SHELL = /bin/sh
########################################################################################
# Makefile to compile the Material subroutine for BVP solution using spectral method
########################################################################################
# Be sure to remove all files compiled with different options by using "make clean"
#
# Uses OpenMP to parallelize the material subroutines
# (set number of threads with "export DAMASK_NUM_THREADS=n" to n)
#
# Install fftw3 (v3.3 is tested):
# + execute
# ./configure --enable-threads --enable-sse2 --enable-shared
# make
# make install
# + specify in the "pathinfo:FFTW" where FFTW was installed.
# We essentially look for two library files "lib/libfftw3_threads" and "lib/libfftw3",
# so you can copy those, for instance,
# into DAMASK_ROOT/lib/fftw/lib/ and specify "./fftw/" as pathinfo:FFTW
# Uses linux threads to parallelize fftw3
#
# Instead of the AMD Core Math Library or the Intel Kernel Math Library
# a standard "lib(64)/liblapack.a/dylib/etc." can be used.
# leave pathinfo:ACML and pathinfo:IKML blank, but specify as pathinfo:LAPACK where the library folder is located
########################################################################################
# OPTIONS = standard (alternative): meaning
#-------------------------------------------------------------
# F90 = ifort (gfortran): compiler, choose Intel or GNU
# COMPILERNAME = overwrite name of Compiler, e.g. using mpich-g90 instead of ifort
# PORTABLE = TRUE (FALSE): decision, if executable is optimized for the machine on which it was built.
# OPTIMIZATION = DEFENSIVE (OFF,AGGRESSIVE,ULTRA): Optimization mode: O2, O0, O3 + further options for most files, O3 + further options for all files
# OPENMP = TRUE (FALSE): OpenMP multiprocessor support
# FFTWROOT = pathinfo:FFTW (will be adjusted by setup_code.py - required in pathinfo)
# IMKLROOT = pathinfo:IMKL (will be adjusted by setup_code.py if present in pathinfo)
# ACMLROOT = pathinfo:ACML (will be adjusted by setup_code.py if present in pathinfo)
# LAPACKROOT = pathinfo:LAPACK (will be adjusted by setup_code.py if present in pathinfo)
# PREFIX = arbitrary prefix
# SUFFIX = arbitrary suffix
# STANDARD_CHECK = checking for Fortran 2008, compiler dependend
########################################################################################
#any values will be overwritten by setup_code.py
FFTWROOT ?= auto
IMKLROOT ?= auto
ACMLROOT ?= auto
LAPACKROOT ?= auto
F90?=ifort
COMPILERNAME ?= $(F90)
INCLUDE_DIRS +=-I$(DAMASK_ROOT)/lib
ifeq "$(FASTBUILD)" "YES"
OPENMP := OFF
OPTIMIZATION := OFF
else
OPENMP ?= ON
OPTIMIZATION ?= DEFENSIVE
endif
ifeq "$(OPTIMIZATION)" "OFF"
OPTI := OFF
MAXOPTI := OFF
endif
ifeq "$(OPTIMIZATION)" "DEFENSIVE"
OPTI := DEFENSIVE
MAXOPTI := DEFENSIVE
endif
ifeq "$(OPTIMIZATION)" "AGGRESSIVE"
OPTI := AGGRESSIVE
MAXOPTI := DEFENSIVE
endif
ifeq "$(OPTIMIZATION)" "ULTRA"
OPTI := AGGRESSIVE
MAXOPTI := AGGRESSIVE
endif
ifndef OPTI
OPTI := DEFENSIVE
MAXOPTI := DEFENSIVE
endif
ifeq "$(PORTABLE)" "FALSE"
PORTABLE_SWITCH =-msse3
endif
# settings for multicore support
ifeq "$(OPENMP)" "ON"
OPENMP_FLAG_ifort =-openmp -openmp-report0 -parallel
OPENMP_FLAG_gfortran =-fopenmp
ACML_ARCH =_mp
LIBRARIES +=-lfftw3_threads -lpthread
endif
LIBRARIES +=-lfftw3
LIB_DIRS +=-L$(FFTWROOT)/lib
ifdef IMKLROOT
LIB_DIRS +=-L$(IMKLROOT)/lib/intel64
LIBRARIES +=-mkl
else
ifdef ACMLROOT
LIB_DIRS +=-L$(ACMLROOT)/$(F90)64$(ACML_ARCH)/lib
LIBRARIES +=-lacml$(ACML_ARCH)
else
ifdef LAPACKROOT
LIB_DIRS +=-L$(LAPACKROOT)/lib64 -L$(LAPACKROOT)/lib
LIBRARIES +=-llapack
endif
endif
endif
ifdef PETSC_DIR
include ${PETSC_DIR}/conf/variables
INCLUDE_DIRS +=${PETSC_FC_INCLUDES} -DPETSc
LIBRARIES +=${PETSC_WITH_EXTERNAL_LIB}
endif
ifdef STANDARD_CHECK
STANDARD_CHECK_ifort =$(STANDARD_CHECK)
STANDARD_CHECK_gfortran =$(STANDARD_CHECK)
endif
ifneq "$(FASTBUILD)" "YES"
STANDARD_CHECK_ifort ?=-stand f08 -standard-semantics -warn stderrors
STANDARD_CHECK_gfortran ?=-std=f2008 -fall-intrinsics
endif
#-fall-intrinsics: all intrinsic procedures (including the GNU-specific extensions) are accepted. This can be useful with -std=f95 to force standard-compliance
#-std=f2008ts: for newer gfortran
# but get access to the full range of intrinsics available with gfortran. As a consequence, -Wintrinsics-std will be ignored and no user-defined
# procedure with the same name as any intrinsic will be called except when it is explicitly declared external
#-Wintrinsics-std: warnings because of "flush" is not longer in the standard, but still an intrinsic fuction of the compilers
OPTIMIZATION_OFF_ifort :=-O0 -no-ip
OPTIMIZATION_OFF_gfortran :=-O0
OPTIMIZATION_DEFENSIVE_ifort :=-O2
OPTIMIZATION_DEFENSIVE_gfortran :=-O2
OPTIMIZATION_AGGRESSIVE_ifort :=-O3 $(PORTABLE_SWITCH) -no-prec-div -fp-model fast=2 -ipo #-static causes trouble at the moment
OPTIMIZATION_AGGRESSIVE_gfortran :=-O3 $(PORTABLE_SWITCH) -ffast-math -funroll-loops -ftree-vectorize
COMPILE_OPTIONS_ifort :=-fpp\
-ftz\
-assume byterecl
ifneq "$(FASTBUILD)" "YES"
COMPILE_OPTIONS_ifort :=$(COMPILE_OPTIONS_ifort)\
-implicitnone\
-diag-enable sc3\
-diag-disable 5268\
-warn declarations\
-warn general\
-warn usage\
-warn interfaces\
-warn ignore_loc\
-warn alignments\
-warn unused\
-warn errors
endif
#-fpp: preprocessor
#-ftz: flush unterflow to zero, automatically set if O<0,1,2,3> >0
#-fimplicit-none: assume "implicit-none" even if not present in source
#-diag-disable: disables warnings, where
# warning ID 5268: the text exceeds right hand column allowed on the line (we have only comments there)
#-warn: enables warnings, where
# declarations: any undeclared names
# general: warning messages and informational messages are issued by the compiler
# usage: questionable programming practices
# interfaces: checks the interfaces of all SUBROUTINEs called and FUNCTIONs invoked in your compilation against an external set of interface blocks
# ignore_loc: %LOC is stripped from an actual argument
# alignments: data that is not naturally aligned
# unused: declared variables that are never used
# errors: warnings are changed to errors
# stderrors: warnings about Fortran standard violations are changed to errors (STANDARD_CHECK)
#
###################################################################################################
#MORE OPTIONS FOR DEBUGGING DURING COMPILING
#-warn: enables warnings, where
# truncated_source: Determines whether warnings occur when source exceeds the maximum column width in fixed-format files. (too many warnings because we have comments beyond character 132)
# uncalled: Determines whether warnings occur when a statement function is never called
# all:
#
#OPTIONS FOR DEGUBBING DURING RUNTIME
# information on http://software.intel.com/en-us/articles/determining-root-cause-of-sigsegv-or-sigbus-errors/
#-g: Generate symbolic debugging information in the object file
#-traceback: Generate extra information in the object file to provide source file traceback information when a severe error occurs at run time.
#-gen-interfaces: Generate an interface block for each routine. http://software.intel.com/en-us/blogs/2012/01/05/doctor-fortran-gets-explicit-again/
#-fp-stack-check: Generate extra code after every function call to ensure that the floating-point (FP) stack is in the expected state.
#-check: checks at runtime, where
# bounds: check if an array index is too small (<1) or too large!
# arg_temp_created: will cause a lot of warnings because we create a bunch of temporary arrays
# format: Checking for the data type of an item being formatted for output.
# output_conversion: Checking for the fit of data items within a designated format descriptor field.
# pointers: Checking for certain disassociated or uninitialized pointers or unallocated allocatable objects.
# uninit: Checking for uninitialized variables.
#-heap-arrays: should not be done for OpenMP, but set "ulimit -s unlimited" on shell. Probably it helps also to unlimit other limits
#-fpe-all0 capture all floating-point exceptions, sets -ftz automatically
###################################################################################################
COMPILE_OPTIONS_gfortran :=-xf95-cpp-input
ifneq "$(FASTBUILD)" "YES"
COMPILE_OPTIONS_gfortran :=$(COMPILE_OPTIONS_gfortran)\
-ffree-line-length-132\
-fimplicit-none\
-pedantic\
-Warray-bounds\
-Wampersand\
-Wno-tabs\
-Wcharacter-truncation\
-Wintrinsic-shadow\
-Waliasing\
-Wconversion\
-Wsurprising\
-Wunderflow\
-Wswitch\
-Wstrict-overflow\
-Wattributes\
-Wunsafe-loop-optimizations\
-Wunused\
-Wall\
-Wextra\
-Wsuggest-attribute=const\
-Wsuggest-attribute=pure\
-Wsuggest-attribute=noreturn\
-Wreal-q-constant\
-pedantic-errors
endif
#-xf95-cpp-input: preprocessor
#-ffree-line-length-132: restrict line length to the standard 132 characters
#-fno-range-check: disables checking if result can be represented by variable. Needs to be set to enable DAMASK_NaN
#-fimplicit-none: assume "implicit-none" even if not present in source
#-pedantic: more strict on standard, enables some of the warnings below
#-Warray-bounds: checks if array reference is out of bounds at compile time. use -fcheck-bounds to also check during runtime
#-Wampersand: checks if a character expression is continued proberly by an ampersand at the end of the line and at the beginning of the new line
#-Wno-tabs: do not allow tabs in source
#-Wcharacter-truncation: warn if character expressions (strings) are truncated
#-Wintrinsic-shadow: warn if a user-defined procedure or module procedure has the same name as an intrinsic
#-Waliasing: warn about possible aliasing of dummy arguments. Specifically, it warns if the same actual argument is associated with a dummy argument with "INTENT(IN)" and a dummy argument with "INTENT(OUT)" in a call with an explicit interface.
#-Wconversion: warn about implicit conversions between different type
#-Wsurprising: warn when "suspicious" code constructs are encountered. While technically legal these usually indicate that an error has been made.
#-Wunderflow: produce a warning when numerical constant expressions are encountered, which yield an UNDERFLOW during compilation
#-Wswitch: warn whenever a "switch" statement has an index of enumerated type and lacks a "case" for one or more of the named codes of that enumeration. (The presence of a "default" label prevents this warning.) "case" labels outside the enumeration range also provokewarnings when this option is used (even if there is a "default" label)
#-Wstrict-overflow:
#-Wattributes: warn about inappropriate attribute usage
#-Wunsafe-loop-optimizations: warn if the loop cannot be optimized due to nontrivial assumptions.
#-Wunused:
# -value:
# -parameter: find usused variables with "parameter" attribute
#-Wextra:
#-Wreal-q-constant: warn about real-literal-constants with 'q' exponent-letter
###################################################################################################
#MORE OPTIONS FOR DEBUGGING DURING COMPILING
#-Wline-truncation: too many warnings because we have comments beyond character 132
#-Warray-temporarieswarnings: because we have many temporary arrays (performance issue?):
#-Wimplicit-interface:
#-pedantic-errors:
#-fmodule-private:
#
#OPTIONS FOR DEGUBBING DURING RUNTIME
#-fcheck-bounds: check if an array index is too small (<1) or too large!
#-ffpe-trap=invalid,\ stop execution if floating point exception is detected (NaN is silent)
# zero,\
# overflow,\
# underflow,\
# precision,\
# denormal
##################################################################################################
PRECISION_ifort :=-real-size 64 -integer-size 32 -DFLOAT=8 -DINT=4
#-real-size 32: set precision to one of those 32/64/128 (= 4/8/16 bytes) for standard real (=8 for pReal)
#-integer-size 16: set precision to one of those 16/32/64 (= 2/4/8 bytes) for standard integer (=4 for pInt)
PRECISION_gfortran :=-fdefault-real-8 -fdefault-double-8 -DFLOAT=8 -DINT=4
#-fdefault-real-8: set precision to 8 bytes for standard real (=8 for pReal). Will set size of double to 16 bytes as long as -fdefault-double-8 is not set
#-fdefault-double-8: set precision to 8 bytes for double real, would be 16 bytes because -fdefault-real-8 is used
#-fdefault-integer-8: Use it to set precision to 8 bytes for integer, don't use it for the standard case of pInt=4 (there is no -fdefault-integer-4)
###################################################################################################
COMPILE =$(OPENMP_FLAG_$(F90)) $(COMPILE_OPTIONS_$(F90)) $(STANDARD_CHECK_$(F90)) $(OPTIMIZATION_$(OPTI)_$(F90)) $(INCLUDE_DIRS) $(PRECISION_$(F90)) -DSpectral
COMPILE_MAXOPTI =$(OPENMP_FLAG_$(F90)) $(COMPILE_OPTIONS_$(F90)) $(STANDARD_CHECK_$(F90)) $(OPTIMIZATION_$(MAXOPTI)_$(F90)) $(INCLUDE_DIRS) $(PRECISION_$(F90)) -DSpectral
###################################################################################################
COMPILED_FILES = prec.o DAMASK_spectral_interface.o IO.o numerics.o debug.o math.o \
FEsolving.o mesh.o material.o lattice.o \
constitutive_dislotwin.o constitutive_j2.o constitutive_phenopowerlaw.o \
constitutive_titanmod.o constitutive_nonlocal.o constitutive_none.o constitutive.o crystallite.o \
homogenization_RGC.o homogenization_isostrain.o homogenization.o CPFEM.o \
DAMASK_spectral_utilities.o DAMASK_spectral_solverBasic.o
ifdef PETSC_DIR
PETSC_FILES = DAMASK_spectral_solverAL.o DAMASK_spectral_solverBasicPETSc.o
COMPILED_FILES += $(PETSC_FILES)
endif
DAMASK_spectral.exe: DAMASK_spectral_driver.o
$(PREFIX) $(COMPILERNAME) $(OPENMP_FLAG_$(F90)) $(OPTIMIZATION_$(MAXOPTI)_$(F90)) $(STANDARD_CHECK_$(F90)) \
-o DAMASK_spectral.exe DAMASK_spectral_driver.o \
$(COMPILED_FILES) $(LIB_DIRS) $(LIBRARIES) $(SUFFIX)
DAMASK_spectral_driver.o: DAMASK_spectral_driver.f90 DAMASK_spectral_solverBasic.o $(PETSC_FILES)
$(PREFIX) $(COMPILERNAME) $(COMPILE_MAXOPTI) -c DAMASK_spectral_driver.f90 $(SUFFIX)
DAMASK_spectral_solverAL.o: DAMASK_spectral_solverAL.f90\
DAMASK_spectral_utilities.o
DAMASK_spectral_solverBasic.o: DAMASK_spectral_solverBasic.f90\
DAMASK_spectral_utilities.o
DAMASK_spectral_solverBasicPETSc.o: DAMASK_spectral_solverBasicPETSc.f90\
DAMASK_spectral_utilities.o
DAMASK_spectral_utilities.o: DAMASK_spectral_utilities.f90\
CPFEM.o
CPFEM.o: CPFEM.f90\
homogenization.o
homogenization.o: homogenization.f90\
homogenization_RGC.o \
homogenization_isostrain.o
homogenization_RGC.o: homogenization_RGC.f90 \
crystallite.o
homogenization_isostrain.o: homogenization_isostrain.f90 \
crystallite.o
crystallite.o: crystallite.f90 \
constitutive.o
constitutive.o: constitutive.f90 \
constitutive_nonlocal.o \
constitutive_titanmod.o \
constitutive_dislotwin.o \
constitutive_phenopowerlaw.o \
constitutive_j2.o \
constitutive_none.o
constitutive_nonlocal.o: constitutive_nonlocal.f90 \
lattice.o
constitutive_titanmod.o: constitutive_titanmod.f90 \
lattice.o
constitutive_dislotwin.o: constitutive_dislotwin.f90 \
lattice.o
constitutive_phenopowerlaw.o: constitutive_phenopowerlaw.f90 \
lattice.o
constitutive_j2.o: constitutive_j2.f90 \
lattice.o
constitutive_none.o: constitutive_none.f90 \
lattice.o
lattice.o: lattice.f90 \
material.o
material.o: material.f90 \
mesh.o
mesh.o: mesh.f90 \
FEsolving.o \
math.o
FEsolving.o: FEsolving.f90 \
debug.o
math.o: math.f90 \
debug.o
debug.o: debug.f90 \
numerics.o
numerics.o: numerics.f90 \
IO.o
IO.o: IO.f90 \
DAMASK_spectral_interface.o
DAMASK_spectral_interface.o: DAMASK_spectral_interface.f90 \
prec.o
ifeq "$(F90)" "gfortran" # fno-range-check because NaN is defined in prec
prec.o: prec.f90
$(PREFIX) $(COMPILERNAME) $(COMPILE) -c -fno-range-check prec.f90 $(SUFFIX)
else
prec.o: prec.f90
$(PREFIX) $(COMPILERNAME) $(COMPILE) -c prec.f90 $(SUFFIX)
endif
%.o : %.f90
$(PREFIX) $(COMPILERNAME) $(COMPILE) -c $< $(SUFFIX)
tidy:
@rm -rf *.o
@rm -rf *.mod
clean:
@rm -rf *.o
@rm -rf *.mod
@rm -rf *.exe