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

Rework cuda file handling for >= 12 #89

Merged
merged 1 commit into from
Oct 14, 2024
Merged

Rework cuda file handling for >= 12 #89

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
2 changes: 1 addition & 1 deletion expui/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@ set(common_INCLUDE $<INSTALL_INTERFACE:include>


if(ENABLE_CUDA)
list(APPEND common_LINKLIB CUDA::cudart)
list(APPEND common_LINKLIB CUDA::toolkit CUDA::cudart)
if (CUDAToolkit_VERSION VERSION_GREATER_EQUAL 12)
list(APPEND common_LINKLIB CUDA::nvtx3)
else ()
Expand Down
7 changes: 4 additions & 3 deletions exputil/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,14 +51,15 @@ set(common_INCLUDE_DIRS $<INSTALL_INTERFACE:include>
${DEP_INC} ${EIGEN3_INCLUDE_DIR} ${HDF5_INCLUDE_DIRS}
${FFTW_INCLUDE_DIRS})

set(common_LINKLIBS ${DEP_LIB} OpenMP::OpenMP_CXX MPI::MPI_CXX
set(common_LINKLIB ${DEP_LIB} OpenMP::OpenMP_CXX MPI::MPI_CXX
yaml-cpp ${VTK_LIBRARIES} ${HDF5_LIBRARIES} ${HDF5_HL_LIBRARIES}
${FFTW_DOUBLE_LIB})

if(ENABLE_CUDA)
list(APPEND common_LINKLIB CUDA::cudart)
list(APPEND common_LINKLIB CUDA::toolkit CUDA::cudart)
if (CUDAToolkit_VERSION VERSION_GREATER_EQUAL 12)
list(APPEND common_LINKLIB CUDA::nvtx3)
set_source_files_properties(${BIORTH_SRC} ${PARTICLE_SRC} PROPERTIES LANGUAGE CUDA)
else ()
list(APPEND common_LINKLIB CUDA::nvToolsExt)
endif ()
Expand All @@ -74,7 +75,7 @@ endif()
add_library(exputil ${exputil_SOURCES})
set_target_properties(exputil PROPERTIES OUTPUT_NAME exputil)
target_include_directories(exputil PUBLIC ${common_INCLUDE_DIRS})
target_link_libraries(exputil PUBLIC ${common_LINKLIBS})
target_link_libraries(exputil PUBLIC ${common_LINKLIB})

install(TARGETS exputil DESTINATION lib)

Expand Down
4 changes: 2 additions & 2 deletions include/BiorthCube.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@
#include <massmodel.H>
#include <yaml-cpp/yaml.h>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaUtil.cuH>
#include <cudaMappingConstants.cuH>
#endif
Expand All @@ -28,7 +28,7 @@
#include <highfive/H5DataSpace.hpp>
#include <highfive/H5Attribute.hpp>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaParticle.cuH>
#include <cudaMappingConstants.cuH>
#endif
Expand Down
6 changes: 3 additions & 3 deletions include/BiorthCyl.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@
#include <massmodel.H>
#include <yaml-cpp/yaml.h>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaUtil.cuH>
#include <cudaMappingConstants.cuH>
#endif
Expand All @@ -28,7 +28,7 @@
#include <highfive/H5DataSpace.hpp>
#include <highfive/H5Attribute.hpp>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaParticle.cuH>
#include <cudaMappingConstants.cuH>
#endif
Expand Down Expand Up @@ -172,7 +172,7 @@ public:
static std::map<std::string, std::string>
cacheInfo(const std::string& cachefile, bool verbose=true);

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
void initialize_cuda(std::vector<cudaArray_t>& cuArray,
thrust::host_vector<cudaTextureObject_t>& tex);

Expand Down
4 changes: 2 additions & 2 deletions include/EmpCylSL.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,7 @@
#include <SLGridMP2.H>
#include <coef.H>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaParticle.cuH>
#include <cudaMappingConstants.cuH>
#endif
Expand Down Expand Up @@ -924,7 +924,7 @@ public:
//! Check orthogonality for basis (pyEXP style)
std::vector<Eigen::MatrixXd> orthoCheck();

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
cudaMappingConstants getCudaMappingConstants();

void initialize_cuda(std::vector<cudaArray_t>& cuArray,
Expand Down
6 changes: 3 additions & 3 deletions include/SLGridMP2.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@
#include <libvars.H>
using namespace __EXP__;

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaUtil.cuH>
#include <cudaMappingConstants.cuH>
#endif
Expand Down Expand Up @@ -167,7 +167,7 @@ public:
//! produce matrices
std::vector<Eigen::MatrixXd> orthoCheck(int knots=40);

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
void initialize_cuda(std::vector<cudaArray_t>& cuArray,
thrust::host_vector<cudaTextureObject_t>& tex);

Expand Down Expand Up @@ -453,7 +453,7 @@ public:

//@}

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
void initialize_cuda(std::vector<cudaArray_t>& cuArray,
thrust::host_vector<cudaTextureObject_t>& tex);

Expand Down
2 changes: 1 addition & 1 deletion pyEXP/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@ set(common_INCLUDE $<INSTALL_INTERFACE:include>
${HDF5_INCLUDE_DIRS} ${EIGEN3_INCLUDE_DIR})

if(ENABLE_CUDA)
list(APPEND common_LINKLIB CUDA::cudart)
list(APPEND common_LINKLIB CUDA::toolkit CUDA::cudart)
if (CUDAToolkit_VERSION VERSION_GREATER_EQUAL 12)
list(APPEND common_LINKLIB CUDA::nvtx3)
else ()
Expand Down
16 changes: 6 additions & 10 deletions src/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,15 +18,6 @@ if (ENABLE_CUDA)
cudaCylinder.cu cudaEmpCylSL.cu cudaComponent.cu NVTX.cc
cudaIncpos.cu cudaIncvel.cu cudaMultistep.cu cudaOrient.cu
cudaBiorthCyl.cu cudaCube.cu cudaSlabSL.cu)

set_source_files_properties(Component.cc ComponentContainer.cc
Cube.cc Cylinder.cc ExternalForce.cc NVTX.cc OrbTrace.cc Orient.cc
OutAscii.cc OutCHKPT.cc OutCHKPTQ.cc OutCalbr.cc OutFrac.cc
OutLog.cc OutPS.cc OutPSN.cc OutPSP.cc OutPSQ.cc OutPSR.cc
OutputContainer.cc PolarBasis.cc PotAccel.cc SlabSL.cc Sphere.cc
SphericalBasis.cc begin.cc incpos.cc incvel.cc step.cc
PROPERTIES LANGUAGE CUDA)

endif()

set(common_INCLUDE_DIRS
Expand All @@ -46,9 +37,10 @@ if(PNG_FOUND)
endif()

if(ENABLE_CUDA)
list(APPEND common_LINKLIB CUDA::cudart)
list(APPEND common_LINKLIB CUDA::toolkit CUDA::cudart)
if (CUDAToolkit_VERSION VERSION_GREATER_EQUAL 12)
list(APPEND common_LINKLIB CUDA::nvtx3)
set_source_files_properties(${exp_SOURCES} PROPERTIES LANGUAGE CUDA)
else ()
list(APPEND common_LINKLIB CUDA::nvToolsExt)
endif ()
Expand All @@ -75,5 +67,9 @@ add_executable(exp expand.cc)
target_include_directories(exp PUBLIC ${common_INCLUDE_DIRS})
target_link_libraries(exp PUBLIC ${common_LINKLIB} EXPlib)

if (ENABLE_CUDA)
set_target_properties(exp PROPERTIES LINKER_LANGUAGE CUDA)
endif ()

install(TARGETS EXPlib DESTINATION lib)
install(TARGETS exp DESTINATION bin)
8 changes: 4 additions & 4 deletions src/Component.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@

#include <config_exp.h>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaParticle.cuH>
#endif

Expand Down Expand Up @@ -569,15 +569,15 @@ public:
//! Compute center of mass and center of velocity (CPU version)
void fix_positions_cpu(unsigned mlevel=0);

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
//! Compute center of mass and center of velocity (GPU version)
void fix_positions_cuda(unsigned mlevel=0);
#endif

//! Compute center of mass and center of velocity
void fix_positions(unsigned mlevel=0)
{
#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
if (use_cuda) fix_positions_cuda(mlevel);
else
#endif
Expand Down Expand Up @@ -926,7 +926,7 @@ public:
//! Compute level from minimum requested time step from last master step
inline bool DTreset() { return dtreset; }

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
//@{
//! CUDA utilities for handling host <===> device exchange

Expand Down
6 changes: 3 additions & 3 deletions src/Cube.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@
#include <Coefficients.H>
#include <PotAccel.H>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <thrust/complex.h>
#include <cudaParticle.cuH>
#include <cudaMappingConstants.cuH>
Expand Down Expand Up @@ -41,7 +41,7 @@ private:
//! Valid keys for YAML configurations
static const std::set<std::string> valid_keys;

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
virtual void determine_coefficients_cuda();
virtual void determine_acceleration_cuda();
virtual void multistep_update_cuda();
Expand Down Expand Up @@ -135,7 +135,7 @@ private:
std::cout << std::string(60, '=') << std::endl;
std::cout << "Time in CPU: "
<< duration0.count()-duration1.count() << std::endl;
#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
if (c->cC->cudaDevice>=0) {
std::cout << "Time in GPU: " << duration1.count() << std::endl;
}
Expand Down
2 changes: 1 addition & 1 deletion src/CylEXP.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,7 @@
#include <expand.H>
#include <EmpCylSL.H>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaParticle.cuH>
#include <cudaMappingConstants.cuH>
#endif
Expand Down
4 changes: 2 additions & 2 deletions src/Cylinder.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@

#include <config_exp.h>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaParticle.cuH>
#include <cudaMappingConstants.cuH>
#endif
Expand Down Expand Up @@ -171,7 +171,7 @@ protected:
int sampT, defSampT;

//! CUDA method for coefficient accumulation
#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
virtual void determine_coefficients_cuda(bool compute_pca);
virtual void determine_acceleration_cuda();
virtual void multistep_update_cuda();
Expand Down
2 changes: 1 addition & 1 deletion src/ExternalForce.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,7 +53,7 @@ public:
//! Finish and clean-up (caching data necessary for restart)
virtual void finish() {}

// #if HAVE_LIBCUDA==1 && defined (__NVCC__)
// #if HAVE_LIBCUDA==1
#if HAVE_LIBCUDA==1
//! Copy particles from device for non-cuda forces
void getParticlesCuda(Component *c);
Expand Down
4 changes: 2 additions & 2 deletions src/FlatDisk.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@

typedef std::shared_ptr<BiorthCyl> CylPtr;

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaUtil.cuH>
#endif

Expand Down Expand Up @@ -61,7 +61,7 @@ private:

virtual double getRtable() { return ortho->getRtable(); }

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
virtual void initialize_cuda()
{
sampT = floor(sqrt(component->CurTotal()));
Expand Down
2 changes: 1 addition & 1 deletion src/NVTX.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,7 @@

#include <config_exp.h>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1

#if __CUDACC_VER_MAJOR__ < 12
#include <nvToolsExt.h>
Expand Down
4 changes: 2 additions & 2 deletions src/PolarBasis.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@

#include <config_exp.h>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaParticle.cuH>
#include <cudaMappingConstants.cuH>
#endif
Expand Down Expand Up @@ -198,7 +198,7 @@ protected:
virtual double getRtable() = 0;

//! CUDA method for coefficient accumulation
#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
virtual void determine_coefficients_cuda(bool compute_pca);
virtual void determine_acceleration_cuda();
virtual void multistep_update_cuda();
Expand Down
2 changes: 1 addition & 1 deletion src/PotAccel.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -195,7 +195,7 @@ public:
//! Execute to finish level shifts for particles
virtual void multistep_update_finish() {}

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
//! Cuda implementation of level shifts
virtual void multistep_update_cuda() {}
#endif
Expand Down
4 changes: 2 additions & 2 deletions src/SlabSL.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,7 +11,7 @@
#include <biorth1d.H>
#include <PotAccel.H>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <thrust/complex.h>
#include <cudaParticle.cuH>
#include <cudaMappingConstants.cuH>
Expand Down Expand Up @@ -56,7 +56,7 @@ private:

SlabSLCoefHeader coefheader;

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
virtual void determine_coefficients_cuda();
virtual void determine_acceleration_cuda();
virtual void multistep_update_cuda();
Expand Down
4 changes: 2 additions & 2 deletions src/Sphere.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@

typedef std::shared_ptr<SLGridSph> SLGridSphPtr;

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaUtil.cuH>
#endif

Expand Down Expand Up @@ -68,7 +68,7 @@ private:
void make_model_bin();
void make_model_plummer();

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
virtual void initialize_cuda()
{
ortho->initialize_cuda(cuInterpArray, tex);
Expand Down
4 changes: 2 additions & 2 deletions src/SphericalBasis.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,7 @@

#include <config_exp.h>

#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
#include <cudaParticle.cuH>
#include <cudaMappingConstants.cuH>
#endif
Expand Down Expand Up @@ -117,7 +117,7 @@ protected:
virtual void determine_coefficients_playback(void);

//! CUDA method for coefficient accumulation
#if HAVE_LIBCUDA==1 && defined (__NVCC__)
#if HAVE_LIBCUDA==1
virtual void determine_coefficients_cuda(bool compute_pca);
virtual void determine_acceleration_cuda();
virtual void multistep_update_cuda();
Expand Down
Loading