Sameer Agarwal comments

Results 143 comments of


                                            Sameer Agarwal

How to use cuda_sparse as the sparse linear algebra library in solving BAL Problems with 2.20 version?

This means that you did not compile Ceres with CUDA linked into it. Please take a look at your cmake log and you will find that cuda was not detected...

How to use cuda_sparse as the sparse linear algebra library in solving BAL Problems with 2.20 version?

ah I did not realize what linear solver you were trying to use. I should have paid more attention to the fact that you are using SPARSE_NORMAL_CHOLESKY. We currently do...

How to use cuda_sparse as the sparse linear algebra library in solving BAL Problems with 2.20 version?

We just added support for cuDSS, you should be able to do this now.

BlockSparseMatrix might benefit from caching column-block partition

cc: @DmitriyKorchemkin

MinGW Compilation fails in debug mode - file too big

@sergiud I think you fixed this right?

Use 64bit integers in ceres internals to enable the solution of larger problems

Sorry but I have not had time to work on this. Our current focus is to modernize the code base and improve the support for linear solvers (GPU and CPU)....

Refactor parallel_for.h/cc parallel_for_cxx.h/cc

cc: @DmitriyKorchemkin

The program crashes when executing solve()

This is a crash inside eigen's sparse cholesky factorization routine. Beyond that it's hard to say why this crash is happening without knowing more about the structure of the problem...

The program crashes when executing solve()

The destructor calling the destructor again does not seem right. This looks like some kind of linking problem, this should not happen. Are you sure that you are not mixing...

The program crashes when executing solve()

Closing because of lack of update.