Commit 14c4aea3 authored by Stefan Paquay's avatar Stefan Paquay
Browse files

Merge branch 'master' into port-enforce2d-kokkos

parents 4bf9a93c 666a13ea

cmake/CMakeLists.txt

+17 −10
Original line number Diff line number Diff line
@@ -2,9 +2,9 @@ 
# CMake build system

# This file is part of LAMMPS

# Created by Christoph Junghans and Richard Berger

cmake_minimum_required(VERSION 3.1)

cmake_minimum_required(VERSION 2.8.12)



project(lammps LANGUAGES CXX)

project(lammps CXX)

set(SOVERSION 0)

set(LAMMPS_SOURCE_DIR ${CMAKE_CURRENT_SOURCE_DIR}/../src)

set(LAMMPS_LIB_SOURCE_DIR ${CMAKE_CURRENT_SOURCE_DIR}/../lib)
@@ -129,7 +129,7 @@ foreach(PKG ${ACCEL_PACKAGES} ${OTHER_PACKAGES}) 
endforeach()



macro(pkg_depends PKG1 PKG2)

  if(PKG_${PKG1} AND NOT PKG_${PKG2})

  if(PKG_${PKG1} AND NOT (PKG_${PKG2} OR BUILD_${PKG2}))

    message(FATAL_ERROR "${PKG1} package needs LAMMPS to be build with ${PKG2}")

  endif()

endmacro()
@@ -137,23 +137,27 @@ endmacro() 
pkg_depends(MPIIO MPI)

pkg_depends(QEQ MANYBODY)

pkg_depends(USER-ATC MANYBODY)

pkg_depends(USER-H5MD MPI)

pkg_depends(USER-LB MPI)

pkg_depends(USER-MISC MANYBODY)

pkg_depends(USER-PHONON KSPACE)

pkg_depends(CORESHELL KSPACE)



######################################################

# packages with special compiler needs or external libs

######################################################

if(PKG_REAX OR PKG_MEAM OR PKG_USER-QUIP OR PKG_USER-QMMM OR PKG_LATTE)

  enable_language(Fortran)

  list(APPEND LAMMPS_LINK_LIBS ${CMAKE_Fortran_IMPLICIT_LINK_LIBRARIES})

endif()



if(PKG_MEAM)

if(PKG_MEAM OR PKG_USER-H5MD OR PKG_USER-QMMM)

  enable_language(C)

endif()



if(PKG_KOKKOS OR PKG_MSCG)

if(PKG_MSCG)

  if (CMAKE_VERSION VERSION_LESS "3.1") 

    message(FATAL_ERROR "For the MSCG package you need at least cmake-3.1")

  endif()

  # starting with CMake 3.1 this is all you have to do to enforce C++11

  set(CMAKE_CXX_STANDARD 11) # C++11...

  set(CMAKE_CXX_STANDARD_REQUIRED ON) #...is required...
@@ -312,7 +316,7 @@ if(PKG_LATTE) 
      message(FATAL_ERROR "LATTE not found, help CMake to find it by setting LATTE_LIBRARY, or set DOWNLOAD_LATTE=ON to download it")

    endif()

  endif()

  list(APPEND LAMMPS_LINK_LIBS ${LATTE_LIBRARIES} ${LAPACK_LIBRARIES} ${CMAKE_Fortran_IMPLICIT_LINK_LIBRARIES})

  list(APPEND LAMMPS_LINK_LIBS ${LATTE_LIBRARIES} ${LAPACK_LIBRARIES})

endif()



if(PKG_USER-MOLFILE)
@@ -355,14 +359,14 @@ endif() 


if(PKG_USER-QUIP)

  find_package(QUIP REQUIRED)

  list(APPEND LAMMPS_LINK_LIBS ${QUIP_LIBRARIES} ${LAPACK_LIBRARIES} ${CMAKE_Fortran_IMPLICIT_LINK_LIBRARIES})

  list(APPEND LAMMPS_LINK_LIBS ${QUIP_LIBRARIES} ${LAPACK_LIBRARIES})

endif()



if(PKG_USER-QMMM)

  message(WARNING "Building QMMM with CMake is still experimental")

  find_package(QE REQUIRED)

  include_directories(${QE_INCLUDE_DIRS})

  list(APPEND LAMMPS_LINK_LIBS ${QE_LIBRARIES} ${CMAKE_Fortran_IMPLICIT_LINK_LIBRARIES})

  list(APPEND LAMMPS_LINK_LIBS ${QE_LIBRARIES})

endif()



if(PKG_USER-VTK)
@@ -657,6 +661,9 @@ if(PKG_USER-INTEL) 
endif()



if(PKG_GPU)

    if (CMAKE_VERSION VERSION_LESS "3.1") 

      message(FATAL_ERROR "For the GPU package you need at least cmake-3.1")

    endif()

    set(GPU_SOURCES_DIR ${LAMMPS_SOURCE_DIR}/GPU)

    set(GPU_SOURCES ${GPU_SOURCES_DIR}/gpu_extra.h

                    ${GPU_SOURCES_DIR}/fix_gpu.h
@@ -753,7 +760,7 @@ if(PKG_GPU) 
      add_library(gpu STATIC ${GPU_LIB_SOURCES})

      target_link_libraries(gpu ${OpenCL_LIBRARIES})

      target_include_directories(gpu PRIVATE ${CMAKE_CURRENT_BINARY_DIR}/gpu ${OpenCL_INCLUDE_DIRS})

      target_compile_definitions(gpu PRIVATE -D_${GPU_PREC} -DMPI_GERYON -DUCL_NO_EXIT)

      target_compile_definitions(gpu PRIVATE -D_${GPU_PREC} -D${OCL_TUNE}_OCL -DMPI_GERYON -DUCL_NO_EXIT)

      target_compile_definitions(gpu PRIVATE -DUSE_OPENCL)



      list(APPEND LAMMPS_LINK_LIBS gpu)

lib/gpu/Makefile.linux_multi

0 → 100644
+53 −0
Original line number Diff line number Diff line 
# /* ----------------------------------------------------------------------   

#  Generic Linux Makefile for CUDA 

#     - Change CUDA_ARCH for your GPU

# ------------------------------------------------------------------------- */



# which file will be copied to Makefile.lammps



EXTRAMAKE = Makefile.lammps.standard



ifeq ($(CUDA_HOME),)

CUDA_HOME = /usr/local/cuda

endif



NVCC = nvcc



# Kepler CUDA

#CUDA_ARCH = -arch=sm_35

# newer CUDA

#CUDA_ARCH = -arch=sm_13

# older CUDA

#CUDA_ARCH = -arch=sm_10 -DCUDA_PRE_THREE



CUDA_ARCH = -gencode arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61



# this setting should match LAMMPS Makefile

# one of LAMMPS_SMALLBIG (default), LAMMPS_BIGBIG and LAMMPS_SMALLSMALL



LMP_INC = -DLAMMPS_SMALLBIG



# precision for GPU calculations

# -D_SINGLE_SINGLE  # Single precision for all calculations

# -D_DOUBLE_DOUBLE  # Double precision for all calculations

# -D_SINGLE_DOUBLE  # Accumulation of forces, etc. in double



CUDA_PRECISION = -D_SINGLE_DOUBLE



CUDA_INCLUDE = -I$(CUDA_HOME)/include

CUDA_LIB = -L$(CUDA_HOME)/lib64

CUDA_OPTS = -DUNIX -O3 -Xptxas -v --use_fast_math $(LMP_INC) -Xcompiler "-fPIC -std=c++98"



CUDR_CPP = mpicxx -DMPI_GERYON -DUCL_NO_EXIT -DMPICH_IGNORE_CXX_SEEK -DOMPI_SKIP_MPICXX=1 -fPIC

CUDR_OPTS = -O2 $(LMP_INC) # -xHost -no-prec-div -ansi-alias



BIN_DIR = ./

OBJ_DIR = ./

LIB_DIR = ./

AR = ar

BSH = /bin/sh



CUDPP_OPT = -DUSE_CUDPP -Icudpp_mini



include Nvidia.makefile_multi

lib/gpu/Nvidia.makefile

+30 −2
Original line number Diff line number Diff line
@@ -77,7 +77,9 @@ OBJS = $(OBJ_DIR)/lal_atom.o $(OBJ_DIR)/lal_ans.o \ 
       $(OBJ_DIR)/lal_coul_debye.o $(OBJ_DIR)/lal_coul_debye_ext.o \

       $(OBJ_DIR)/lal_zbl.o $(OBJ_DIR)/lal_zbl_ext.o \

       $(OBJ_DIR)/lal_lj_cubic.o $(OBJ_DIR)/lal_lj_cubic_ext.o \

       $(OBJ_DIR)/lal_ufm.o $(OBJ_DIR)/lal_ufm_ext.o 

       $(OBJ_DIR)/lal_ufm.o $(OBJ_DIR)/lal_ufm_ext.o \

       $(OBJ_DIR)/lal_dipole_long_lj.o $(OBJ_DIR)/lal_dipole_long_lj_ext.o \

       $(OBJ_DIR)/lal_lj_expand_coul_long.o $(OBJ_DIR)/lal_lj_expand_coul_long_ext.o



CBNS = $(OBJ_DIR)/device.cubin $(OBJ_DIR)/device_cubin.h \

       $(OBJ_DIR)/atom.cubin $(OBJ_DIR)/atom_cubin.h \
@@ -133,7 +135,9 @@ CBNS = $(OBJ_DIR)/device.cubin $(OBJ_DIR)/device_cubin.h \ 
       $(OBJ_DIR)/coul_debye.cubin $(OBJ_DIR)/coul_debye_cubin.h \

       $(OBJ_DIR)/zbl.cubin $(OBJ_DIR)/zbl_cubin.h \

       $(OBJ_DIR)/lj_cubic.cubin $(OBJ_DIR)/lj_cubic_cubin.h \

       $(OBJ_DIR)/ufm.cubin $(OBJ_DIR)/ufm_cubin.h

       $(OBJ_DIR)/ufm.cubin $(OBJ_DIR)/ufm_cubin.h \

       $(OBJ_DIR)/dipole_long_lj.cubin $(OBJ_DIR)/dipole_long_lj_cubin.h \

       $(OBJ_DIR)/lj_expand_coul_long.cubin $(OBJ_DIR)/lj_expand_coul_long_cubin.h



all: $(OBJ_DIR) $(GPU_LIB) $(EXECS)
@@ -809,6 +813,30 @@ $(OBJ_DIR)/lal_lj_cubic.o: $(ALL_H) lal_lj_cubic.h lal_lj_cubic.cpp $(OBJ_DIR)/l 
$(OBJ_DIR)/lal_lj_cubic_ext.o: $(ALL_H) lal_lj_cubic.h lal_lj_cubic_ext.cpp lal_base_atomic.h

	$(CUDR) -o $@ -c lal_lj_cubic_ext.cpp -I$(OBJ_DIR)



$(OBJ_DIR)/dipole_long_lj.cubin: lal_dipole_long_lj.cu lal_precision.h lal_preprocessor.h

	$(CUDA) --cubin -DNV_KERNEL -o $@ lal_dipole_long_lj.cu



$(OBJ_DIR)/dipole_long_lj_cubin.h: $(OBJ_DIR)/dipole_long_lj.cubin $(OBJ_DIR)/dipole_long_lj.cubin

	$(BIN2C) -c -n dipole_long_lj $(OBJ_DIR)/dipole_long_lj.cubin > $(OBJ_DIR)/dipole_long_lj_cubin.h



$(OBJ_DIR)/lal_dipole_long_lj.o: $(ALL_H) lal_dipole_long_lj.h lal_dipole_long_lj.cpp $(OBJ_DIR)/dipole_long_lj_cubin.h $(OBJ_DIR)/lal_base_dipole.o

	$(CUDR) -o $@ -c lal_dipole_long_lj.cpp -I$(OBJ_DIR)



$(OBJ_DIR)/lal_dipole_long_lj_ext.o: $(ALL_H) lal_dipole_long_lj.h lal_dipole_long_lj_ext.cpp lal_base_dipole.h

	$(CUDR) -o $@ -c lal_dipole_long_lj_ext.cpp -I$(OBJ_DIR)



$(OBJ_DIR)/lj_expand_coul_long.cubin: lal_lj_expand_coul_long.cu lal_precision.h lal_preprocessor.h

	$(CUDA) --cubin -DNV_KERNEL -o $@ lal_lj_expand_coul_long.cu



$(OBJ_DIR)/lj_expand_coul_long_cubin.h: $(OBJ_DIR)/lj_expand_coul_long.cubin $(OBJ_DIR)/lj_expand_coul_long.cubin

	$(BIN2C) -c -n lj_expand_coul_long $(OBJ_DIR)/lj_expand_coul_long.cubin > $(OBJ_DIR)/lj_expand_coul_long_cubin.h



$(OBJ_DIR)/lal_lj_expand_coul_long.o: $(ALL_H) lal_lj_expand_coul_long.h lal_lj_expand_coul_long.cpp $(OBJ_DIR)/lj_expand_coul_long_cubin.h $(OBJ_DIR)/lal_base_charge.o

	$(CUDR) -o $@ -c lal_lj_expand_coul_long.cpp -I$(OBJ_DIR)



$(OBJ_DIR)/lal_lj_expand_coul_long_ext.o: $(ALL_H) lal_lj_expand_coul_long.h lal_lj_expand_coul_long_ext.cpp lal_base_charge.h

	$(CUDR) -o $@ -c lal_lj_expand_coul_long_ext.cpp -I$(OBJ_DIR)



$(BIN_DIR)/nvc_get_devices: ./geryon/ucl_get_devices.cpp $(NVD_H)

	$(CUDR) -o $@ ./geryon/ucl_get_devices.cpp -DUCL_CUDADR $(CUDA_LIB) -lcuda

lib/gpu/Nvidia.makefile_multi

0 → 100644
+854 −0

File added.

Preview size limit exceeded, changes collapsed.

lib/gpu/Opencl.makefile

+24 −3
Original line number Diff line number Diff line
@@ -66,7 +66,9 @@ OBJS = $(OBJ_DIR)/lal_atom.o $(OBJ_DIR)/lal_answer.o \ 
       $(OBJ_DIR)/lal_coul_debye.o $(OBJ_DIR)/lal_coul_debye_ext.o \

       $(OBJ_DIR)/lal_zbl.o $(OBJ_DIR)/lal_zbl_ext.o \

       $(OBJ_DIR)/lal_lj_cubic.o $(OBJ_DIR)/lal_lj_cubic_ext.o \

       $(OBJ_DIR)/lal_ufm.o $(OBJ_DIR)/lal_ufm_ext.o

       $(OBJ_DIR)/lal_ufm.o $(OBJ_DIR)/lal_ufm_ext.o \

       $(OBJ_DIR)/lal_dipole_long_lj.o $(OBJ_DIR)/lal_dipole_long_lj_ext.o \

       $(OBJ_DIR)/lal_lj_expand_coul_long.o $(OBJ_DIR)/lal_lj_expand_coul_long_ext.o



KERS = $(OBJ_DIR)/device_cl.h $(OBJ_DIR)/atom_cl.h \

       $(OBJ_DIR)/neighbor_cpu_cl.h $(OBJ_DIR)/pppm_cl.h \
@@ -95,7 +97,8 @@ KERS = $(OBJ_DIR)/device_cl.h $(OBJ_DIR)/atom_cl.h \ 
       $(OBJ_DIR)/tersoff_mod_cl.h $(OBJ_DIR)/coul_cl.h \

       $(OBJ_DIR)/coul_debye_cl.h $(OBJ_DIR)/zbl_cl.h \

       $(OBJ_DIR)/lj_cubic_cl.h $(OBJ_DIR)/vashishta_cl.h \

       $(OBJ_DIR)/ufm_cl.h

       $(OBJ_DIR)/ufm_cl.h  $(OBJ_DIR)/dipole_long_lj_cl.h \

       $(OBJ_DIR)/lj_expand_coul_long_cl.h





OCL_EXECS = $(BIN_DIR)/ocl_get_devices
@@ -588,7 +591,25 @@ $(OBJ_DIR)/lal_ufm.o: $(ALL_H) lal_ufm.h lal_ufm.cpp $(OBJ_DIR)/ufm_cl.h $(OBJ_ 
$(OBJ_DIR)/lal_ufm_ext.o: $(ALL_H) lal_ufm.h lal_ufm_ext.cpp lal_base_atomic.h

	$(OCL) -o $@ -c lal_ufm_ext.cpp -I$(OBJ_DIR)



$(BIN_DIR)/ocl_get_devices: ./geryon/ucl_get_devices.cpp

$(OBJ_DIR)/dipole_long_lj_cl.h: lal_dipole_long_lj.cu $(PRE1_H)

	$(BSH) ./geryon/file_to_cstr.sh dipole_long_lj $(PRE1_H) lal_dipole_long_lj.cu $(OBJ_DIR)/dipole_long_lj_cl.h;



$(OBJ_DIR)/lal_dipole_long_lj.o: $(ALL_H) lal_dipole_long_lj.h lal_dipole_long_lj.cpp  $(OBJ_DIR)/dipole_long_lj_cl.h $(OBJ_DIR)/lj_expand_coul_long_cl.h $(OBJ_DIR)/lal_base_charge.o

	$(OCL) -o $@ -c lal_dipole_long_lj.cpp -I$(OBJ_DIR)



$(OBJ_DIR)/lal_dipole_long_lj_ext.o: $(ALL_H) lal_dipole_long_lj.h lal_dipole_long_lj_ext.cpp lal_base_dipole.h

	$(OCL) -o $@ -c lal_dipole_long_lj_ext.cpp -I$(OBJ_DIR)



$(OBJ_DIR)/lj_expand_coul_long_cl.h: lal_lj_expand_coul_long.cu $(PRE1_H)

	$(BSH) ./geryon/file_to_cstr.sh lj_expand_coul_long $(PRE1_H) lal_lj_expand_coul_long.cu $(OBJ_DIR)/lj_expand_coul_long_cl.h;



$(OBJ_DIR)/lal_lj_expand_coul_long.o: $(ALL_H) lal_lj_expand_coul_long.h lal_lj_expand_coul_long.cpp  $(OBJ_DIR)/lj_expand_coul_long_cl.h $(OBJ_DIR)/lj_expand_coul_long_cl.h $(OBJ_DIR)/lal_base_charge.o

	$(OCL) -o $@ -c lal_lj_expand_coul_long.cpp -I$(OBJ_DIR)



$(OBJ_DIR)/lal_lj_expand_coul_long_ext.o: $(ALL_H) lal_lj_expand_coul_long.h lal_lj_expand_coul_long_ext.cpp lal_base_charge.h

	$(OCL) -o $@ -c lal_lj_expand_coul_long_ext.cpp -I$(OBJ_DIR)



$(BIN_DIR)/ocl_get_devices: ./geryon/ucl_get_devices.cpp $(OCL_H)

	$(OCL) -o $@ ./geryon/ucl_get_devices.cpp -DUCL_OPENCL $(OCL_LINK) 



$(OCL_LIB): $(OBJS) $(PTXS)

CRA Git | Maintained and supported by SUSTech CRA and CCSE