Files
MachineTom 5f76cb916d SWDEV-555888 - Refactor Numa code (#1191)
1. Create a set of mini numa interface.
In Linux, the interface is based on system call rather than libnuma.
In Windows, the interface can also work, but the policy class is dummy.
Different from Linux, Windows doesn't provide numactl tool or numa lib to setup numa policy, thus
the default policy is followed in Windows, that is, using the closest host numa node to allocate
pinned host memory in hipHostMalloc().
To get the closest host numa node of a GPU device, you need query the new attribute
hipDeviceAttributeHostNumaId. Then you can create a thread with CPU affinity on the numa node.
For example, reference the test in hip-tests/catch/perftests/memory/hipPerfHostNumaAllocWin.cc.

2. Remove pfnSetThreadGroupAffinity and pfnGetNumaNodeProcessorMaskEx as the functions have been exposed since Win7 and Win server 2008.

3. Other minor fixes.
2025-10-23 21:56:15 -04:00

171 строка
6.0 KiB
C++
Исходник Постоянная ссылка Ответственный История

Этот файл содержит невидимые символы Юникода
This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
/*
Copyright (c) 2023 Advanced Micro Devices, Inc. All rights reserved.
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in
all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT 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.
*/
/**
* @addtogroup hipMemcpyKernel hipMemcpyKernel
* @{
* @ingroup perfMemoryTest
* `hipMemcpy(void* dst, const void* src, size_t count, hipMemcpyKind kind)` -
* Copies data between host and device.
*/
#include <unistd.h>
#include <numaif.h>
#include <numa.h>
#include <hip_test_common.hh>
// #define ENABLE_DEBUG 1
// To run it correctly, we must not export HIP_VISIBLE_DEVICES.
// And we must explicitly link libnuma because of numa api move_pages().
#define NUM_PAGES 100
char* h = nullptr;
char* d_h = nullptr;
char* m = nullptr;
char* d_m = nullptr;
int page_size = 0;
const int mode[] = {MPOL_DEFAULT, MPOL_BIND, MPOL_PREFERRED, MPOL_INTERLEAVE};
const char* modeStr[] = {"MPOL_DEFAULT", "MPOL_BIND", "MPOL_PREFERRED", "MPOL_INTERLEAVE"};
bool test(int cpuId, int gpuId, int numaMode, unsigned int hostMallocflags) {
void* pages[NUM_PAGES];
int status[NUM_PAGES];
int ret_code;
CONSOLE_PRINT("set cpu %d, gpu %d, numaMode %d, hostMallocflags %u\n", cpuId, gpuId, numaMode,
hostMallocflags);
if (gpuId >= 0) {
HIP_CHECK(hipSetDevice(gpuId));
}
if (cpuId >= 0) {
unsigned long nodeMask = 1 << cpuId; // NOLINT
unsigned long maxNode = sizeof(nodeMask) * 8; // NOLINT
// Will override existing numa policy in memory
if (set_mempolicy(numaMode, numaMode == MPOL_DEFAULT ? NULL : &nodeMask,
numaMode == MPOL_DEFAULT ? 0 : maxNode) == -1) {
WARN("set_mempolicy() failed with err " << errno << "\n");
return false;
}
}
posix_memalign(reinterpret_cast<void**>(&m), page_size, page_size * NUM_PAGES);
HIP_CHECK(hipHostRegister(m, page_size * NUM_PAGES, hipHostRegisterMapped));
HIP_CHECK(hipHostGetDevicePointer(reinterpret_cast<void**>(&d_m), m, 0));
status[0] = -1;
pages[0] = m;
for (int i = 1; i < NUM_PAGES; i++) {
pages[i] = reinterpret_cast<char*>(pages[0]) + page_size;
}
ret_code = move_pages(0, NUM_PAGES, pages, NULL, status, 0);
CONSOLE_PRINT("Memory (malloc) ret %d at %p (dev %p) is at node: ", ret_code, m, d_m);
for (int i = 0; i < NUM_PAGES; i++) {
CONSOLE_PRINT("%d ", status[i]); // Don't verify as it's out of our control
}
CONSOLE_PRINT("\n");
HIP_CHECK(hipHostMalloc(reinterpret_cast<void**>(&h), page_size * NUM_PAGES, hostMallocflags));
pages[0] = h;
for (int i = 1; i < NUM_PAGES; i++) {
pages[i] = reinterpret_cast<char*>(pages[0]) + page_size;
}
ret_code = move_pages(0, NUM_PAGES, pages, NULL, status, 0);
d_h = nullptr;
if (hostMallocflags & hipHostMallocMapped) {
HIP_CHECK(hipHostGetDevicePointer(reinterpret_cast<void**>(&d_h), h, 0));
CONSOLE_PRINT("Memory (hipHostMalloc) ret %d at %p (dev %p) is at node: ", ret_code, h, d_h);
} else {
CONSOLE_PRINT("Memory (hipHostMalloc) ret %d at %p is at node: ", ret_code, h);
}
for (int i = 0; i < NUM_PAGES; i++) {
CONSOLE_PRINT("%d ", status[i]); // Always print it even if it's wrong. Verify later
}
CONSOLE_PRINT("\n");
HIP_CHECK(hipHostFree(reinterpret_cast<void*>(h)));
HIP_CHECK(hipHostUnregister(m));
free(m);
if (cpuId >= 0 && (numaMode == MPOL_BIND || numaMode == MPOL_PREFERRED)) {
for (int i = 0; i < NUM_PAGES; i++) {
if (status[i] != cpuId) { // Now verify
WARN("Failed at " << i << " status[i] = " << status[i] << " cpuId " << cpuId << "\n");
return false;
}
}
}
return true;
}
bool runTest(const int& cpuCount, const int& gpuCount, unsigned int hostMallocflags,
const std::string& str) {
CONSOLE_PRINT("Test- %s\n", str.c_str());
for (int m = 0; m < sizeof(mode) / sizeof(mode[0]); m++) {
CONSOLE_PRINT("Testing %s\n", modeStr[m]);
for (int i = 0; i < cpuCount; i++) {
for (int j = 0; j < gpuCount; j++) {
if (!test(i, j, mode[m], hostMallocflags)) {
return false;
}
}
}
}
return true;
}
/**
* Test Description
* ------------------------
*  - Verify hipPerfHostNumaAlloc status.
* Test source
* ------------------------
*  - perftests/memory/hipPerfHostNumaAlloc.cc
* Test requirements
* ------------------------
*  - HIP_VERSION >= 5.6
*/
TEST_CASE("Perf_hipPerfHostNumaAlloc_test") {
int gpuCount = 0;
HIP_CHECK(hipGetDeviceCount(&gpuCount));
int cpuCount = numa_max_node() + 1; // number of numa nodes
page_size = getpagesize();
CONSOLE_PRINT("Cpu count %d, Gpu count %d, page_size %d\n", cpuCount, gpuCount, page_size);
if (cpuCount < 0 || gpuCount < 0) {
SUCCEED(
"Skipped testcase hipPerfHostNumaAlloc as "
"there is no device to test.\n");
return;
}
REQUIRE(true == runTest(cpuCount, gpuCount, hipHostMallocDefault | hipHostMallocNumaUser,
"Testing hipHostMallocDefault | hipHostMallocNumaUser......"));
REQUIRE(true == runTest(cpuCount, gpuCount, hipHostMallocMapped | hipHostMallocNumaUser,
"Testing hipHostMallocMapped | hipHostMallocNumaUser......."));
}
/**
* End doxygen group perfMemoryTest.
* @}
*/