Cluster sequence return type #38
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The tiling parameters are internal to the TiledAlgoLL strategy, so they should not by part of the ClusterSequence object. Factorise them out and pass the tiling explicitly to the functions that need it in the TiledAlgoLL reconstruction.
Now the calling functions need to call the inclusive_jets() function with their pt cut Adapt the example reco and the testcases to cope with this (N2Plain still is TODO)
This is to be used by multiple algorithms now
We need to return the same object from all algorithms, which is the ClusterSequence. This allows us to implement post-facto filtering, pruning, cleaning, etc. steps after the initial reconstruction is done. Explicitly convert to use PseudoJet internally, as the ClusterSequence stores these objects. Internal algorithm is now renamed with _ as it should not be used by users. Maintain compact arrays for the reconstruction process, but not for the jets themselves - these are stored into the ClusterSequence linearly, keeping an index pointer (clusterseq_index) to the correct position from the compact array. History of recombination is filled appropriately for both cluster-cluster merges and for beam (final) merge steps.
This is now working properly, with a ClusterSequence being the common return type
Make the --profile option take an argument, which is the sub-directory where the output will be stored, which allows different profiles to be more easily compared
These cause a real slowdown in the runtime (~20% degradation) even when DEBUG isn't enabled There are quite a few complaints about this in the forums, e.g., JuliaLang/julia#28147. It does seem to have improved, but for code like this it's still a serious regression
Use the @__DIR__ macro and the joinpath() function
Automate the unpacking of LZMA compressed data files for tests (this needs "xzcat" to be present on the system). Ensure files are closed on reading events.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR makes the return type of all algorithms the FastJet-ish$d_{ij}$ values, etc.).
ClusterSequencewhich encapsulates all of the intermediate reconstruction steps and the record of cluster merges (parents, children,This unifies the
N2TiledandN2Plainreturn values,Closes #29
Closes #36
The tiling parameter has been factorised out of the
ClusterSequenceas this is an internal parameter forN2Tiled.Closes #37
For the
N2Plainalgorithm, we now exclusively usePseudoJetas the cluster structure type - conversions are made from in the publicplain_jet_reconstructfunction toPseudoJetbefore calling the internal implementation function_plain_jet_reconstruct.The
PseudoJetvector, of all clusters, is now held inside the cluster sequence, and an extra vector is used which maps the compact index (used for heavy calculations) to the relevant index of the cluster sequence jets.Note that
@debugcode needs to be completely commented out as even having it in the algorithm causes noticeable performance degradations (see JuliaLang/julia#28147, even though with mitigations the performance still drops by a noticeable amount).To help with performance testing, the
--profileoption now takes an argument, which controls where the profiling output is written - this allows different versions of the code to be looked at side-by-side.