/* 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 #include #include #include #include /** * Conformance test for checking functionality of * hipError_t hipGetDeviceFlags(unsigned int* flags); * hipError_t hipSetDeviceFlags(unsigned flags); * * * hipGetDeviceFlags and hipSetDeviceFlags tests. * Scenario1: Validates if hipGetDeviceFlags returns hipErrorInvalidValue for flags = nullptr. * Scenario2: Validates if hipSetDeviceFlags returns hipErrorInvalidValue for invalid flags. * Scenario3: Validates if flags returned by hipGetDeviceFlags are valid. * Scenario4: Validates that flags set with hipSetDeviceFlags can be retrieved with * hipGetDeviceFlags. * Scenario5: Validates that flags set with hipSetDeviceFlags can be retrieved on a seperate thread * with hipGetDeviceFlags. */ TEST_CASE("Unit_hipGetSetDeviceFlags_NullptrFlag") { // Scenario1 HIP_CHECK_ERROR(hipGetDeviceFlags(nullptr), hipErrorInvalidValue); } TEST_CASE("Unit_hipGetSetDeviceFlags_InvalidFlag") { #if HT_AMD HipTest::HIP_SKIP_TEST("EXSWCPHIPT-115"); return; #endif // Scenario2 const unsigned int invalidFlag = GENERATE(0b011, // schedule flags should not overlap 0b101, // schedule flags should not overlap 0b110, // schedule flags should not overlap 0b111, // schedule flags should not overlap 0b100000, // out of bounds 0xFFFF); CAPTURE(invalidFlag); HIP_CHECK_ERROR(hipSetDeviceFlags(invalidFlag), hipErrorInvalidValue); } std::array getValidFlags() { constexpr std::array scheduleFlags{hipDeviceScheduleAuto, hipDeviceScheduleSpin, hipDeviceScheduleYield, hipDeviceScheduleBlockingSync}; constexpr std::array hostMapFlags{0, hipDeviceMapHost}; constexpr std::array localMemResizeFlags{0, 0x10}; // FIXME EXSWCPHIPT-110 constexpr size_t size = scheduleFlags.size() * hostMapFlags.size() * localMemResizeFlags.size(); std::array validFlags; int i = 0; for (auto sf : scheduleFlags) { for (auto hf : hostMapFlags) { for (auto lf : localMemResizeFlags) { validFlags[i] = sf | hf | lf; i += 1; } } } return validFlags; } TEST_CASE("Unit_hipGetSetDeviceFlags_ValidFlag") { // Scenario3 auto validFlags = getValidFlags(); unsigned int flag = 0; HIP_CHECK(hipGetDeviceFlags(&flag)); REQUIRE(std::find(std::begin(validFlags), std::end(validFlags), flag) != std::end(validFlags)); } TEST_CASE("Unit_hipGetSetDeviceFlags_SetThenGet") { // Scenario4 auto validFlags = getValidFlags(); auto devNo = GENERATE(range(0, HipTest::getDeviceCount())); HIP_CHECK(hipSetDevice(devNo)); const unsigned int flag = GENERATE_COPY(from_range(std::begin(validFlags), std::end(validFlags))); HIP_CHECK(hipSetDeviceFlags(flag)); unsigned int getFlag; HIP_CHECK(hipGetDeviceFlags(&getFlag)); // flags other than hipDeviceSchedule* are ignore on the ROCm backend #if HT_NVIDIA // CUDA backend will sometimes set other flags getFlag = getFlag & hipDeviceScheduleMask; #endif REQUIRE((flag & hipDeviceScheduleMask) == getFlag); } TEST_CASE("Unit_hipGetSetDeviceFlags_Threaded") { // Scenario5 auto validFlags = getValidFlags(); auto devNo = GENERATE(range(0, HipTest::getDeviceCount())); HIP_CHECK(hipSetDevice(devNo)); std::mutex mut; std::condition_variable cv; bool ready = false; // required to avoid spurious wakeups const unsigned int flag = GENERATE_COPY(from_range(std::begin(validFlags), std::end(validFlags))); std::thread test_thread([&mut, &ready, &cv, devNo, flag]() { std::unique_lock lock(mut); cv.wait(lock, [&ready] { return ready; }); unsigned int getFlag; HIP_CHECK_THREAD(hipSetDevice(devNo)); HIP_CHECK_THREAD(hipGetDeviceFlags(&getFlag)); // flags other than hipDeviceSchedule* are ignore on the ROCm backend #if HT_NVIDIA // CUDA backend will set other flags we aren't concerned about getFlag = getFlag & hipDeviceScheduleMask; #endif REQUIRE_THREAD((flag & hipDeviceScheduleMask) == getFlag); }); { std::lock_guard lock(mut); HIP_CHECK(hipSetDeviceFlags(flag)); ready = true; } cv.notify_one(); test_thread.join(); HIP_CHECK_THREAD_FINALIZE(); }