experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Update benchmarks to newer API

Browse Source 
PiperOrigin-RevId: 356584743
Change-Id: I3c50bc0be4482171533b948b2a44c8eade403578
This commit is contained in:
A. Unique TensorFlower 2021-02-09 13:57:57 -08:00 committed by TensorFlower Gardener
parent 57cdc3ae73
commit 65f038e14d
1 changed files with 42 additions and 32 deletions

74
tensorflow/core/common_runtime/gpu/gpu_bfc_allocator_test.cc
View File

@ -333,7 +333,7 @@ TEST_P(GPUBFCAllocatorTest, DISABLED_AllocatorReceivesZeroMemory) {
INSTANTIATE_TEST_SUITE_P(GPUBFCAllocatorTestSuite, GPUBFCAllocatorTest,
TestSuiteValues());
static void BM_Allocation(int iters) {
static void BM_Allocation(::testing::benchmark::State& state) {
GPUBFCAllocator a(CreateSubAllocator(1ul << 36), 1uLL << 33, "GPU_0_bfc");
// Exercise a few different allocation sizes
std::vector<size_t> sizes = {256, 4096, 16384, 524288,
@ -341,7 +341,7 @@ static void BM_Allocation(int iters) {
1048576000, 2048576000};
int size_index = 0;
while (--iters > 0) {
for (auto s : state) {
size_t bytes = sizes[size_index++ % sizes.size()];
void* p = a.AllocateRaw(1, bytes);
a.DeallocateRaw(p);
@ -349,43 +349,53 @@ static void BM_Allocation(int iters) {
}
BENCHMARK(BM_Allocation);
static void BM_AllocationThreaded(int iters, int num_threads) {
GPUBFCAllocator a(CreateSubAllocator(1ul << 36), 1uLL << 33, "GPU_0_bfc");
thread::ThreadPool pool(Env::Default(), "test", num_threads);
std::atomic_int_fast32_t count(iters);
mutex done_lock;
condition_variable done;
bool done_flag = false;
static void BM_AllocationThreaded(::testing::benchmark::State& state) {
int num_threads = state.range(0);
int sub_iters = 500; // Pick a reasonably large number.
for (int t = 0; t < num_threads; t++) {
pool.Schedule([&a, &count, &done_lock, &done, &done_flag, iters]() {
// Exercise a few different allocation sizes
std::vector<int> sizes = {256, 4096, 16384, 524288,
512, 1048576, 10485760, 104857600};
int size_index = 0;
for (int i = 0; i < iters; i++) {
int bytes = sizes[size_index++ % sizes.size()];
void* p = a.AllocateRaw(1, bytes);
a.DeallocateRaw(p);
if (count.fetch_sub(1) == 1) {
mutex_lock l(done_lock);
done_flag = true;
done.notify_all();
break;
for (auto s : state) {
state.PauseTiming();
GPUBFCAllocator a(CreateSubAllocator(1ul << 36), 1uLL << 33, "GPU_0_bfc");
thread::ThreadPool pool(Env::Default(), "test", num_threads);
std::atomic_int_fast32_t count(sub_iters);
mutex done_lock;
condition_variable done;
bool done_flag = false;
state.ResumeTiming();
for (int t = 0; t < num_threads; t++) {
pool.Schedule([&a, &count, &done_lock, &done, &done_flag, sub_iters]() {
// Exercise a few different allocation sizes
std::vector<int> sizes = {256, 4096, 16384, 524288,
512, 1048576, 10485760, 104857600};
int size_index = 0;
for (int i = 0; i < sub_iters; i++) {
int bytes = sizes[size_index++ % sizes.size()];
void* p = a.AllocateRaw(1, bytes);
a.DeallocateRaw(p);
if (count.fetch_sub(1) == 1) {
mutex_lock l(done_lock);
done_flag = true;
done.notify_all();
break;
}
}
}
});
}
mutex_lock l(done_lock);
if (!done_flag) {
done.wait(l);
});
}
mutex_lock l(done_lock);
if (!done_flag) {
done.wait(l);
}
}
}
BENCHMARK(BM_AllocationThreaded)->Arg(1)->Arg(4)->Arg(16);
// A more complex benchmark that defers deallocation of an object for
// "delay" allocations.
static void BM_AllocationDelayed(int iters, int delay) {
static void BM_AllocationDelayed(::testing::benchmark::State& state) {
int delay = state.range(0);
GPUBFCAllocator a(CreateSubAllocator(1ull << 32), 1 << 30, "GPU_0_bfc");
// Exercise a few different allocation sizes
std::vector<int> sizes = {256, 4096, 16384, 4096, 512, 1024, 1024};
@ -397,7 +407,7 @@ static void BM_AllocationDelayed(int iters, int delay) {
ptrs.push_back(nullptr);
}
int pindex = 0;
while (--iters > 0) {
for (auto s : state) {
if (ptrs[pindex] != nullptr) {
a.DeallocateRaw(ptrs[pindex]);
ptrs[pindex] = nullptr;