::RAND_R(&seed) % (numDevices-1)+1));
int device;
HIP_CHECK(hipGetDevice(&device));
std::cout <<"hipMemcpy is set to happen between device 0 and device "
<<device << std::endl;
HipTest::initArrays<TestType>(&X_d, &Y_d, &Z_d, nullptr,
nullptr, nullptr, NUM_ELM, false);
for (int j = 0; j < NUM_ELM; j++) {
A_h[j] = 0;
B_h[j] = 0;
C_h[j] = 0;
}
HIP_CHECK(hipMemcpy(A_h, A_d, Nbytes, hipMemcpyDeviceToHost));
HIP_CHECK(hipMemcpy(X_d, A_h, Nbytes, hipMemcpyHostToDevice));
HIP_CHECK(hipMemcpy(B_h, B_d, Nbytes, hipMemcpyDeviceToHost));
HIP_CHECK(hipMemcpy(Y_d, B_h, Nbytes, hipMemcpyHostToDevice));
hipLaunchKernelGGL(HipTest::vectorADD, dim3(blocks), dim3(threadsPerBlock),
0, 0, static_cast<const TestType*>(X_d),
static_cast<const TestType*>(Y_d), Z_d, NUM_ELM);
HIP_CHECK(hipGetLastError());
HIP_CHECK(hipMemcpy(C_h, Z_d, Nbytes, hipMemcpyDeviceToHost));
HipTest::checkVectorADD(A_h, B_h, C_h, NUM_ELM);
if (MallocPinType) {
HipTest::freeArrays<TestType>(A_d, B_d, C_d, A_h, B_h, C_h, true);
} else {
HIP_CHECK(hipHostUnregister(A_h));
free(A_h);
HIP_CHECK(hipHostUnregister(B_h));
free(B_h);
HIP_CHECK(hipHostUnregister(C_h));
free(C_h);
HipTest::freeArrays<TestType>(A_d, B_d, C_d, nullptr,
nullptr, nullptr, false);
}
HipTest::freeArrays<TestType>(X_d, Y_d, Z_d, nullptr,
nullptr, nullptr, false);
}
}