tx258/lammps-gran-kokkos
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Kept the .cubin files so that they are only rebuilt whenever changes are made to the relevant source files

Browse Source
This commit is contained in:
Trung Nguyen
2021-03-29 13:46:16 -05:00
parent 7662cfe28d
commit dca9cd9c1c
4 changed files with 4 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
lib/gpu/Makefile.cuda
View File

@ -45,7 +45,7 @@ CUDA_INCLUDE = -I$(CUDA_HOME)/include
CUDA_LIB = -L$(CUDA_HOME)/lib64 -L$(CUDA_HOME)/lib64/stubs
CUDA_OPTS = -DUNIX -O3 --use_fast_math $(LMP_INC) -Xcompiler -fPIC
CUDA_LINK = $(CUDA_LIB) -lcudart
CUDA = $(NVCC) $(CUDA_INCLUDE) $(CUDA_OPTS) -Icudpp_mini $(CUDA_ARCH) \
CUDA = $(NVCC) $(CUDA_INCLUDE) $(CUDA_OPTS) $(CUDA_ARCH) \
$(CUDA_PRECISION)
BIN2C = $(CUDA_HOME)/bin/bin2c
@ -103,7 +103,6 @@ $(OBJ_DIR)/pppm_d_cubin.h: $(OBJ_DIR)/pppm_d.cubin
$(OBJ_DIR)/%_cubin.h: lal_%.cu $(ALL_H)
$(CUDA) --fatbin -DNV_KERNEL -o $(OBJ_DIR)/$*.cubin $(OBJ_DIR)/lal_$*.cu
$(BIN2C) -c -n $* $(OBJ_DIR)/$*.cubin > $@
@rm $(OBJ_DIR)/$*.cubin
# host code compilation

3
lib/gpu/Makefile.cuda_mps
View File

@ -44,7 +44,7 @@ CUDA_INCLUDE = -I$(CUDA_HOME)/include
CUDA_LIB = -L$(CUDA_HOME)/lib64 -L$(CUDA_HOME)/lib64/stubs
CUDA_OPTS = -DUNIX -O3 --use_fast_math $(LMP_INC) -Xcompiler -fPIC
CUDA_LINK = $(CUDA_LIB) -lcudart
CUDA = $(NVCC) $(CUDA_INCLUDE) $(CUDA_OPTS) -Icudpp_mini $(CUDA_ARCH) \
CUDA = $(NVCC) $(CUDA_INCLUDE) $(CUDA_OPTS) $(CUDA_ARCH) \
$(CUDA_PRECISION)
BIN2C = $(CUDA_HOME)/bin/bin2c
@ -102,7 +102,6 @@ $(OBJ_DIR)/pppm_d_cubin.h: $(OBJ_DIR)/pppm_d.cubin
$(OBJ_DIR)/%_cubin.h: lal_%.cu $(ALL_H)
$(CUDA) --fatbin -DNV_KERNEL -o $(OBJ_DIR)/$*.cubin $(OBJ_DIR)/lal_$*.cu
$(BIN2C) -c -n $* $(OBJ_DIR)/$*.cubin > $@
@rm $(OBJ_DIR)/$*.cubin
# host code compilation

5
lib/gpu/Nvidia.makefile
View File

@ -30,7 +30,7 @@ $(OBJ_DIR):
# Compilers and linkers
CUDA = $(NVCC) $(CUDA_INCLUDE) $(CUDA_OPTS) -Icudpp_mini $(CUDA_ARCH) \
CUDA = $(NVCC) $(CUDA_INCLUDE) $(CUDA_OPTS) $(CUDA_ARCH) \
$(CUDA_PRECISION)
CUDR = $(CUDR_CPP) $(CUDR_OPTS) $(CUDA_PRECISION) $(CUDA_INCLUDE) \
$(CUDPP_OPT)
@ -46,7 +46,6 @@ $(OBJ_DIR)/pppm_f.cubin: lal_pppm.cu lal_precision.h lal_preprocessor.h \
$(OBJ_DIR)/pppm_f_cubin.h: $(OBJ_DIR)/pppm_f.cubin
$(BIN2C) -c -n pppm_f $(OBJ_DIR)/pppm_f.cubin > $(OBJ_DIR)/pppm_f_cubin.h
rm $(OBJ_DIR)/pppm_f.cubin
$(OBJ_DIR)/pppm_d.cubin: lal_pppm.cu lal_precision.h lal_preprocessor.h \
lal_pre_cuda_hip.h
@ -54,12 +53,10 @@ $(OBJ_DIR)/pppm_d.cubin: lal_pppm.cu lal_precision.h lal_preprocessor.h \
$(OBJ_DIR)/pppm_d_cubin.h: $(OBJ_DIR)/pppm_d.cubin
$(BIN2C) -c -n pppm_d $(OBJ_DIR)/pppm_d.cubin > $(OBJ_DIR)/pppm_d_cubin.h
rm $(OBJ_DIR)/pppm_d.cubin
$(OBJ_DIR)/%_cubin.h: lal_%.cu $(ALL_H)
$(CUDA) --fatbin -DNV_KERNEL -o $(OBJ_DIR)/$*.cubin $(OBJ_DIR)/lal_$*.cu
$(BIN2C) -c -n $* $(OBJ_DIR)/$*.cubin > $@
@rm $(OBJ_DIR)/$*.cubin
# host code compilation

3
lib/gpu/Nvidia.makefile_multi
View File

@ -29,7 +29,7 @@ $(OBJ_DIR):
# Compilers and linkers
CUDA = $(NVCC) $(CUDA_INCLUDE) $(CUDA_OPTS) -Icudpp_mini $(CUDA_ARCH) \
CUDA = $(NVCC) $(CUDA_INCLUDE) $(CUDA_OPTS) $(CUDA_ARCH) \
$(CUDA_PRECISION)
CUDR = $(CUDR_CPP) $(CUDR_OPTS) $(CUDA_PRECISION) $(CUDA_INCLUDE) \
$(CUDPP_OPT)
@ -54,7 +54,6 @@ $(OBJ_DIR)/pppm_d_cubin.h: $(OBJ_DIR)/pppm_d.cubin
$(OBJ_DIR)/%_cubin.h: lal_%.cu $(ALL_H)
$(CUDA) --fatbin -DNV_KERNEL -o $(OBJ_DIR)/$*.cubin $(OBJ_DIR)/lal_$*.cu
$(BIN2C) -c -n $* $(OBJ_DIR)/$*.cubin > $@
@rm $(OBJ_DIR)/$*.cubin
# host code compilation