Skip to content

Fix GPU dispatch for PolymorphicArenaAllocator in DefaultInitialization #6808

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
ax3l merged 1 commit into from
Apr 27, 2026
Merged

Fix GPU dispatch for PolymorphicArenaAllocator in DefaultInitialization #6808

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
15 changes: 10 additions & 5 deletions Source/Particles/ParticleCreation/DefaultInitialization.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -144,6 +144,11 @@ void DefaultInitializeRuntimeAttributes (PTile& ptile,
constexpr int lev = 0;
const amrex::Real t = WarpX::GetInstance().gett_new(lev);

// Check: does the particle data reside on CPU or GPU?
amrex::Arena const * const pc_arena = pc.arena();
const bool run_on_gpu =
pc_arena->isManaged() || pc_arena->isDevice();

// Initialize the last NumRuntimeRealComps() - n_external_attr_real runtime real attributes
for (int j = PIdx::nattribs + n_external_attr_real; j < ptile.NumRealComps() ; ++j)
{
Expand All @@ -158,7 +163,7 @@ void DefaultInitializeRuntimeAttributes (PTile& ptile,
const QuantumSynchrotronGetOpticalDepth quantum_sync_get_opt =
p_qs_engine->build_optical_depth_functor();
// If the particle tile was allocated in a memory pool that can run on GPU, launch GPU kernel
if constexpr (amrex::RunOnGpu<typename PTile::template AllocatorType<amrex::Real>>::value) {
if (run_on_gpu) {
amrex::ParallelForRNG(stop - start,
[=] AMREX_GPU_DEVICE (int i, amrex::RandomEngine const& engine) noexcept {
const int ip = i + start;
Expand All @@ -185,7 +190,7 @@ void DefaultInitializeRuntimeAttributes (PTile& ptile,
const BreitWheelerGetOpticalDepth breit_wheeler_get_opt =
p_bw_engine->build_optical_depth_functor();;
// If the particle tile was allocated in a memory pool that can run on GPU, launch GPU kernel
if constexpr (amrex::RunOnGpu<typename PTile::template AllocatorType<amrex::Real>>::value) {
if (run_on_gpu) {
amrex::ParallelForRNG(stop - start,
[=] AMREX_GPU_DEVICE (int i, amrex::RandomEngine const& engine) noexcept {
const int ip = i + start;
Expand Down Expand Up @@ -214,7 +219,7 @@ void DefaultInitializeRuntimeAttributes (PTile& ptile,
const amrex::ParserExecutor<7> user_real_attrib_parserexec =
user_real_attrib_parser[ia]->compile<7>();
// If the particle tile was allocated in a memory pool that can run on GPU, launch GPU kernel
if constexpr (amrex::RunOnGpu<typename PTile::template AllocatorType<amrex::Real>>::value) {
if (run_on_gpu) {
amrex::ParallelFor(stop - start,
[=] AMREX_GPU_DEVICE (int i) noexcept {
const int ip = i + start;
Expand Down Expand Up @@ -246,7 +251,7 @@ void DefaultInitializeRuntimeAttributes (PTile& ptile,
if (it_ioniz != particle_icomps.end() &&
std::distance(particle_icomps.begin(), it_ioniz) == j)
{
if constexpr (amrex::RunOnGpu<typename PTile::template AllocatorType<int>>::value) {
if (run_on_gpu) {
amrex::ParallelFor(stop - start,
[=] AMREX_GPU_DEVICE (int i) noexcept {
const int ip = i + start;
Expand All @@ -268,7 +273,7 @@ void DefaultInitializeRuntimeAttributes (PTile& ptile,
{
const amrex::ParserExecutor<7> user_int_attrib_parserexec =
user_int_attrib_parser[ia]->compile<7>();
if constexpr (amrex::RunOnGpu<typename PTile::template AllocatorType<int>>::value) {
if (run_on_gpu) {
amrex::ParallelFor(stop - start,
[=] AMREX_GPU_DEVICE (int i) noexcept {
const int ip = i + start;
Expand Down
Loading