Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add timeout for distributed tests. #10315

Merged
merged 2 commits into from
May 23, 2024
Merged

Add timeout for distributed tests. #10315

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2863da1
Add timeout for distributed tests.
trivialfis May 22, 2024
0da2936
Fix.
trivialfis May 22, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion src/collective/coll.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -191,7 +191,7 @@ Result BroadcastAllgatherV(NCCLComm const* comm, common::Span<std::int8_t const>
for (std::int32_t r = 0; r < comm->World(); ++r) {
auto as_bytes = sizes[r];
auto rc = stub->Broadcast(data.data(), recv.subspan(offset, as_bytes).data(), as_bytes,
ncclInt8, r, comm->Handle(), dh::DefaultStream());
ncclInt8, r, comm->Handle(), comm->Stream());
if (!rc.OK()) {
return rc;
}
Expand Down
20 changes: 13 additions & 7 deletions tests/cpp/collective/test_worker.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -147,7 +147,8 @@ inline auto MakeDistributedTestConfig(std::string host, std::int32_t port,
}

template <typename WorkerFn>
void TestDistributedGlobal(std::int32_t n_workers, WorkerFn worker_fn, bool need_finalize = true) {
void TestDistributedGlobal(std::int32_t n_workers, WorkerFn worker_fn, bool need_finalize = true,
std::chrono::seconds test_timeout = std::chrono::seconds{30}) {
system::SocketStartup();
std::chrono::seconds timeout{1};

Expand All @@ -163,12 +164,17 @@ void TestDistributedGlobal(std::int32_t n_workers, WorkerFn worker_fn, bool need

for (std::int32_t i = 0; i < n_workers; ++i) {
workers.emplace_back([=] {
auto config = MakeDistributedTestConfig(host, port, timeout, i);
Init(config);
worker_fn();
if (need_finalize) {
Finalize();
}
auto fut = std::async(std::launch::async, [=] {
auto config = MakeDistributedTestConfig(host, port, timeout, i);
Init(config);
worker_fn();
if (need_finalize) {
Finalize();
}
});
auto status = fut.wait_for(test_timeout);
CHECK(status == std::future_status::ready) << "Test timeout";
fut.get();
});
}

Expand Down
Loading