Skip to content

[WIP] Allow NUTS to do eager evaluation on forward and backward trajectory in parallel #3103

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 11 commits into
base: develop
Choose a base branch
from
Open

[WIP] Allow NUTS to do eager evaluation on forward and backward trajectory in parallel #3103

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
e4ab35b
Merge branch 'develop' of https://github.com/stan-dev/stan into develop
May 7, 2018
54b927c
Merge branch 'develop' of https://github.com/stan-dev/stan into develop
Jun 24, 2018
cc44118
Merge branch 'develop' of https://github.com/stan-dev/stan into develop
wds15 Feb 11, 2022
586989f
prototype speculative NUTS
wds15 Feb 11, 2022
5d9a09a
get rid of mutex when sampling uniforms for parallel nuts
SteveBronder Feb 15, 2022
7669f6d
clang format
SteveBronder Feb 15, 2022
4cda135
update with some comments
SteveBronder Feb 15, 2022
dfdc55a
tests running for parallel adaptation
SteveBronder Feb 16, 2022
f18ae74
clang format
SteveBronder Feb 16, 2022
ff668c9
update to not use get_num_theads() to detect thread level
SteveBronder Feb 17, 2022
334d242
clang format
SteveBronder Feb 17, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 13 additions & 6 deletions src/stan/mcmc/hmc/nuts/adapt_diag_e_nuts.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,18 +12,25 @@ namespace mcmc {
* with a Gaussian-Euclidean disintegration and adaptive
* diagonal metric and adaptive step size
*/
template <class Model, class BaseRNG>
class adapt_diag_e_nuts : public diag_e_nuts<Model, BaseRNG>,
template <class Model, class BaseRNG, bool ParallelBase = false>
class adapt_diag_e_nuts : public diag_e_nuts<Model, BaseRNG, ParallelBase>,
public stepsize_var_adapter {
public:
template <bool ParallelBase_ = ParallelBase,
std::enable_if_t<!ParallelBase_>* = nullptr>
adapt_diag_e_nuts(const Model& model, BaseRNG& rng)
: diag_e_nuts<Model, BaseRNG>(model, rng),
: diag_e_nuts<Model, BaseRNG, ParallelBase>(model, rng),
stepsize_var_adapter(model.num_params_r()) {}

~adapt_diag_e_nuts() {}
template <bool ParallelBase_ = ParallelBase,
std::enable_if_t<ParallelBase_>* = nullptr>
adapt_diag_e_nuts(const Model& model, std::vector<BaseRNG>& thread_rngs)
: diag_e_nuts<Model, BaseRNG, ParallelBase>(model, thread_rngs),
stepsize_var_adapter(model.num_params_r()) {}

sample transition(sample& init_sample, callbacks::logger& logger) {
sample s = diag_e_nuts<Model, BaseRNG>::transition(init_sample, logger);
inline sample transition(sample& init_sample, callbacks::logger& logger) {
sample s = diag_e_nuts<Model, BaseRNG, ParallelBase>::transition(
init_sample, logger);

if (this->adapt_flag_) {
this->stepsize_adaptation_.learn_stepsize(this->nom_epsilon_,
Expand Down
Loading