Local state-space iteration at order 3 with and without pruning

assigned to @normann

changed the description

added 15 commits

5bab00ee...584e661f - 10 commits from branch Dynare:master
ed4ac470 - Local state space iteration at order 3: single-thread 2nd-order version in Fortran
0830d0e6 - Local state space iteration at order 3: multi-thread 2nd-order version in Fortran
02429533 - Local state space iteration at order 3: multi-thread 2nd-order version in...
70444f73 - Local state space iteration at order 3: attempted optimization of Fortran...
79627239 - Local state-space iteration at order 3: multi-thread 3rd-order version with and without pruning

added 7 commits

79627239...0107365f - 2 commits from branch Dynare:master
07b7fd2a - Local state space iteration at order 3: single-thread 2nd-order version in Fortran
5ca68194 - Local state space iteration at order 3: multi-thread 2nd-order version in Fortran
6d451a13 - Local state space iteration at order 3: multi-thread 2nd-order version in...
1bfa4536 - Local state space iteration at order 3: attempted optimization of Fortran...
14948890 - Local state-space iteration at order 3: multi-thread 3rd-order version with and without pruning

Thanks @normann, it looks good overall.

A few questions/remarks:

this MR has 5 commits, but since some commits seem to undo what previous ones have done, it looks to me like they should rather be merged into a single commit
I wonder why pparticle_3.F08 is part of libkordersim. It is used only by the 3rd order particle filter iteration, so I have the feeling that it should rather be moved out of the library and directly into the MEX source (either remaining a separate source file, or merged into the main one, whatever seems more logical to you). Actually I realize that the same remark applies to pparticle.F08 (for k-order)
the changes to the MATLAB code (for actually using the new MEX) are not included in this MR, but I guess this is for a further one?

assigned to @sebastien and unassigned @normann

marked this merge request as draft

added 31 commits

14948890...9e467fcf - 26 commits from branch Dynare:master
dc52c1ba - Local state space iteration at order 3: single-thread 2nd-order version in Fortran
0a41796b - Local state space iteration at order 3: multi-thread 2nd-order version in Fortran
063ff416 - Local state space iteration at order 3: multi-thread 2nd-order version in...
c2c8ea1d - Local state space iteration at order 3: attempted optimization of Fortran...
cfeae5bd - Local state-space iteration at order 3: multi-thread 3rd-order version with and without pruning

added 1 commit

added 1 commit

2a1ff3fc - Integrates the local_state_space_iteration_3 mex into dynare's bayesian estimation eco-system

added 1 commit

84e44d64 - Integrates the local_state_space_iteration_3 mex into dynare's bayesian estimation eco-system

added 65 commits

84e44d64...982ec2e6 - 57 commits from branch Dynare:master
3b74372a - Local state space iteration at order 3: single-thread 2nd-order version in Fortran
d8dd6517 - Local state space iteration at order 3: multi-thread 2nd-order version in Fortran
c5f41a65 - Local state space iteration at order 3: multi-thread 2nd-order version in...
263e0a2d - Local state space iteration at order 3: attempted optimization of Fortran...
97ec4ee3 - Local state-space iteration at order 3: multi-thread 3rd-order version with and without pruning
b2af0671 - Fix: merging pruning latent variables into one
0fe060e4 - Integrates the local_state_space_iteration_3 mex into dynare's bayesian estimation eco-system
c82a874c - Local state-space iteration at order 3: optimizes memory management

added 1 commit

c6d5c48f - Local state-space iteration at order 3: multi-thread 3rd-order version with and without pruning

marked this merge request as ready

mentioned in merge request !2071 (merged)

mentioned in commit 4bfd687b

merged

Activity