Skip to content

[curand] Use host_task rather than codeplay_host_task #137

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

Merged
mkrainiuk merged 1 commit into from
Oct 22, 2021
Merged

[curand] Use host_task rather than codeplay_host_task #137

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
40 changes: 20 additions & 20 deletions src/rng/backends/curand/mrg32k3a.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -100,7 +100,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
queue_
.submit([&](cl::sycl::handler& cgh) {
auto acc = r.get_access<cl::sycl::access::mode::read_write>(cgh);
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
auto r_ptr =
reinterpret_cast<float*>(ih.get_native_mem<cl::sycl::backend::cuda>(acc));
curandStatus_t status;
Expand All @@ -117,7 +117,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
queue_
.submit([&](cl::sycl::handler& cgh) {
auto acc = r.get_access<cl::sycl::access::mode::read_write>(cgh);
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
auto r_ptr =
reinterpret_cast<double*>(ih.get_native_mem<cl::sycl::backend::cuda>(acc));
curandStatus_t status;
Expand All @@ -135,7 +135,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
queue_
.submit([&](cl::sycl::handler& cgh) {
auto acc = ib.get_access<cl::sycl::access::mode::read_write>(cgh);
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
auto ib_ptr = reinterpret_cast<std::uint32_t*>(
ih.get_native_mem<cl::sycl::backend::cuda>(acc));
curandStatus_t status;
Expand All @@ -152,7 +152,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
queue_
.submit([&](cl::sycl::handler& cgh) {
auto acc = r.get_access<cl::sycl::access::mode::read_write>(cgh);
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
auto r_ptr =
reinterpret_cast<float*>(ih.get_native_mem<cl::sycl::backend::cuda>(acc));
curandStatus_t status;
Expand All @@ -169,7 +169,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
queue_
.submit([&](cl::sycl::handler& cgh) {
auto acc = r.get_access<cl::sycl::access::mode::read_write>(cgh);
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
auto r_ptr =
reinterpret_cast<double*>(ih.get_native_mem<cl::sycl::backend::cuda>(acc));
curandStatus_t status;
Expand All @@ -186,7 +186,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
queue_
.submit([&](cl::sycl::handler& cgh) {
auto acc = r.get_access<cl::sycl::access::mode::read_write>(cgh);
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
auto r_ptr =
reinterpret_cast<float*>(ih.get_native_mem<cl::sycl::backend::cuda>(acc));
curandStatus_t status;
Expand All @@ -203,7 +203,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
queue_
.submit([&](cl::sycl::handler& cgh) {
auto acc = r.get_access<cl::sycl::access::mode::read_write>(cgh);
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
auto r_ptr =
reinterpret_cast<double*>(ih.get_native_mem<cl::sycl::backend::cuda>(acc));
curandStatus_t status;
Expand Down Expand Up @@ -236,7 +236,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
queue_
.submit([&](cl::sycl::handler& cgh) {
auto acc = r.get_access<cl::sycl::access::mode::read_write>(cgh);
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
auto r_ptr =
reinterpret_cast<float*>(ih.get_native_mem<cl::sycl::backend::cuda>(acc));
curandStatus_t status;
Expand All @@ -253,7 +253,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
queue_
.submit([&](cl::sycl::handler& cgh) {
auto acc = r.get_access<cl::sycl::access::mode::read_write>(cgh);
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
auto r_ptr =
reinterpret_cast<double*>(ih.get_native_mem<cl::sycl::backend::cuda>(acc));
curandStatus_t status;
Expand Down Expand Up @@ -313,7 +313,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
queue_
.submit([&](cl::sycl::handler& cgh) {
auto acc = r.template get_access<cl::sycl::access::mode::read_write>(cgh);
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
auto r_ptr = reinterpret_cast<std::uint32_t*>(
ih.get_native_mem<cl::sycl::backend::cuda>(acc));
curandStatus_t status;
Expand All @@ -331,7 +331,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
cl::sycl::event::wait_and_throw(dependencies);
queue_
.submit([&](cl::sycl::handler& cgh) {
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
curandStatus_t status;
CURAND_CALL(curandGenerateUniform, status, engine_, r, n);
});
Expand All @@ -346,7 +346,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
cl::sycl::event::wait_and_throw(dependencies);
queue_
.submit([&](cl::sycl::handler& cgh) {
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
curandStatus_t status;
CURAND_CALL(curandGenerateUniformDouble, status, engine_, r, n);
});
Expand All @@ -364,7 +364,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
n * sizeof(std::uint32_t), queue_.get_device(), queue_.get_context());
queue_
.submit([&](cl::sycl::handler& cgh) {
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
curandStatus_t status;
CURAND_CALL(curandGenerate, status, engine_, ib, n);
});
Expand All @@ -379,7 +379,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
cl::sycl::event::wait_and_throw(dependencies);
queue_
.submit([&](cl::sycl::handler& cgh) {
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
curandStatus_t status;
CURAND_CALL(curandGenerateUniform, status, engine_, r, n);
});
Expand All @@ -394,7 +394,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
cl::sycl::event::wait_and_throw(dependencies);
queue_
.submit([&](cl::sycl::handler& cgh) {
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
curandStatus_t status;
CURAND_CALL(curandGenerateUniformDouble, status, engine_, r, n);
});
Expand All @@ -409,7 +409,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
std::int64_t n, float* r, const std::vector<cl::sycl::event>& dependencies) override {
cl::sycl::event::wait_and_throw(dependencies);
return queue_.submit([&](cl::sycl::handler& cgh) {
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
curandStatus_t status;
CURAND_CALL(curandGenerateNormal, status, engine_, r, n, distr.mean(),
distr.stddev());
Expand All @@ -423,7 +423,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
std::int64_t n, double* r, const std::vector<cl::sycl::event>& dependencies) override {
cl::sycl::event::wait_and_throw(dependencies);
return queue_.submit([&](cl::sycl::handler& cgh) {
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
curandStatus_t status;
CURAND_CALL(curandGenerateNormalDouble, status, engine_, r, n, distr.mean(),
distr.stddev());
Expand Down Expand Up @@ -455,7 +455,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
std::int64_t n, float* r, const std::vector<cl::sycl::event>& dependencies) override {
cl::sycl::event::wait_and_throw(dependencies);
return queue_.submit([&](cl::sycl::handler& cgh) {
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
curandStatus_t status;
CURAND_CALL(curandGenerateLogNormal, status, engine_, r, n, distr.m(), distr.s());
});
Expand All @@ -468,7 +468,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
std::int64_t n, double* r, const std::vector<cl::sycl::event>& dependencies) override {
cl::sycl::event::wait_and_throw(dependencies);
return queue_.submit([&](cl::sycl::handler& cgh) {
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
curandStatus_t status;
CURAND_CALL(curandGenerateLogNormalDouble, status, engine_, r, n, distr.m(),
distr.s());
Expand Down Expand Up @@ -535,7 +535,7 @@ class mrg32k3a_impl : public oneapi::mkl::rng::detail::engine_impl {
const std::vector<cl::sycl::event>& dependencies) override {
cl::sycl::event::wait_and_throw(dependencies);
return queue_.submit([&](cl::sycl::handler& cgh) {
cgh.codeplay_host_task([=](cl::sycl::interop_handle ih) {
cgh.host_task([=](cl::sycl::interop_handle ih) {
curandStatus_t status;
CURAND_CALL(curandGenerate, status, engine_, r, n);
});
Expand Down
Loading