getMembraneTractionNonTensor modified to use neo-Hookean constitutive equation for membrane stress.
Sign error corrected in getMembraneTractionNonTensor.m
Bending component included in getMembraneTractionNonTensor.m
getMembraneTractionNonTensor.m modified to use the "singularity-free" formulation.
VesicleInfFlow runs through with getMembraneTractionNonTensor that uses divgcm. Haven't tested things like area/volume/force balance with the results.
Non-debug versions of mex files (had debug flags on in last commit)
Just a small error in vhagcm, pu and pv were switched around.
vhagc working in from fortran test code (see DivergenceTest.f90), but not from MEX. Why?
divgcm.mexw64 committed
getMembraneTractionNonTensor.m added, but not working because it seems like divergence is not being computed correctly.
Building mex for vhagcm and divgcm to calculate divergence. These two mexes have not been tested yet.
never committed integrateForce.m and testForceIntegral.m
getVolumeArea.m was never committed
MATLAB 2019a with Intel Fortran 19.0 has default /integer-size:64 option. This is a problem for spherepack (where variables are not explicitly declared) and other places where variables are declared as "integer" to mean "integer*4". There is no way to override this (see discussion in https://www.mathworks.com/matlabcentral/answers/474520-overriding-compiler-switch-in-mex?s_tid=prof_contriblnk). So had to write a Makefile to do this.
Two more mex binaries added; this is not good, need a better way for this.
number of longitude points changed to 2*nlat - 1 (i.e., nlat = N+1, nlon = 2*N + 1) in
Error corrections in get_9NodeQuad_GtKu_nonsingular.F90 and get_9NodeQuad_GtKu_singular.F90
In BEM direct, instead of forming matrices and then multiplying, we try to multiply the kernel by vector then integrate. get_9nodeQuad_BEM_GtKu.m is only partly done. Ali is going to complete this.
Updated check_conservation.m for parallelization
Tried to use parfor to parallelize in solveBIE.m; this actually hurts performance. Probably MATLAB was doing enough multithreading that using parfor in fact hurt. Going to revert to non-parallel version after this commit.
Created
Bug fix: In VesicleInfFlow.m, was using undeformed geometry to calculate uinf; corrected to use deformed geometry.
Fixed a bug in VesicleInfFlow.m (was passing undeformed instead of deformed configuration to solveBIE).
Modify solveBIE.m to rotations the new way.
Modifying to avoid multiple masks speeds up a bit more. For N = 20, speedup is 10x. For N = 40, speedup is about 6x.
Rotation of SH coefficients is the most time-consuming part. Try new organization for doing this calculation in testSHCoeffRotation.m. A bit better (2-3x faster), but may be avoiding multiple masks can give further speedup.
VesicleInfFlow.m modified: rotation matrices can be computed once upfront (and stored in a cell array)
Radical changes...
Correction: Added minus sign in getSHBIEGK.f90
A long time since the last commit. Apparently, ...
A brute force implementation of the constant vo...
Monitor volume conservation in visualizeRBC.m. ...
Runs to 1s with 1e-2 time step when ED is set t...
Activated dealiasing because computation was no...
Modified runRBCsimulation.m to work with Stress...
Modified StressResultantShell.m so that undefor...
G computation is modified (in getSHBIEGK.F90) s...
Error fixed in nullspace (normal vector) calcul...
Put constitutive equations in. Results for
Forgot to save changes in ShellFormulation.tex ...
Modified stress-resultant shell implementation ...
Tried to do derivative of (sqrt(a) s_ll) becaus...
Christoffel symbol for ellipsoid does not work ...
Computing Christoffel symbol using derivative o...
Got an idea to test, at least computation of a_...
Stress resultant shell formulation added. Not t...
Reverted get_9nodequad_solid_angle.m to revisio...
committed from CCR machine, put timing in diffe...
There were 2 problems.
Moved matrix rearranging into spmd, doing distr...
Modified runRBCsimulation.m to use spmd instead...
Using a 3 parfor structure with 2 sets of cell ...
Tried to break the construction of G and K into...
Modified runRBCsimulation.m with revised call t...
ABAQUS test works correctly, so KK appears to b...
Modified shagsd.f, actually quite simple. Gives...
The plan is to make a new version of shags, cal...
Added shagsim and shagsm for precomputing assoc...
Made a new mex, shagsm. Changed shagcm call to ...
Modified getSHBIEGK.F90 to compute only lower t...
Wrote computation of G and K matrices in mex.
Modified ExtStokesFlow.m and getSHBIEmatrices.m...
Reverted ExtStokesFlow.m and getSHBIEmatrices.m...
Added some plotting in getSHBIEmatrices.m to vi...
Reverted ExtStokesFlow.m and getSHBIEmatrices.m...
Using shagcm does not work well. Since G is no ...
getSHBIEmatrices.m modified and checked that us...
Took ExtStokesFlow.m and getSHBIEmatrices.m to ...
Still messing around with getSHBIEmatrices.m.
Modified only ExtStokesFlow.m. Using Gauss quad...
SparseGrid doesn't work for the RBC geometry (o...
Modified runRBCsimulation.m to run one time ste...
Everything works in CCR. Ran on CCR, got rbcsim...
modifications to run on CCR cluster, commited f...
getSHBIEmatrices.m was screwed up in the last c...
Went back to Matlab version of getSHBIEmatrices...
Tried to rewrite getSHBIEmatricesMEX.F90, still...
Revised getSHBIEmatrices.m in an attempt to mak...
PuttingItTogether works (just had the wrong sig...
Put the RBC forces in as well, but the RBC is d...
Started PuttingItTogether!
Rewrote computing G and K in getSHBIEmatrices.m...
Tested ExtStokesFlow.m with undeformed/deformed...
Modified ExtStokesFlow.m in BIEsphere to also c...
ExtStokesFlow.m modified to use SHRotateMEX -> ...
1. SHRotate mexed to SHRotateMEX.f90; speedup m...
New traction conditions for inlet and outlet ed...
Fortan for both singular and nonsingular parts ...
The nested functions in get_9nodeQuad_BEM_matri...
Orienting normal vectors made general in BEMdir...
Boundary integral with spherical harmonics is w...
Committing for solid angle stuff.
Added solid angle computation
1. Checked RBC force calculation using finite d...
Number of major changes in shell formulation (R...
Working on LearnSphericalHarmonics directory, n...
Did force balance check
Modified BEMdirect to treat traction unknowns a...
PatranImport modified --- reads INLET, OUTLET a...
PatranImport modified --- reads INLET and OUTLE...