Open Hub
Black Duck Open Hub

 Settings |  Report Duplicate

0
I Use This!
Very Low Activity

Commits : Listings

Analyzed 1 day ago. based on code collected 2 days ago.
Jun 19, 2023 — Jun 19, 2024
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Added solve/tt_gmres and related routines (tt_iterapp, tt_axpy2, und so weiter...) Specially for Kazeev =)
dc
More... almost 13 years ago
Merge branch '1.07' Conflicts: solve/dmrg_solve2.m Now it has both "verb" as a parameter, stopping criterium via local residual, and returns number of sweeps conducted
dc
More... almost 13 years ago
some files I forgot in the previous commit
dc
More... almost 13 years ago
test_steps2 and test_full_KN2 - for Poisson equation tests/blockts_video - for nice videos for timedep. (+ a lot of images for Poisson) deleted: exp/tt_kron2.m - conflicts with core/tt_kron2 modified: solve/dmrg_solve2.m - stopping criterium via local residual
dc
More... almost 13 years ago
Certain files required to run the "newbab" benchmark
Ivan Oseledets
More... almost 13 years ago
Merge branch 'master' of 83.149.207.38:TT
Ivan Oseledets
More... almost 13 years ago
"Full" was not fully correct in the case of r(1) not equal to 1
Ivan Oseledets
More... almost 13 years ago
Added exp/parmldivide mex - for solving parallel dense systems [A1,A2,...].
Sergey Dolgov
More... almost 13 years ago
Parametric system solution + Babuska example. Huge need for a block diagonal solver
Ivan Oseledets
More... almost 13 years ago
It seems, it works (I mean, dmrg_parb)
Ivan Oseledets
More... almost 13 years ago
dmrg_parb is closer to the working state (but still works only for physical dimension equal to 1)
Ivan Oseledets
More... almost 13 years ago
Various fixes around basic functions, to make it work with the "block" TT-tensors. Also fixed a nasty bug in the ttm function
Ivan Oseledets
More... almost 13 years ago
Added two working functions, useful in block computations: add two TT-tensors to a block columnwise and rowwise, i.e. you can you [a;b] and [a,b] now Enjoy. The file dmrg_parb is now in the development, not functional, and added only not to mix with merges
Ivan Oseledets
More... almost 13 years ago
Small modification to dmrg_solve2 output Now if verb is set to false (it is now an input argument!!) then gmres does not produce junk into the channel (like I did not converge etc) tt_tensor now also is more silent in the case accuracy is absent
Ivan Oseledets
More... almost 13 years ago
Added simple function for the reorthogonalization of a new basis to a given one
Ivan Oseledets
More... almost 13 years ago
Added (untested) routine to perform tensor-by-matrix product
Ivan Oseledets
More... almost 13 years ago
Added test file for the henon-heiles potential
Ivan Oseledets
More... about 13 years ago
Small fixes: now you can write a+b, kron(a,b) where one of the operands is empty. Helpful for ``long'' summations. Also added function axpx which computes the product of A and projects in into the basis spanned by X (i.e., one step of the ALS). This is required for time-stepping subroutines
Ivan Oseledets
More... about 13 years ago
Minor
Ivan Oseledets
More... about 13 years ago
Some minor modifications Added chunk.m function - it can take a contigious chunk of the train
Ivan Oseledets
More... about 13 years ago
Added @tt_tensor/col.m for the case r(d+1) is not 1. Helpful for block methods. kron for tt_tensors has been made slightly faster, since now it does not depend on the TT1.0 code dmrg_eigb modified, however we SHOULD THINK: The current truncation destroys the orthogonality of the approximation Sometimes the convergence even stucks into the solution with collinear columns! The approximation problem has a nice matrix formulation (ask me if you are interested). Or you can just check something like dot(col(v,1), col(v,2)) on your solution
Ivan Oseledets
More... about 13 years ago
Added cg_trunc - wround by nonlinear cg =) test: tests/test_cgtrunc
dc
More... about 13 years ago
Few more tweaking of dmrg_eigb
Ivan Oseledets
More... about 13 years ago
Cleaned minor bugs in dmrg_eigb.m Added test_eigb for the creation of the FP matrix and computing its nullspace. However, it seems that B'*B matrix IS NOT positive definite anymore. Maybe even it is non-symmetric (?) However, the produced solution seems to be more (or less) adequate
Ivan Oseledets
More... about 13 years ago
Merge branch 'master' of 83.149.207.38:TT
Ivan Oseledets
More... about 13 years ago
Cleaned up dmrg_eigb.m Now it returns the solution as a TT-tensor with the blocking index pushed up to the right (i.e. r(d+1)=k). This seems reasonable, and certain subroutines have to be modified to work correctly in cases where r(1) and/or r(d+1) are not equal to 1
Ivan Oseledets
More... about 13 years ago
Cleaned up dmrg_eigb.m Now it returns the solution as a TT-tensor with the blocking index pushed up to the right (i.e. r(d+1)=k). This seems reasonable, and certain subroutines have to be modified to work correctly in cases where r(1) and/or r(d+1) are not equal to 1
Ivan Oseledets
More... about 13 years ago
More numerics on the dumbbell
Sergei Dolgov
More... about 13 years ago
Merge branch '8.06'
dc
More... about 13 years ago
Some new tests on the dumbbell + newton rank truncation (in principle it sucks, esp. for high ranks; but may be it could be modified to greedy alg.)
dc
More... about 13 years ago