Improve memory usage when calculating normalization integrals#139
Merged
Conversation
…ions; in many places reduce size of event indices from long or long long to int, which should be sufficient
…f events rather than all of them
…ks instead of for the entire data set at once; this is intended to reduce memory consumption
…entire data set, which reduces memory consumption
…nk size to increment the factor arrays; improve some debugging outputs
…a sets copied to the GPU; adjust initData call order to be compatible with GPU operations
…data into memory in addition to data being loaded into memory on the follower nodes; use new loadMC method to control single process and MPI loading; change some unsigned long int to unsigned int
…e GPU in parallel with AmpVecs data sharing; NormIntInterface cleans up memory allocated for terms after gen MC integral calculation
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 is a number of changes that are focused on reducing memory (GPU and CPU) use. Some existing inefficiencies were eliminated. The major algorithmic change is to compute the normalization integrals for the generated MC in batches at the end of a fit. There is currently no way to toggle off/on this behavior -- it is on by default. The code should be completely backwards compatible with existing fitters. One needs to just recompile and relink with this new version of the AmpTools core library.