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

call_f refactor #3452

Merged
merged 6 commits into from
Jul 27, 2023
Merged

call_f refactor #3452

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
27d0ce8
use if constexpr in DenseBins
atmyers Jul 26, 2023
df81d7a
use if constexpr in ParticleMesh
atmyers Jul 26, 2023
f39d78e
also do ParticleReduce
atmyers Jul 26, 2023
5132971
a little cleanup
atmyers Jul 26, 2023
fe330b1
Merge branch 'development' into call_f_refactor
atmyers Jul 26, 2023
e682522
Update Src/Particle/AMReX_DenseBins.H
atmyers Jul 26, 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
18 changes: 6 additions & 12 deletions Src/Particle/AMReX_DenseBins.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -93,18 +93,12 @@ private:

template <typename F, typename I>
AMREX_GPU_HOST_DEVICE
static auto call_f(F const& f, const_pointer_input_type v, I& index)
noexcept -> decltype(f(v,index))
{
return f(v,index);
}

template <typename F, typename I>
AMREX_GPU_HOST_DEVICE
static auto call_f(F const& f, const_pointer_input_type v, I& index)
noexcept -> decltype(f(v[index]))
{
return f(v[index]);
static auto call_f (F const& f, const_pointer_input_type v, I& index) {
if constexpr (IsCallable<F, decltype(v), I>::value) {
return f(v, index);
} else {
return f(v[index]);
}
}

public:
Expand Down
27 changes: 26 additions & 1 deletion Src/Particle/AMReX_ParticleMesh.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,8 +6,33 @@
#include <AMReX_MultiFab.H>
#include <AMReX_ParticleUtil.H>

namespace amrex
namespace amrex {

namespace particle_detail {

template <typename F, typename T, typename T_ParticleType, template<class, int, int> class PTDType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const PTDType<T_ParticleType, NAR, NAI>& p,
const int i, Array4<T> const& fabarr,
GpuArray<Real,AMREX_SPACEDIM> const& plo,
GpuArray<Real,AMREX_SPACEDIM> const& dxi) noexcept
{
if constexpr (IsCallable<F, decltype(p.m_aos[i]), decltype(fabarr), decltype(plo), decltype(dxi)>::value) {
return f(p.m_aos[i], fabarr, plo, dxi);
} else if constexpr (IsCallable<F, decltype(p.m_aos[i]), decltype(fabarr)>::value) {
return f(p.m_aos[i], fabarr);
} else if constexpr (IsCallable<F, decltype(p.getSuperParticle(i)), decltype(fabarr), decltype(plo), decltype(dxi)>::value) {
return f(p.getSuperParticle(i), fabarr, plo, dxi);
} else if constexpr (IsCallable<F, decltype(p.getSuperParticle(i)), decltype(fabarr)>::value) {
return f(p.getSuperParticle(i), fabarr);
} else if constexpr (IsCallable<F, decltype(p), decltype(fabarr), decltype(plo), decltype(dxi)>::value) {
return f(p, i, fabarr, plo, dxi);
} else {
return f(p, i, fabarr);
}
}
}

template <class PC, class MF, class F, std::enable_if_t<IsParticleContainer<PC>::value, int> foo = 0>
void
Expand Down
19 changes: 18 additions & 1 deletion Src/Particle/AMReX_ParticleReduce.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,8 +13,25 @@

#include <limits>

namespace amrex
namespace amrex {

namespace particle_detail {

template <typename F, typename T_ParticleType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ConstParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i) noexcept
{
if constexpr (IsCallable<F, decltype(p.m_aos[i])>::value) {
return f(p.m_aos[i]);
} else if constexpr (IsCallable<F, decltype(p.getSuperParticle(i))>::value) {
return f(p.getSuperParticle(i));
} else {
return f(p, i);
}
}
}

/**
* \brief A general reduction method for the particles in a ParticleContainer that can run on either CPUs or GPUs.
Expand Down
178 changes: 1 addition & 177 deletions Src/Particle/AMReX_ParticleUtil.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,183 +17,7 @@

#include <limits>

namespace amrex
{

namespace particle_detail {

// The next several functions are used by ParticleReduce

// Lambda takes a Particle
template <typename F, typename T_ParticleType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ConstParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i) noexcept
-> decltype(f(p.m_aos[i]))
{
return f(p.m_aos[i]);
}

// Lambda takes a SuperParticle
template <typename F, typename T_ParticleType, int NAR, int NAI,
typename std::enable_if<NAR != 0 || NAI != 0, int>::type = 0>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ConstParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i) noexcept
-> decltype(f(p.getSuperParticle(i)))
{
return f(p.getSuperParticle(i));
}

// Lambda takes a ConstParticleTileData
template <typename F, typename T_ParticleType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ConstParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i) noexcept
-> decltype(f(p, i))
{
return f(p, i);
}

// These next several functions are used by ParticleToMesh and MeshToParticle

// Lambda takes a Particle
template <typename F, typename T, typename T_ParticleType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ConstParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i, Array4<T> const& fabarr,
GpuArray<Real,AMREX_SPACEDIM> const& plo,
GpuArray<Real,AMREX_SPACEDIM> const& dxi) noexcept
-> decltype(f(p.m_aos[i], fabarr, plo, dxi))
{
return f(p.m_aos[i], fabarr, plo, dxi);
}

// Lambda takes a Particle
template <typename F, typename T, typename T_ParticleType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ConstParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i, Array4<T> const& fabarr,
GpuArray<Real,AMREX_SPACEDIM> const&,
GpuArray<Real,AMREX_SPACEDIM> const&) noexcept
-> decltype(f(p.m_aos[i], fabarr))
{
return f(p.m_aos[i], fabarr);
}

// Lambda takes a Particle
template <typename F, typename T, typename T_ParticleType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i, Array4<const T> const& fabarr,
GpuArray<Real,AMREX_SPACEDIM> const& plo,
GpuArray<Real,AMREX_SPACEDIM> const& dxi) noexcept
-> decltype(f(p.m_aos[i], fabarr, plo, dxi))
{
return f(p.m_aos[i], fabarr, plo, dxi);
}

// Lambda takes a Particle
template <typename F, typename T, typename T_ParticleType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i, Array4<const T> const& fabarr,
GpuArray<Real,AMREX_SPACEDIM> const&,
GpuArray<Real,AMREX_SPACEDIM> const&) noexcept
-> decltype(f(p.m_aos[i], fabarr))
{
return f(p.m_aos[i], fabarr);
}

// Lambda takes a SuperParticle
template <typename F, typename T, typename T_ParticleType, int NAR, int NAI,
typename std::enable_if<(NAR != 0) || (NAI != 0), int>::type = 0>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ConstParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i, Array4<T> const& fabarr,
GpuArray<Real,AMREX_SPACEDIM> const& plo,
GpuArray<Real,AMREX_SPACEDIM> const& dxi) noexcept
-> decltype(f(p.getSuperParticle(i), fabarr, plo, dxi))
{
return f(p.getSuperParticle(i), fabarr, plo, dxi);
}

// Lambda takes a SuperParticle
template <typename F, typename T, typename T_ParticleType, int NAR, int NAI,
typename std::enable_if<(NAR != 0) || (NAI != 0), int>::type = 0>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ConstParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i, Array4<T> const& fabarr,
GpuArray<Real,AMREX_SPACEDIM> const&,
GpuArray<Real,AMREX_SPACEDIM> const&) noexcept
-> decltype(f(p.getSuperParticle(i), fabarr))
{
return f(p.getSuperParticle(i), fabarr);
}

// Lambda takes a ConstParticleTileData
template <typename F, typename T, typename T_ParticleType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ConstParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i, Array4<T> const& fabarr,
GpuArray<Real,AMREX_SPACEDIM> const&,
GpuArray<Real,AMREX_SPACEDIM> const&) noexcept
-> decltype(f(p, i, fabarr))
{
return f(p, i, fabarr);
}

// Lambda takes a ConstParticleTileData
template <typename F, typename T, typename T_ParticleType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ConstParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i, Array4<T> const& fabarr,
GpuArray<Real,AMREX_SPACEDIM> const& plo,
GpuArray<Real,AMREX_SPACEDIM> const& dxi) noexcept
-> decltype(f(p, i, fabarr, plo, dxi))
{
return f(p, i, fabarr, plo, dxi);
}

// Lambda takes a ParticleTileData
template <typename F, typename T, typename T_ParticleType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i, Array4<const T> const& fabarr,
GpuArray<Real,AMREX_SPACEDIM> const& plo,
GpuArray<Real,AMREX_SPACEDIM> const& dxi) noexcept
-> decltype(f(p, i, fabarr, plo, dxi))
{
return f(p, i, fabarr, plo, dxi);
}

// Lambda takes a ParticleTileData
template <typename F, typename T, typename T_ParticleType, int NAR, int NAI>
AMREX_GPU_HOST_DEVICE AMREX_FORCE_INLINE
auto call_f (F const& f,
const ParticleTileData<T_ParticleType, NAR, NAI>& p,
const int i, Array4<const T> const& fabarr,
GpuArray<Real,AMREX_SPACEDIM> const&,
GpuArray<Real,AMREX_SPACEDIM> const&) noexcept
-> decltype(f(p, i, fabarr))
{
return f(p, i, fabarr);
}


}
namespace amrex {

/**
* \brief Returns the number of particles that are more than nGrow cells
Expand Down
Loading