* Also added some more openmp directives to increase percentage of parallelized code.
* "implicit none" was missing in two subroutines of homogenization and constitutive.
0 : only version infos and all from "hypela2"/"umat"
1 : basic outputs from "CPFEM.f90", basic output from initialization routines, debug_info
2 : extensive outputs from "CPFEM.f90", extensive output from initialization routines
3 : basic outputs from "homogenization.f90"
4 : extensive outputs from "homogenization.f90"
5 : basic outputs from "crystallite.f90"
6 : extensive outputs from "crystallite.f90"
7 : basic outputs from the constitutive files
8 : extensive outputs from the constitutive files
If verbosity is equal to zero, all counters in debug are not set during calculation (e.g. debug_StressLoopDistribution or debug_cumDotStateTicks). This might speed up parallel calculation, because all these need critical statements which extremely slow down parallel computation.
In order to keep it like that, please follow these simple rules:
DON'T use implicit array subscripts:
example: real, dimension(3,3) :: A,B
A(:,2) = B(:,1) <--- DON'T USE
A(1:3,2) = B(1:3,1) <--- BETTER USE
In many cases the use of explicit array subscripts is inevitable for parallelization. Additionally, it is an easy means to prevent memory leaks.
Enclose all write statements with the following:
!$OMP CRITICAL (write2out)
<your write statement>
!$OMP END CRITICAL (write2out)
Whenever you change something in the code and are not sure if it affects parallelization and leads to nonconforming behavior, please ask me and/or Franz to check this.
* removed input variables in constitutive_collectDotState and constitutive_postResults that are not needed anymore (because of recent changes in constitutive_nonlocal)
Now it is possible to compile a single precision spectral solver/crystal plasticity by replacing mesh.f90 and prec.f90 with mesh_single.f90 and prec_single.f90.
For the spectral method, just call "make precision=single" instead of "make". Use "make clean" evertime you switch precision
First try of implement single precision crystal plasticity, not working yet.
polishing text about geometry construction.
polishing postResults, still having problems concerning machines without MSC installation
* dislocation flux and internal stress calculation now consistent with new definition of slip system lattice according to paper (polarity of screws inverted)
* now complaining when encountering an unknown nonlocal parameter in material.config
* use same error ID for all material parameters out of bounds
* symmetric flux calculation in side dotState can now be omitted (because of new treatment of periodicity)
* switching back to "local flux balance" (add leaving and entering fluxes at central MP, don't touch neighbor) instead of "flux distribution" (subtract leaving fluxes from central MP and add them at neighboring MP). This has the advantage that there is almost no need for CRITICAL statements in parallelization, so hopefully this results in some speed up.
To enable this feature one has to add the following somewhere in the marc input file:
$mpie periodic x y z
for having periodicity in all directions
$mpie periodic z x
for having periodicity in x and z direction
etc.
Note that this only works for regular meshes!!!
postprocessing: renamed name of python/f2py modul from "reconstruct" to "postprocessingMath", added some numerical operations to use for postprocessing.
* need to recalculate dislocation velocity in postResults, otherwise we take values of last perturbed state! So the following outputs were up to now showing the perturbed state: shearrate, dislocation velocity, all density rates!
mpie_spectral and numerics: added switch to prevent pre calculation of gamma_hat. slower, but saves memory
3Dvisualize: started to add support for gmsh (not fully working yet)
reconstruct: new version of f2py/Fortran subroutines for output of results from spectral method
removed storage of full cauchy stres field from mpie_spectral.f90, only average is stored now
added cauchy stress and von mises equivalent calculation to spectral post.
renamed mpie_spectral2.f90 to mpie_spectral2d.f90 (testing file, not properly working at the moment)
changed file extension and variable names in mpie_spectral.f90 and mpie_spectral_interface.f90 from "mesh" to "geom". Removed direct output from mpie_spectral.f90, all output is now base on materialpoint_results(:,1,:)
* default value of the OMP_NUM_THREADS variable has to be restored at the end of mpie subroutine, since marc also seems to use and change(!) this
* usage: "export MPIE_NUM_THREADS=<number of threads>" to set variable in shell, then restart mentat and compile with option 3 (at the moment this does only work on ws 6, since all other workstations use compiler option "-save"; this puts all local variables by default in static memory, which is a killer for parallelization!)
* better use SINGLE (having an implicit barrier at the end) instead of MASTER construct
* deleted all explicit BARRIERs after do loops since parallel loop construct implies barrier at the end
* had to add some BARRIER constructs
* only the master thread is allowed to increase the state counter
yet parallelization seems not to give a significant decrease in calculation time with nonlocal model (because of too many CRITICAL statements?)
* also put a call to constitutive_microstructure at the start of each crystallite_integration subroutine like it was before. need that for nonlocal model in case of crystallite cutback
numerics: polishing
mpie_cpfem_marc: polishing
..powerlaw: aware of symmetryType function
crystallite: aware of symmetryType function, smaller leapfrog acceleration
IO: new warning 101
CPFEM: range of odd stress is now -1e15...+1e15, H_sym is used for stiffness
Major changes:
CPFEM.f90 =>
1. Moving the initialization out of CPFEM_general into a separate subroutine, which is directly called by the hypela2 (Beware, the Abaqus version must also be modified in order to adapt with this change).
2. Restore primary state variables in CPFEM_init from binary files when requested (Marc flag: restart read).
3. Writing primary state variables into binary files (Marc flag: restart write).
FEsolving.f90 =>
1. Adding functions to recognize Marc restart flags: read and write and the corresponding restart file (parent job).
2. Change the initial value of cycleCounter = -1 in conjuction with the change made the ping-pong scheme
homogenization_RGC.f90 =>
1. Just syntax polishing.
IO.f90 =>
1. Adding functions/subroutines to open binary files for writing the primary state variables for restart purpose.
mpie_cpfem_marc.f90
1. Modification of the general scheme for collection and calculation in order to accommodate the newly added restart feature.
* in Fixed Point Iteration: update dependent states after state preguess was missing; on the other hand, the first call to constitutive_microstructure was obsolete
* fluxes are now again calculated and distributed only! by the originating material point. this means that the central MP might change the dotState of its neighbor. have to see whether locks slow down parallel computation
* detection of grain boundary in constitutive_nonlocal_microstructure with the help of transmissivity
* enforce positive densities in constitutive_nonlocal_microstructure (needed because dotState does not create cutbacks for negative densities anymore)
* reset single mobile densities below certain threshold to zero (also done in constitutive_nonlocal_microstructure)
* constitutive_nonlocal_kinetics only gets local state variable as input, no need for the entire array here
* dv_dtau is always positive
* multiplication is only active when there is already some initial density of the respective type
added mpie_spectral2.f90, a version that should get the new algorithm proposed in 2010. until now, it is the same as mpie_spectral.f90 (large strain formulation by suquet et al) but with c2c, c2c FFT
added some parameters for spectral method to numerics.f90 (tolerance)
changed error message concerning spectral method in IO.f90
corrected calculation of stress BC in mpie_spectral.f90
* now remembering stiffness similar to how we do it for Lp etc.; avoids undefined stiffness values for nonconverged stiffness calculation
* non-local stuff:
* changed non-local kinetics (Gilman2002)
* enforce zero shearrate for overall carrrier density below relevant density
* enforce zero density for those states that become negative and were below relevant density before
* dislocation velocity is not limited by V^(1/3) / dt anymore
mpie_spectral.f90 uses both functions
math_Plain99to3333 is used for inversion of c0
math_mul33x3_complex is used for equilibrium check in fourier space
also did some cleaning up on mpie_spectral.f90
2) local stiffness calculation is now standard for non-local grains
3) stressLoopDistribution discriminates between (a) central solution and (b) stiffness perturbation
4) debugger is switched on as standard... (but verboseDebugger not!)
changes on mpie_spectral.f90:
new structure, changed variable names, now using defgrad instead of disgrad, cleaned up, removed augmented Lagrange.
ToDo: Implement Augmented Lagrange again (but then a working version), implement Large strain, think about complex-to real-transform backwards, try to implement MP-support
mesh:
elemType identification based on lower case
Abaqus now reports more errors
IO:
new function to inquire whether inputfile contains "parts"
new function to assemble multiply included inputfile into a flat one
awareness of range generation in element numbers
error reporting
started to commend Ricardos code, layoutet loops and removed some redundant variables.
until now, no error calculation is done. at the moment calculations are in an infinite loop
rather perturb all components at once (and optionally decrease the frequency of the Jacobian update with the iJaco parameter) than perturbing only a single component per cycle
restructured nonlocal_dotstate, to be able to easily switch on and off particular effects in the microstructure evolution
nonlocal_dotstate now enforces a cutback when single density runs the risk of becoming negative; in the case of a state already below the relevantState dotState is set to zero
introduced two new output variables: rho_dot_edge and rho_dot_screw
So switched back to the former order of calls which also works for marc2010:
el 1 IP 1 | el 1 IP 2
|
inc mode lovl | inc mode lovl
------------------ | ------------------
0 2 6 | 0 2 6
1 6 4 | 1 6 4
1 4 6 | 1 3 6
1 6 4 | 1 6 4
1 1 6 | 1 2 6
Still to be checked:
according to Franz the lovl order could be different in contact analysis (e.g. starting with lovl 4 instead of 6)
- grainrotation calculation now is done with symmetryID 0, i.e. without symmetry reduction since we want the absolute misorientation.
- While math has everything in radians, post results eulerangles and axisangle are given in degrees.
And: grain rotation seems OK after the previous changes in math module.
beware that crystallite output "orientation" now by default returns the orientation as quaternion. if you want euler angles instead, you have to add "eulerangles" as a crystallite output in your material.config file (see material.config template).
for input of orientations in the texture block of the material.config you still have to specify the rotation in terms of euler angles, quaternions are not yet supported for input.
- nonlocal stiffness calculation: we perturb all material points at the same time, so instead of N^2 loops we just need N
- set "forceLocalStiffnessCalculation" to false as standard
- flux density interpolation now depends on the position of the interface between ttwo neighboring material points
- simplified flux calculation scheme
- introduced sanity check for dislocation velocity to ensure v*dt< cellsize
--> new "crystallite" part in config file
--> new "crystallite" option for microstructures
--> new output file "...job.outputCrystallite" to be used in conjunction with marc_addUserOutput for meaningful naming of User Defined Vars.
- reworked contribution of immobile dislocation density for rate equations
- flux is now calculated on the basis of interpolated velocities and densities at the interface; both incoming and outgoing fluxes are considered, so every material point only changes his own dotState
- dislocation velocity is now globally defined and calculated by subroutine constitutive_nonlocal_kinetics; the subroutine is called inside _LpAndItsTangent as well as _microstructure; therefore, microstructure now needs Tstar_v as additional input; in the future one should perhaps create a subroutine constitutive_kinetics that calls constitutive_nonlocal_kinetics separately, to clearly distinguish between microstructural and kinetic variables
- better use flux density vector as output variable instead of scalar flux values for each interface
- added output variables internal and external resolved stress
crystallite:
- added flag to force local stiffness calculation in case of nonlocal model
- misorientation angle is explicitly set to zero when no neighbor can be found
debug:
- added flag "selectiveDebugger" that is used when debugging statements should only affect a specific element, ip and grain; these are specified with the new variables debug_e, debug_i and debug_g
- debugger can now be used in its original sense
- remobilization of immobile singles immediately increases the mobile single dislocation density and therefore directly affects all other mechanisms in dotState
- changed nomenclature (rho -> rhoSgl) to distinguish precisely between single dislocation density and total dislocation density
- changed material.config accordingly
The calculation of the misorientation is now done once in crystallite init and at the end of every FE increment. This saves a lot of time compared to doing it in dotState for every crystallite subinc.
- corrected flux term
- multiplication is now aware of dislocation type
- corrected change rate for "dipole size" dupper
- corrected term for dipole dissociation by stress change
- added transmissivity term in fluxes which accounts for misorientation between two neighboring grains (yet hardcoded transmissivity according to misorientation angle)
- added more output variables
constitutive:
- 2 additional variables "previousDotState" and "previousDotState2", which are used to store the previous and second previous dotState (used in crystallite for acceleration/stabilization of state integration)
- timer for dotState now measures the time for calls to constitutive_ collectState (used to reside in crystallite_updateState, which is not critical in terms of calculation time anymore)
crystallite:
- convergence check for nonlocal elments is now done at end of crystallite loop, not at the beginning; we simple set all elements to not converged if there is at least one nonlocal element that did not converge
- need call to microstructure before first call to collect dotState for dependent states
- stiffness calculation (jacobian): if there are nonlocal elements, we also have to consider changes in our neighborhood's states; so for every perturbed component in a single ip, we have to loop over all elements; since this is extremely time-consuming, we just perturb one component per cycle, starting with the one that changes the most during regular time step.
- updateState gets a damping prefactor for our dotState that helps to improve convergence; prefactor is calculated according to change of dotState
IO:
- additional warning message for unknown crystal symmetry
- new subroutine "math_misorientation" which calculates misorientations (based on old subroutine "misorientation", which was not used previously)
- rendered some functions pure
previously was:
if (the_sum /= 1.0_pReal) then (error 170)
this condition is too strict. it may give problem with some numerical truncation error.
now becomes:
if (abs(the_sum - 1.0_pReal) >= 1.0e-10_pReal) then (error 170)
(1) subStepSizeHomog and subStepSizeCryst := size of substep when cut-back is applied (initially was hard-coded).
(2) stepIncreaseHomog and stepIncreaseCryst := step increase when calculation for substep converge (was also hardcoded).
introduced a possibility to choose different finite difference scheme, i.e., forward-, backward- and central-difference, for computing grain numerical tangent. note that central-difference scheme will slow down the computation significantly. please use it only if necessary.
parameters to set these new features have been included in numerics.f90 and numerics.config, whereas corresponding error messages have been introduced in the IO.f90
in constitutive_nonlocal.f90:
Derivatives of shear rates w.r.t. resolved shear stress HAVE to be positive.
Computation of dgdot_dtauslip is now correct.
cleaning up of twin system nomenclature
introduced lookup mechanism to calculate the twinning shears depending on the c/a ratio of the present structure (was [wrongly!!] hardcoded to Titanium)
- read in activation energy for dislocation glide from material.config
- changed naming of dDipMin/Max to dLower/dUpper
- added new outputs: rho_dot, rho_dot_dip, rho_dot_gen, rho_dot_sgl2dip, rho_dot_dip2sgl, rho_dot_ann_ath, rho_dot_ann_the, rho_dot_flux, d_upper_edge, d_upper_screw, d_upper_dot_edge, d_upper_dot_screw
- poisson's ratio is now calculated from elastic constants
- microstrucutre has state as first argument, since this is our output variable
- periodic boundary conditions are taken into account for fluxes and internal stresses. for the moment, flag has to be set in constitutive_nonlocal.
- corrected calculation for dipole formation by glide
- added terms for dipole formation/annihilation by stress decrease/increase
constitutive:
- passing of arguments is adapted for constitutive_nonlocal model
crystallite:
- in stiffness calculation: call to collect_dotState used wrong arguments
- crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults
homogenization:
- crystallite_postResults uses own Tstar_v and temperature, no need for passing them from materialpoint_postResults
IO:
- changed error message 229
material.config:
- changed example for nonlocal constitution according to constitutive_nonlocal
all:
- added some flush statements
crystallite does not accelerate anymore, since, typically, longer step immediately fails and uses resources in vain. (future: remember number of successful steps to increase step size after x of those...)
Christoph Kords