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

FillRandom: Use MKL host API #3536

Merged
merged 2 commits into from
Nov 1, 2023
Merged

FillRandom: Use MKL host API #3536

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
3a01ff1
FillRandom: Use MKL host API
WeiqunZhang Aug 25, 2023
f0cdd0b
Merge branch 'development' into mkl_rng
WeiqunZhang Oct 23, 2023
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: 14 additions & 26 deletions Src/Base/AMReX_Random.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@ namespace
namespace amrex {
#ifdef AMREX_USE_SYCL
sycl_rng_descr* rand_engine_descr = nullptr;
//xxxxx oneapi::mkl::rng::philox4x32x10* gpu_rand_generator = nullptr;
oneapi::mkl::rng::philox4x32x10* gpu_rand_generator = nullptr;
#else
amrex::randState_t* gpu_rand_state = nullptr;
amrex::randGenerator_t gpu_rand_generator = nullptr;
Expand All @@ -44,8 +44,8 @@ void ResizeRandomSeed (amrex::ULong gpu_seed)
rand_engine_descr = new sycl_rng_descr
(Gpu::Device::streamQueue(), sycl::range<1>(N), gpu_seed, 1);

//xxxxx gpu_rand_generator = new std::remove_pointer_t<decltype(gpu_rand_generator)>
// (Gpu::Device::streamQueue(), gpu_seed+1234ULL);
gpu_rand_generator = new std::remove_pointer_t<decltype(gpu_rand_generator)>
(Gpu::Device::streamQueue(), gpu_seed+1234ULL);

#elif defined(AMREX_USE_CUDA) || defined(AMREX_USE_HIP)

Expand Down Expand Up @@ -212,11 +212,11 @@ DeallocateRandomSeedDevArray ()
Gpu::streamSynchronize();
rand_engine_descr = nullptr;
}
//xxxxx if (gpu_rand_generator != nullptr) {
// delete gpu_rand_generator;
// Gpu::streamSynchronize();
// gpu_rand_generator = nullptr;
// }
if (gpu_rand_generator != nullptr) {
delete gpu_rand_generator;
Gpu::streamSynchronize();
gpu_rand_generator = nullptr;
}
#else
if (gpu_rand_state != nullptr)
{
Expand Down Expand Up @@ -258,15 +258,9 @@ void FillRandom (Real* p, Long N)

#elif defined(AMREX_USE_SYCL)

//xxxxx oneapi::mkl::rng::uniform<Real> distr;
// auto event = oneapi::mkl::rng::generate(distr, gpu_rand_generator, N, p);
// event.wait();

amrex::ParallelForRNG(N, [=] AMREX_GPU_DEVICE (Long i, RandomEngine const& eng)
{
p[i] = Random(eng);
});
Gpu::streamSynchronize();
oneapi::mkl::rng::uniform<Real> distr;
auto event = oneapi::mkl::rng::generate(distr, *gpu_rand_generator, N, p);
event.wait();

#else
std::uniform_real_distribution<Real> distribution(Real(0.0), Real(1.0));
Expand Down Expand Up @@ -299,15 +293,9 @@ void FillRandomNormal (Real* p, Long N, Real mean, Real stddev)

#elif defined(AMREX_USE_SYCL)

//xxxxx oneapi::mkl::rng::gaussian<Real> distr(mean, stddev);
// auto event = oneapi::mkl::rng::generate(distr, gpu_rand_generator, N, p);
// event.wait();

amrex::ParallelForRNG(N, [=] AMREX_GPU_DEVICE (Long i, RandomEngine const& eng)
{
p[i] = RandomNormal(mean, stddev, eng);
});
Gpu::streamSynchronize();
oneapi::mkl::rng::gaussian<Real> distr(mean, stddev);
auto event = oneapi::mkl::rng::generate(distr, *gpu_rand_generator, N, p);
event.wait();

#else

Expand Down
2 changes: 1 addition & 1 deletion Tools/CMake/AMReXSYCL.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,7 +53,7 @@ endif()
#
target_link_options( SYCL
INTERFACE
$<${_cxx_sycl}:-fsycl -fsycl-device-lib=libc,libm-fp32,libm-fp64> )
$<${_cxx_sycl}:-qmkl=sequential -fsycl -fsycl-device-lib=libc,libm-fp32,libm-fp64> )


# TODO: use $<LINK_LANG_AND_ID:> genex for CMake >=3.17
Expand Down
2 changes: 1 addition & 1 deletion Tools/GNUMake/comps/dpcpp.mak
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -123,7 +123,7 @@ ifneq ($(BL_NO_FORT),TRUE)
endif
endif

LDFLAGS += -fsycl-device-lib=libc,libm-fp32,libm-fp64
LDFLAGS += -qmkl=sequential -fsycl-device-lib=libc,libm-fp32,libm-fp64

ifdef SYCL_PARALLEL_LINK_JOBS
LDFLAGS += -fsycl-max-parallel-link-jobs=$(SYCL_PARALLEL_LINK_JOBS)
Expand Down