Skip to content

[SYCL][NATIVECPU] Implement events on Native CPU #15926

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.

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

Merged
merged 8 commits into from
Nov 13, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
12 changes: 6 additions & 6 deletions sycl/cmake/modules/UnifiedRuntimeTag.cmake
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
# commit 9a209aa5fe6f438b682b3a999b9ee505e202f9b7
# Merge: 2eae687a 9339e374
# commit cd92e72bbc4ebddef63c63c0f7e66a410f4b9552
# Merge: 9a209aa5 b1222f08
# Author: Callum Fare <[email protected]>
# Date: Tue Nov 12 11:47:03 2024 +0000
# Merge pull request #2179 from Maetveis/wrap_icx_linker_flags
# Wrap linker flags on Windows for IntelLLLVM
set(UNIFIED_RUNTIME_TAG 9a209aa5fe6f438b682b3a999b9ee505e202f9b7)
# Date: Wed Nov 13 09:57:16 2024 +0000
# Merge pull request #2254 from PietroGhg/pietro/events_rr
# [NATIVECPU] Implement events on Native CPU
set(UNIFIED_RUNTIME_TAG cd92e72bbc4ebddef63c63c0f7e66a410f4b9552)
75 changes: 49 additions & 26 deletions sycl/test/native_cpu/scalar_args.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -12,39 +12,62 @@ const size_t N = 10;

template <typename T> class init_a;

template <typename T> bool test(queue myQueue) {
{
buffer<float, 1> a(range<1>{N});
const T test = rand() % 10;

myQueue.submit([&](handler &cgh) {
auto A = a.get_access<access::mode::write>(cgh);
cgh.parallel_for<init_a<T>>(range<1>{N},
[=](id<1> index) { A[index] = test; });
});

auto A = a.get_access<access::mode::read>();
std::cout << "Result:" << std::endl;
for (size_t i = 0; i < N; i++) {
if (A[i] != test) {
std::cout << "ERROR at pos " << i << " expected " << test << ", got "
<< A[i] << "\n";
return false;
}
}
template <typename T, int M>
bool compare(const sycl::vec<T, M> a, const sycl::vec<T, M> truth) {
bool res = true;
for (int i = 0; i < M; i++) {
res &= a[i] == truth[i];
}
return res;
}

std::cout << "Good computation!" << std::endl;
template <typename T> bool compare(const T a, const T truth) {
return a == truth;
}

template <typename T> bool check(buffer<T, 1> &result, const T truth) {
auto A = result.get_host_access();
for (size_t i = 0; i < N; i++) {
if (!compare(A[i], truth)) {
return false;
}
}
return true;
}

template <typename T> bool test(queue myQueue) {
buffer<T, 1> a(range<1>{N});
const T test{42};

myQueue.submit([&](handler &cgh) {
auto A = a.template get_access<access::mode::write>(cgh);
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It might be worth replicating this test also for access::mode::read_write to see if timing/events handling is correct for different access modes.

cgh.parallel_for<init_a<T>>(range<1>{N},
[=](id<1> index) { A[index] = test; });
});

return check(a, test);
}

int main() {
queue q;
int res1 = test<int>(q);
int res2 = test<unsigned>(q);
int res3 = test<float>(q);
int res4 = test<double>(q);
if (!(res1 && res2 && res3 && res4)) {

std::vector<bool> res;
res.push_back(test<int>(q));
res.push_back(test<unsigned>(q));
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Performance. Using a vector here likely causes allocations and there's the overhead of iterating over the vector.
It's probably faster to instead check the test values and return straight away if there's an error.

res.push_back(test<float>(q));
res.push_back(test<double>(q));
res.push_back(test<sycl::vec<int, 2>>(q));
res.push_back(test<sycl::vec<int, 3>>(q));
res.push_back(test<sycl::vec<int, 4>>(q));
res.push_back(test<sycl::vec<int, 8>>(q));
res.push_back(test<sycl::vec<int, 16>>(q));
res.push_back(test<sycl::vec<double, 2>>(q));
res.push_back(test<sycl::vec<double, 3>>(q));
res.push_back(test<sycl::vec<double, 4>>(q));
res.push_back(test<sycl::vec<double, 8>>(q));
res.push_back(test<sycl::vec<double, 16>>(q));

if (std::any_of(res.begin(), res.end(), [](bool b) { return !b; })) {
return 1;
}
return 0;
Expand Down
Loading